Family 8335+06 IBM Power System AC922 optimised for Linux workloads

IBM Canada Sales Manual
Revised: September 8, 2021


Table of contents
Product life cycle datesProduct life cycle datesTechnical descriptionTechnical description
AbstractAbstractPublicationsPublications
HighlightsHighlightsFeaturesFeatures
DescriptionDescriptionAccessoriesAccessories
ModelsModelsSuppliesSupplies


Product life cycle dates

Top rule

Type Model Announced Available Marketing Withdrawn Service Discontinued
8335-GTX 2018-05-08 2018-05-25 - -


Back to topBack to top

Abstract

Top rule

The IBM Power System AC922 server is co-designed with OpenPOWER Foundation ecosystem members for the demanding needs of deep learning and AI, high-performance analytics, and high-performance computing users. It will be deployed in the most powerful supercomputers on the planet through a partnership between IBM, NVIDIA, and Mellanox, among others.

The Power AC922 server delivers up to six Tesla V100 with NVLink GPUs with 16 GB (#EC4H) and 32 GB (#EC4K) local GPU memory versions, supported in two processor sockets. The server is only offered in an air-cooled system design. It offers a 36-core at 3.15 GHz (3.45 GHz turbo) or 44-core at 2.80 GHz (3.10 GHz turbo) IBM POWER9 NVLink 2.0 technology configuration in a 19-inch rack-mount, 2U (EIA units) drawer configuration. All the cores are activated.

The Power AC922 server provides two hot-swap and redundant power supplies (for all GPU configurations) and sixteen DIMM memory slots. Supported memory features are 8 GB (#EM60), 16 GB (#EM61), 32 GB (#EM63), 64 GB (#EM64), and 128 GB (#EM65), allowing for a maximum system memory of over 2 TB.

With the availability of CUDA 11.0, Red Hat Enterprise Linux (RHEL) 7.6, or later, is supported with POWER9 based servers with GPUs. This support is expected to continue with subsequent RHEL 8 releases which have Extended Update Support (EUS). More information on the RHEL lifecycle and EUS releases is available from Red Hat:

  • https://access.redhat.com/support/policy/updates/errata

The Power AC922 server also offers:

  • High throughput and performance for high-value Linux workloads such as LAMP, BDA, or industry applications
  • Up to six Tesla V100 with NVLink GPUs with 16 GB and 32 GB GPU memory
  • Potentially low acquisition cost through system optimization (industry standard memory, limited configurations, limited I/O and expansion, and industry standard warranty)
  • Multiple I/O options in the system unit, including:
    • Two PCIe x16 4.0 LP slots, CAPI enabled
    • One PCIe x16 4.0 LP slot, CAPI enabled or one PCIe x8x8 shared 4.0 slot, CAPI enabled
    • One PCIe x4 4.0 LP slot
  • Two SFF bays for up to two HDDs or SSDs
  • Integrated SATA controller (does not support RAID functionality)
  • Optional Mellanox InfiniBand or high-performance Ethernet
  • One front and one rear USB 3.0 port
  • 19-inch rack-mount 2U configuration
  • Operating systems:
    • Red Hat Enterprise Linux 7.6 for Power LE (POWER9), or later

Model abstract 8335-GTX

The next generation of IBM Power Systems, with POWER9 technology, is built with innovations that deliver high speeds for the AI era. Our Linux scale-out systems embrace acceleration for the next generation of datacenter computing and help provide differentiated performance, scalability, and potentially low acquisition cost. The IBM Power System AC922 for High Performance Computing provides:

  • 2x POWER9 CPUs with next generation NVLink
  • Water-cooled server design
  • A differentiated platform for GPU acceleration and superior I/O with PCIe Gen 4 and CAPI 2.0
  • POWER9 with Next generation NVLink Technology: Only processor with NVLink from CPU to GPU
  • GPUs are connected to CPUs coherently, enabling simpler GPU acceleration and access to system memory for larger model size


Back to topBack to top

Highlights

Top rule

The IBM Power System AC922 (8335-GTX) server is an AI, supercomputing, powerful 2-socket server that offers 36 or 44 fully activated cores and I/O configuration flexibility to help meet accelerated computing needs. The server features are designed for high-performance data analytics and high-performance computing workloads with up to six Tesla V100 with NVLink GPUs in a water-cooled system design.

The Power AC922 model offers:

  • Two IBM POWER9 Single Chip Module (SCM) processors that offer high performance with 36 or 44 fully activated cores
  • Up to six Tesla V100 with NVLink GPUs with 16 GB and 32 GB GPU memory
  • Up to 1 TB of memory
  • Four PCIe slots and three CAPI enabled for future CAPI-enabled devices (a maximum of three CAPI devices can be used concurrently)
  • Two 2.5-inch SATA drives for a maximum of 4 TB hard disk drive (HDD) or 7.68 TB of solid-state drive (SDD) storage
  • Two integrated USB 3.0 ports
  • Two hot-swap and redundant power supplies: (for all 4 GPU configurations) 2200 W 200 - 240 and 277 V AC
  • 19-inch rack-mount hardware (2U)


Back to topBack to top

Description

Top rule

The Power AC922 server delivers up to six Tesla V100 with NVLink GPUs with 16 GB (#EC4H) and 32 GB (#EC4K) local GPU memory versions, supported in two processor sockets. The server is only offered in an water-cooled system design. It offers a 36-core at 3.15 GHz (3.45 GHz turbo) or 44-core at 2.80 GHz (3.10 GHz turbo) IBM POWER9 NVLink 2.0 technology configuration in a 19-inch rack-mount, 2U (EIA units) drawer configuration. All the cores are activated. The Power AC922 server provides two hot-swap and redundant power supplies (for all GPU configurations) and sixteen DIMM memory slots. Supported memory features are 8 GB (#EM60), 16 GB (#EM61), 32 GB (#EM63), 64 GB (#EM64), and 128 GB (#EM65), allowing for a maximum system memory of over 2 TB.

Summary of standard features for the Power AC922 server:

  • Power Systems server built with POWER9 processor modules
    • 18-core, 3.15 GHz (3.45 GHz turbo)
    • 22-core, 2.80 GHz (3.10 GHz turbo)
  • High-performance DDR4 memory
    • 128 GB (#EM60 x16), 256 GB (#EM61 x16), 512 GB (#EM63 x16), 1.02 TB (#EM64 x16), or 2.04 TB (#EM65 x16) memory features
  • Storage bays
    • Two SFF 4-bays for up to two HDDs or SSDs (No mixing of HDD and SSD drives allowed)
    • Up to two 1 TB 7200 RPM SATA disk drives (#ELD0)
    • Up to two 2 TB 7200 RPM SATA disk drives (#ES6A)
    • Up to two 960 GB SATA SSD (#ELU4)
    • Up to two 1.92 TB SATA SSD (#ELU5)
    • Up to two 3.84 TB SATA SSD (#ELU6)
  • PCIe Gen4 slots
    • Two PCIe x16 4.0 LP slots, CAPI enabled
    • One PCIe x16 4.0 LP slot, CAPI enabled or one PCIe x8x8 shared 4.0 slot, CAPI enabled
    • One PCIe x4 4.0 LP slot
  • Compute Intensive Accelerator
    • Four or six 16 GB SXM2 NVIDIA Tesla V100 GPUs with NVLink Water-Cooled (#EC4H)
    • Four or six 32 GB SXM2 NVIDIA Tesla V100 GPUs with NVLink Water-Cooled (#EC4K)
  • Integrated:
    • One front USB 3.0 port and one rear USB 3.0 port
  • Two hot-swap redundant power supplies (for all GPU configurations)
  • 19-inch rack-mount hardware (2U)

Power AC922 (8335-GTX) system configuration

The minimum Power AC922 initial order must include two processor modules, 128 GB of memory, two power supplies, two line cords, rack- mounting hardware, a system software indicator, a rack integrator specify, and a Language Group Specify.

Linux is the operating system. The minimum defined initial order configuration is as follows:

Feature       Description
number
------------- ----------------------------------------------------------
EP0Q x 2      18-core 3.15 GHz (3.45 GHz Turbo) POWER9 Processor Module
or
EP0S x 2      22-core 2.80 GHz (3.10 GHz Turbo) POWER9 Processor Module
EM60 x 16     8 GB DIMMs, 2666 MHz, 4 Gb DDR4 DRAM
EC16 x 1      Open Power Abstraction Layer (OPAL)
EC4H x 4      NVIDIA Tesla V100 GPU with NVLink Water-Cooled (16 GB)
or
EC4K x 4      NVIDIA Tesla V100 GPU with NVLink Water-Cooled (32 GB)
4650 x 1      Rack Indicator -- Not Factory Integrated
EB2X x 2      AC Power Supply, 2200 Watt (200-240V/277V)
2 power cords Select two power cords from supported list. Feature EPAM
              is defaulted. (#EPAM is Power Cord 4.3 m (14-ft), Drawer
              to Wall/IBM PDU (250V/16A))
2147 x 1      Primary OS Linux
EJTX x 1      Rack-mount fixed rail kit
93xx x 1      Language Group Specify (select one from announced
              features)
EL3Z          PCIe2 LP 2-port 10/1GbE BaseT RJ45 Adapter

Note: Adapter feature EL3Z is the default 10/1Gb Ethernet adapter. Other options of a 10 Gb Ethernet adapter include one of either EC2R, EC2T, EN0T, or EN0V.
Note: If a rack is wanted, it must be ordered as an MTM rack on initial system orders. If on the same system order, it will be shipped at the same time in the same shipment, but in separate packing material. IBM does not offer IBM Manufacturing rack integration of the server into the rack before shipping at this time.

Processor modules

A minimum of two identical processor modules are required with eighteen processor cores (#EP0Q) or twenty-two processor cores (#EP0S) allowed. All processor cores are activated.

There are no processor activation features used or orderable on the 8335-GTX. All processor cores/"n-ways" are always fully activated.

Mixing of different processor features on the same system is not allowed.

System memory

There are 16 memory slots in the server. Each memory slot can hold one DDR4 memory feature. A minimum of 16 memory features is required, filling all the memory slots. With 8 GB DIMMs, 128 GB of memory is required. With 16 GB DIMMs, 256 GB of memory is required. With 32 GB DIMMs, 512 GB of memory is required. With 64 GB DIMMs, 1.02 TB is required. With 128 GB DIMMs, 2.04 TB of memory is required.
Note: Mixing of memory features is not allowed.

Care should be exercised in mixing DIMMS of different size, density, and type in the system. For additional information refer to Knowledge Center memory plugging rules at https://www.ibm.com/support/knowledgecenter/8335-GTH/p9eik/p9eik_memor y_p lacemnt.htm

DIMM memory features

Feature name         Feature Maximum
                     number  quantity
-------------------- ------- --------
8 GB DDR4 2666 MHz   EM60    16
16 GB DDR4 2666 MHz  EM61    16
32 GB DDR4 2666 MHz  EM63    16
64 GB DDR4 2666 MHz  EM64    16
128 GB DDR4 2666 MHz EM65    16

I/O adapters

The system contains the following adapter slots:

  • Two PCIe x16 G4 HHHL slots, CAPI enabled
  • One PCIe x16 G4HHHL slot, CAPI enabled or one PCIe x8x8 shared 4.0 slot, CAPI enabled
  • One PCIe x4 G4 HHHL slot

Only LP adapters can be placed in LP slots. An x8 adapter can be placed in an x16 slot, but an x16 adapter cannot be placed in an x8 slot.

Power supply

  • Two hot-swap and redundant power supplies (for all GPU configurations): 2200 W 200-240V/277V AC (2 x #EB2X)

Power cords

Two power cords are required. Two feature EPAM cords are defaulted. See the feature listing for other options.

Reliability, Availability, and Serviceability

The Power AC922 server brings POWER9 processor and memory RAS functionality into a cloud data centre with Open Source Linux technology supplying the operating system and virtualization. The OPAL firmware provides a hypervisor and operating system-independent layer that exploits the error detection and self-healing functions built into the POWER9 processor.

The processor address paths and data paths, the control logic, state machines, and computational units are protected with parity or error correction codes (ECC). The processor core soft errors or intermittent errors are recovered with processor instruction retry. Unrecoverable errors are reported as machine check (MC) while errors that affect the integrity of data lead to system check-stop.

The Level 1 (L1) data and instruction caches in each processor core are parity protected and data are stored through to L2 immediately. L1 caches have a retry capability for intermittent errors and a cache set delete mechanism for handling solid failures. The L2 and L3 caches in the POWER9 processor are protected with double-bit detect, single-bit correct ECC. In addition, a threshold of correctable errors detected on cache lines can result in the data in the cache lines being purged and the cache lines removed from further access without requiring a reboot. An uncorrectable error detected in these caches can also trigger a purge and delete of cache lines. This does not impact the current operation if the cache lines contained data unmodified from what was stored in system memory.

The memory subsystem has proactive memory scrubbing to help prevent the accumulation of multiple single-bit errors. The ECC scheme can correct the complete failure of any one memory module within an ECC word. After marking the module as unusable, the ECC logic can still correct single symbol (two adjacent bit) errors. An uncorrectable error of data of any layer of cache up to the main memory is marked to prevent usage of fault data. The processor's memory controller has retry capabilities for certain fetch and store faults. Special Uncorrectable Error handling

Special Uncorrectable Error (SUE) handling prevents an uncorrectable error in memory or cache from immediately causing the system to terminate. Rather, the system tags the data and determines whether it will ever be used again. If the error is irrelevant, it will not force a checkstop. If the data is used, termination may be limited to the program/kernel or hypervisor owning the data; or freeze of the I/O adapters controlled by an I/O hub controller if data would be transferred to an I/O device.

Thermal management, current/voltage monitoring

The On Chip Controller (OCC) monitors various temperature sensors in the processor module, memory modules, and environmental temperature sensors and steers the throttling of processor cores and memory channels should the temperature rise over thresholds defined by the design. The power supplies have their own independent thermal sensors and monitoring. Power supplies and voltage regulator modules monitor over-voltage, under-voltage, and over-current conditions. They report into a power good tree that is monitored by the Service Processor.

PCI extended error handling

PCI extended error handling (EEH)-enabled adapters respond to a special data packet generated from the affected PCI slot hardware by calling system firmware, which will examine the affected bus, allow the device driver to reset it, and continue without a system reboot. For Linux, EEH support extends to the majority of frequently used devices, although some third-party PCI devices may not provide native EEH support.

Graphics processing unit (GPU) acceleration

GPUs are attached with second-generation NVLink to the system power processors and provide cache coherence capabilities. The GPUs should be run in compute mode, which enables Single Error Correction and Double Error Detection (SECDEC) Error Correction Code (ECC) for the GPU Memory, SM register file, L1 cache, and L2 cache to improve data integrity for GPU-accelerated workloads.

Chassis policy after input power loss and autorestart after system-check-stop

The boot parameter "chassis policy" controls whether the server returns to the previous state or powers up axiomatically after an input power loss. The system automatically reboots after a system- check-stop, and it is up to the system management software to decide whether to use the server with potentially fewer resources.

Serviceability

The server is designed for system install and setup, feature install and remove, proactive maintenance, and corrective repair performed by the client: Warranty Service Upgrades are offered for an On Site Repair (OSR) by an IBM System Services Representative (SSR) or an authorized warranty service provider.

IBM Knowledge Center provides up-to-date documentation to effectively service the system with:

  • Quick Install Guide
  • User's Guide
  • Troubleshooting Guide
  • Boot Configuration Guide

The documentation can be downloaded in PDF format or used on-line with an internet connection.

Service Processor

The Service Processor supports the Intelligent Platform Management Interface (IPMI 2.0) and Data Center Management Interface (DCMI 1.5) and Simple Network Management Protocol (SNMP V2 and V3) for system monitoring and management. The Service Processor provides platform system functions such as power on/off, power sequencing, power fault monitoring, power reporting, fan/thermal control, fault monitoring, VPD inventory collection, Serial over LAN (SOL), Service Indicator LED management, code update, and event reporting through system event logs (SEL).

All SELs can be retrieved either directly from the Service Processor or from the host operating system (Linux). The Service Processor monitors the operation of the firmware during the boot process and also monitors the hypervisor for termination. The firmware code update is supported through the Service Processor and IPMI interface. The firmware image can be updated or flashed regardless of its current state.

Service interface

The service interface enables the client and the support personnel to communicate with the service support applications in a server by connecting directly or remotely through a web browser or command-line interface. It provides access to various service applications and available actions. The service interface enables client and support personnel to manage system resources, inventory, and service information in an efficient and effective way.

Different service interfaces are used, depending on the state of the system and its operating environment. The primary service interfaces are:

  • Service processor: Ethernet Service Network with IPMI version 2.0, systems management GUI through web browser
  • Service indicator LEDs: System attention and system identification (front and back)
  • Host operating system: Command-line interface

The primary service applications are:

  • System event logs (SEL)
  • Operating system event logs
  • Sensor status GUI LEDs for Problem Determination (PD) when next to the system, locally

Concurrent maintenance

The following components can be replaced without powering off the server:

  • Hard drives
  • Fans

Error handling and reporting

In the event of a system hardware failure or environmentally induced failure, the system error capture capability systematically analyzes the hardware error signature to help determine the cause of failure. The processor and memory recoverable errors are handled through Processor Runtime Diagnostics (PRD) in the OPAL layer and generate a system event log (SEL). An extended SEL (eSEL) is associated with each SEL. It contains additional First Failure Data Capture (FFDC) information for use by the support structure.

For system check-stop errors, the On Chip Controller (OCC) collects Failure Information Register (FIR) data and saves it in nonvolatile memory. PRD analyzes the data upon reboot and creates a SEL and an eSEL. The host Linux operating system can monitor the SELs on the service processor through the IPMI tool. Hardware and firmware failures are recorded in the SELs and can be retrieved through the IPMI interface. The system has the ability to report errors associated with PCIe adapters/devices through the host operating system.

Warranty and spare parts

The system comes with a three-year warranty on parts. The replacement parts can be ordered through the Advanced Part Exchange Warranty Service.

Section 508 of the US Rehabilitation Act

Power AC922 (8335-GTX) is capable as of May 25 2018, when used in accordance with IBM's associated documentation, of satisfying the applicable requirements of Section 508 of the Rehabilitation Act, provided that any assistive technology used with the product properly interoperates with it. A US Section 508 Voluntary Product Accessibility Template (VPAT) can be found on the Product accessibility information website.



Back to topBack to top

Models

Top rule

Model summary matrix

Model Processor Speed Memory
GTX 18-, 22-core 2.6 GHz (3.09 GHz turbo) to 2.0 GHz (2.87 GHz turbo) 128 GB to up to 2 TB

Customer setup (CSU)

Yes.

Devices supported

IBM 7226-1U3 Multi-Media Storage Enclosure.

  • LTO Ultrium Fiber Channel (FC) Tape Drive feature numbers
    • #8348 LTO Ultrium 6 FC tape drive. Capacity up to 6.25 TB
    • #8446 LTO Ultrium 7 FC tape drive. Capacity up to 15 TB
    • #8546 LTO Ultrium 8 FC tape drive. Capacity up to 30 TB
    • DVD-RAM USB 3.0 drive feature #5757
    • RDX (USB 3.0) Removable hard drives. Drive capacity: 500 GB, 1 TB, 2 TB

Model conversions

Not available.



Back to topBack to top

Technical description

Top rule

Physical specifications Physical specifications Hardware requirements Hardware requirements
Operating environment Operating environment Software requirements Software requirements
Limitations Limitations


The Power AC922 server delivers up to six Tesla V100 with NVLink GPUs with 16 GB (#EC4H) and 32 GB (#EC4K) local GPU memory versions, supported in two processor sockets. The server is only offered in an water-cooled system design. It offers a 36-core at 3.15 GHz (3.45 GHz turbo) or 44-core at 2.80 GHz (3.10 GHz turbo) IBM POWER9 NVLink 2.0 technology configuration in a 19-inch rack-mount, 2U (EIA units) drawer configuration. All the cores are activated. The Power AC922 server provides two hot-swap and redundant power supplies (for all GPU configurations) and sixteen DIMM memory slots. Supported memory features are 8 GB (#EM60), 16 GB (#EM61), 32 GB (#EM63), 64 GB (#EM64), and 128 GB (#EM65), allowing for a maximum system memory of over 2 TB.

Summary of standard features for the Power AC922 server:

  • Power Systems server built with POWER9 processor modules
    • 18-core, 3.15 GHz (3.45 GHz turbo)
    • 22-core, 2.80 GHz (3.10 GHz turbo)
  • High-performance DDR4 memory
    • 128 GB (#EM60 x16), 256 GB (#EM61 x16), 512 GB (#EM63 x16), 1.02 TB (#EM64 x16), or 2.04 TB (#EM65 x16) memory features
  • Storage bays
    • Two SFF 4-bays for up to two HDDs or SSDs (No mixing of HDD and SSD drives allowed)
    • Up to two 1 TB 7200 RPM SATA disk drives (#ELD0)
    • Up to two 2 TB 7200 RPM SATA disk drives (#ES6A)
    • Up to two 960 GB SATA SSD (#ELU4)
    • Up to two 1.92 TB SATA SSD (#ELU5)
    • Up to two 3.84 TB SATA SSD (#ELU6)
  • PCIe Gen4 slots
    • Two PCIe x16 4.0 LP slots, CAPI enabled
    • One PCIe x16 4.0 LP slot, CAPI enabled or one PCIe x8x8 shared 4.0 slot, CAPI enabled
    • One PCIe x4 4.0 LP slot
  • Compute Intensive Accelerator
    • Four or six 16 GB SXM2 NVIDIA Tesla V100 GPUs with NVLink Water-Cooled (#EC4H)
    • Four or six 32 GB SXM2 NVIDIA Tesla V100 GPUs with NVLink Water-Cooled (#EC4K)
  • Integrated:
    • One front USB 3.0 port and one rear USB 3.0 port
  • Two hot-swap redundant power supplies (for all GPU configurations)
  • 19-inch rack-mount hardware (2U)

Power AC922 (8335-GTX) system configuration

The minimum Power AC922 initial order must include two processor modules, 128 GB of memory, two power supplies, two line cords, rack- mounting hardware, a system software indicator, a rack integrator specify, and a Language Group Specify.

Linux is the operating system. The minimum defined initial order configuration is as follows:

Feature       Description
number
------------- ----------------------------------------------------------
EP0Q x 2      18-core 3.15 GHz (3.45 GHz Turbo) POWER9 Processor Module
or
EP0S x 2      22-core 2.80 GHz (3.10 GHz Turbo) POWER9 Processor Module
EM60 x 16     8 GB DIMMs, 2666 MHz, 4 Gb DDR4 DRAM
EC16 x 1      Open Power Abstraction Layer (OPAL)
EC4H x 4      NVIDIA Tesla V100 GPU with NVLink Water-Cooled (16 GB)
or
EC4K x 4      NVIDIA Tesla V100 GPU with NVLink Water-Cooled (32 GB)
4650 x 1      Rack Indicator -- Not Factory Integrated
EB2X x 2      AC Power Supply, 2200 Watt (200-240V/277V)
2 power cords Select two power cords from supported list. Feature EPAM
              is defaulted. (#EPAM is Power Cord 4.3 m (14-ft), Drawer
              to Wall/IBM PDU (250V/16A))
2147 x 1      Primary OS Linux
EJTX x 1      Rack-mount fixed rail kit
93xx x 1      Language Group Specify (select one from announced
              features)

Note: If a rack is wanted, it must be ordered as an MTM rack on initial system orders. If on the same system order, it will be shipped at the same time in the same shipment, but in separate packing material. IBM does not offer IBM Manufacturing rack integration of the server into the rack before shipping at this time.

Processor modules

A minimum of two identical processor modules are required with eighteen processor cores (#EP0Q) or twenty-two processor cores (#EP0S) allowed. All processor cores are activated.

There are no processor activation features used or orderable on the 8335-GTX. All processor cores/"n-ways" are always fully activated.

Mixing of different processor features on the same system is not allowed.

System memory

There are 16 memory slots in the server. Each memory slot can hold one DDR4 memory feature. A minimum of 16 memory features is required, filling all the memory slots. With 8 GB DIMMs, 128 GB of memory is required. With 16 GB DIMMs, 256 GB of memory is required. With 32 GB DIMMs, 512 GB of memory is required. With 64 GB DIMMs, 1.02 TB is required. With 128 GB DIMMs, 2.04 TB of memory is required.
Note: Mixing of memory features is not allowed.

DIMM memory features

Feature name         Feature Maximum
                     number  quantity
-------------------- ------- --------
8 GB DDR4 2666 MHz   EM60    16
16 GB DDR4 2666 MHz  EM61    16
32 GB DDR4 2666 MHz  EM63    16
64 GB DDR4 2666 MHz  EM64    16
128 GB DDR4 2666 MHz EM65    16

I/O adapters

The system contains the following adapter slots:

  • Two PCIe x16 G4 HHHL slots, CAPI enabled
  • One PCIe x16 G4HHHL slot, CAPI enabled or one PCIe x8x8 shared 4.0 slot, CAPI enabled
  • One PCIe x4 G4 HHHL slot

Only LP adapters can be placed in LP slots. An x8 adapter can be placed in an x16 slot, but an x16 adapter cannot be placed in an x8 slot.

Power supply

  • Two hot-swap and redundant power supplies (for all GPU configurations): 2200 W 200-240V/277V AC (2 x #EB2X)

Power cords

Two power cords are required. Two feature EPAM cords are defaulted. See the feature listing for other options.

Physical specifications

  • Width: 443 mm (17.4 in.)
  • Depth: 850.9 mm (33.5 in.)
  • Height: 86 mm (3.4 in.)
  • Weight: 30 kg (66 lb)

To assure installability and serviceability in non-IBM industry- standard racks, review the installation planning information for any product-specific installation requirements.

Operating environment

For some standards and guidelines for operating environment, see the ASHRAE-A2 website.

Limitations

  • The integrated system ports are supported for modem and asynchronous terminal connections with Linux. Any other application using serial ports requires a serial port adapter to be installed in a PCI slot. The integrated system ports do not support HACMP configurations.
  • The VGA port does not support cable lengths that exceed 3 meters.

Hardware requirements

Not applicable.

Software requirements

  • Red Hat Enterprise Linux 7.5 for Power LE (POWER9), or later


Back to topBack to top

Publications

Top rule

Power Systems hardware documentation provides clients with the following topical information:

  • Licenses, notices, safety, and warranty information
  • Planning for the system
  • Installing and configuring the system
  • Troubleshooting, service, and support
  • Installing operating systems
  • Glossary

Product documentation is also available on DVD (SK5T-7087).

The following information is shipped with the 8335-GTX server:

  • Supporting information for IBM
  • IBM Power System AC922 (8335-GTX) (GI11-9916)
  • Warranty GI11-4340
  • License Agreement for Machine Code Z125-5468

You can access the product documentation at IBM Knowledge Center.

IBM Knowledge Center provides you with a single information centre where you can access product documentation for IBM systems hardware, operating systems, and server software. Through a consistent framework, you can efficiently find information and personalize your access.

To access the IBM Publications Center Portal, go to the IBM Publications Center website.

The Publications Center is a worldwide central repository for IBM product publications and marketing material with a catalogue of 70,000 items. Extensive search facilities are provided. A large number of publications are available on-line in various file formats, which can currently be downloaded.



Back to topBack to top

Features

Top rule

Features - No charge Features - No charge Feature availability matrix Feature availability matrix
Features - Chargeable Features - Chargeable Feature descriptions Feature descriptions


Features - No charge

  • NONE

Features - Chargeable

Special Features - Initial Orders
  • Adapters
    • (#EC4H) -NVIDIA Tesla V100 GPU with NVLink Water-Cooled (16 GB)
    • (#EC4K) -NVIDIA Tesla V100 GPU with NVLink Water-Cooled (32 GB)
  • Administrative
    • (#ECP0) -Cloud Private Solution
    • (#ESC5) -S&H-a
  • Languages
    • (#9300) -Language Group Specify - US English
    • (#9700) -Language Group Specify - Dutch
    • (#9703) -Language Group Specify - French
    • (#9704) -Language Group Specify - German
    • (#9705) -Language Group Specify - Polish
    • (#9706) -Language Group Specify - Norwegian
    • (#9707) -Language Group Specify - Portuguese
    • (#9708) -Language Group Specify - Spanish
    • (#9711) -Language Group Specify - Italian
    • (#9712) -Language Group Specify - Canadian French
    • (#9714) -Language Group Specify - Japanese
    • (#9715) -Language Group Specify - Traditional Chinese (Taiwan)
    • (#9716) -Language Group Specify - Korean
    • (#9718) -Language Group Specify - Turkish
    • (#9719) -Language Group Specify - Hungarian
    • (#9720) -Language Group Specify - Slovakian
    • (#9721) -Language Group Specify - Russian
    • (#9722) -Language Group Specify - Simplified Chinese (PRC)
    • (#9724) -Language Group Specify - Czech
    • (#9725) -Language Group Specify - Romanian
    • (#9726) -Language Group Specify - Croatian
    • (#9727) -Language Group Specify - Slovenian
    • (#9728) -Language Group Specify - Brazilian Portuguese
    • (#9729) -Language Group Specify - Thai
  • Miscellaneous
    • (#5000) -Software Preload Required
    • (#8143) -Linux Software Pre-install
    • (#8144) -Linux Software Pre-install (Business Partners)
    • (#9461) -Month Indicator
    • (#9462) -Day Indicator
    • (#9463) -Hour Indicator
    • (#9464) -Minute Indicator
    • (#9465) -Qty Indicator
    • (#9466) -Countable Member Indicator
    • (#EHLU) -IBM Power Systems for SAS Viya (Linux)
  • Packaging
    • (#ERB0) -Bulk Packaging Request ID
    • (#ERB1) -Bulk Packaging ID #1
    • (#ERB2) -Bulk Packaging ID #2
    • (#ERB3) -Bulk Packaging ID #3
    • (#ERB4) -Bulk Packaging ID #4
    • (#ERB5) -Bulk Packaging ID #5
    • (#ERB6) -Bulk Packaging ID #6
    • (#ERB7) -Bulk Packaging ID #7
    • (#ERB8) -Bulk Packaging ID #8
    • (#ERB9) -Bulk Packaging ID #9
    • (#ERBA) -Bulk Packaging ID #10
    • (#ERBB) -Bulk Packaging ID #11
    • (#ERBC) -Bulk Packaging ID #12
    • (#ERBD) -Bulk Packaging ID #13
    • (#ERBE) -Bulk Packaging ID #14
    • (#ERBF) -Bulk Packaging ID #15
    • (#ERBG) -Bulk Packaging ID #16
    • (#ERBH) -Bulk Packaging ID #17
    • (#ERBJ) -Bulk Packaging ID #18
    • (#ERBK) -Bulk Packaging ID #19
    • (#ERBL) -Bulk Packaging ID #20
    • (#ERBZ) -No Bulk Packaging Specify
  • Power
    • (#EB2X) -AC Power Supply - 2200 WATT (200-240V/277V)
  • Processor
    • (#EP0Q) -18-core 3.15 GHz (3.45 GHz Turbo) POWER9
    • (#EP0S) -22-core 2.80 GHz (3.10 GHz Turbo) POWER9
  • Services
    • (#ERF1) -RFID Tags for Servers, Compute Nodes, Chassis, Racks, and HMCs
  • Specify Codes
    • (#4650) -Rack Indicator-Not Factory Integrated
    • (#4651) -Rack Indicator, Rack #1
    • (#4652) -Rack Indicator, Rack #2
    • (#4653) -Rack Indicator, Rack #3
    • (#4654) -Rack Indicator, Rack #4
    • (#4655) -Rack Indicator, Rack #5
    • (#4656) -Rack Indicator, Rack #6
    • (#4657) -Rack Indicator, Rack #7
    • (#4658) -Rack Indicator, Rack #8
    • (#4659) -Rack Indicator, Rack #9
    • (#4660) -Rack Indicator, Rack #10
    • (#4661) -Rack Indicator, Rack #11
    • (#4662) -Rack Indicator, Rack #12
    • (#4663) -Rack Indicator, Rack #13
    • (#4664) -Rack Indicator, Rack #14
    • (#4665) -Rack Indicator, Rack #15
    • (#4666) -Rack Indicator, Rack #16

      One and only one rack indicator feature is required onall orders (#4650 to #4666).

    • (#9442) -New Red Hat License Core Counter
    • (#9445) -Other Linux License Core Counter
    • (#9446) -3rd Party Linux License Core Counter
    • (#9449) -Other License Core Counter
  • Virtualization Engine
    • (#EC16) -Open Power non-virtualized configuration
Special Features - Plant and/or Field Installable
  • Adapters
    • (#EC2R) -PCIe3 LP 2-Port 10Gb NIC&ROCE SR/Cu Adapter
    • (#EC2T) -PCIe3 LP 2-Port 25/10Gb NIC&ROCE SR/Cu Adapter
    • (#EC3L) -PCIe3 LP 2-port 100GbE (NIC& RoCE) QSFP28 Adapter x16
    • (#EC5A) -PCIe3 LP 1.6 TB SSD NVMe adapter
    • (#EC5C) -PCIe3 LP 3.2 TB SSD NVMe adapter
    • (#EC5E) -PCIe3 LP 6.4 TB SSD NVMe adapter
    • (#EC5G) -PCIe3 LP 1.6 TB SSD NVMe Adapter
    • (#EC62) -PCIe4 LP 1-port 100Gb EDR IB CAPI adapter
    • (#EC64) -PCIe4 LP 2-port 100Gb EDR IB CAPI adapter
    • (#EC67) -PCIe4 LP 2-port 100Gb ROCE EN LP adapter
    • (#EC6E) -PCIe4 LP 1-port HDR100 IB Adapter x16
    • (#EC6G) -PCIe4 LP 2-port HDR100 IB Adapter x16
    • (#EL3Z) -PCIe2 LP 2-port 10/1GbE BaseT RJ45 Adapter
    • (#EL43) -PCIe3 LP 2-port 16Gb Fibre Channel Adapter
    • (#EL4M) -PCIe2 LP 4-port 1GbE Adapter
    • (#EL5V) -PCIe3 LP 2-port 32Gb Fibre Channel Adapter
    • (#EN0T) -PCIe2 LP 4-Port (10Gb+1GbE) SR+RJ45 Adapter
    • (#EN0V) -PCIe2 LP 4-port (10Gb+1GbE) Copper SFP+RJ45 Adapter
  • Administrative
    • (#ESC0) -S&H - No Charge
  • Cable
    • (#4256) -Extender Cable - USB Keyboards, 1.8M
    • (#EB27) -QSFP+ 40GbE Base-SR Transceiver
    • (#EB2B) -1m (3.3-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)
    • (#EB2H) -3m (9.8-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)
    • (#EB2J) -10m (30.3-ft), IBM Passive QSFP+ MTP Optical Cable
    • (#EB2K) -30m (90.3-ft), IBM Passive QSFP+ MTP Optical Cable
    • (#EB46) -10GbE Optical Transceiver SFP+ SR
    • (#EB47) -25GbE Optical Transceiver SFP28
    • (#EB4J) -0.5m SFP28/25GbE copper Cable
    • (#EB4K) -1.0m SFP28/25GbE copper Cable
    • (#EB4L) -1.5m SFP28/25GbE copper Cable
    • (#EB4M) -2.0m SFP28/25GbE copper Cable
    • (#EB4P) -2.0m QSFP28/100GbE copper split Cable to SFP28 4x25GbE
    • (#EB50) -0.5m EDR IB Copper Cable QSFP28
    • (#EB52) -2.0M EDR IB Copper Cable QSFP28
    • (#EB54) -1.5M EDR IB Copper Cable QSFP28
    • (#EB57) -QSFP+ 40GbE Base-SR4 Transceiver
    • (#EB59) -100GbE Optical Transceiver QSFP28
    • (#EB5A) -3M EDR IB Optical Cable QSFP28
    • (#EB5B) -5M EDR IB Optical Cable QSFP28
    • (#EB5C) -10M EDR IB Optical Cable QSFP28
    • (#EB5D) -15M EDR IB Optical Cable QSFP28
    • (#EB5E) -20M EDR IB Optical Cable QSFP28
    • (#EB5F) -30M EDR IB Optical Cable QSFP28
    • (#EB5G) -50M EDR IB Optical Cable QSFP28
    • (#EB5H) -100M EDR IB Optical Cable QSFP28
    • (#EB5J) -0.5M 100GbE Copper Cable QSFP28
    • (#EB5K) -1.0M 100GbE Copper Cable QSFP28
    • (#EB5L) -1.5M 100GbE Copper Cable QSFP28
    • (#EB5M) -2.0M 100GbE Copper Cable QSFP28
    • (#EB5N) -25M EDR IB Optical Cable QSFP28
    • (#EB5R) -3M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5S) -5M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5T) -10M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5U) -15M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5V) -20M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5W) -30M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5X) -50M 100GbE Optical Cable QSFP28 (AOC)
    • (#EB5Y) -100M 100GbE Optical Cable QSFP28 (AOC)
    • (#EBB1) - 1m Passive Copper cable IB HDR QSFP56
    • (#EBB2) - 1.5m Passive Copper cable IB HDR, QSFP/QSFP
    • (#EBB3) - 2m Passive Copper cable IB HDR, QSFP/QSFP
    • (#EBBA) -3m Active Optical cable IB HDR QSFP56
    • (#EBBB) -5m Active Optical cable IB HDR, QSFPP/QSFP
    • (#EBBC) -10m Active Optical cable IB HDR, QSFP/QSFP
    • (#EBBE) -20m Active Optical cable IB HDR, QSFP/QSFP
    • (#EBBF) -30m Active Optical cable IB HDR, QSFP/QSFP
    • (#EBBK) -1m Passive Copper Splitter cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#EBBL) -1.5m Passivee Copper Split cable, IB HDR 200GB/s to 2X100GB/s, QSFP56 to 2xQSFP56
    • (#EBBM) -2m Passive Copper Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#EBBQ) -3m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#EBBR) -5m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#EBBS) -10m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#EBBU) -20m Active Optical Split cable, 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#EBBV) -30m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56
    • (#ECBN) -5m (16.4-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)
    • (#ECW0) -Optical Wrap Plug
    • (#EN01) -1m (3.3-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper
    • (#EN02) -3m (9.8-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper
    • (#EN03) -5m (16.4-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper
  • Disk
    • (#ELD0) -1 TB 7.2k RPM 5xx SATA SFF-4 Disk Drive
    • (#ES6A) -2 TB 7.2k RPM 5xx SATA SFF-4 Disk Drive
  • Linecords
    • (#EPAJ) -Power Cord 2m (6.56-foot), Drawer to IBM PDU, 250V/16A
    • (#EPAL) -Power Cord 2.8m (9.18-foot), Drawer to IBM PDU, 250V/16A
    • (#EPAM) -Power Cord 4.3m (14.10-foot), Drawer to IBM PDU, 250V/ 16A
  • Media Devices
    • (#EUA5) -Stand-alone USB DVD drive w/cable
  • Memory
    • (#EM60) -8 GB DDR4 Memory
    • (#EM61) -16 GB DDR4 Memory
    • (#EM63) -32 GB DDR4 Memory
    • (#EM7B) -64 GB DDR4 Memory Dimm
    • (#EM7C) -128 GB DDR4 Memory Dimm
  • Miscellaneous
    • (#2147) -Primary OS - Linux
    • (#ECSF) -Custom Service Specify, Montpellier, France
    • (#EHL9) -Marketing Specify - NVIDIA GPU
    • (#EJ31) -Four Water-Cooled Planar Assembly
    • (#EJ34) -Six Water-Cooled Planar Assembly
  • Processor
    • (#EP72) -18-core 3.15 GHz (3.45 GHz Turbo) POWER9 Processor
    • (#EP74) -22-core 2.80 GHz (3.10 GHz Turbo) POWER9 Processor
  • Rack Related
    • (#EJTX) -Rack-mount Fixed Rail Kit
  • Services
    • (#0010) -One CSC Billing Unit
    • (#0011) -Ten CSC Billing Units
  • Solid State Drive
    • (#ELU4) -960 GB 2.5-inch SATA/SSD Disk Drive
    • (#ELU5) -1.92 TB 2.5-inch SATA/SSD Disk Drive
    • (#ELU6) -3.84 TB 2.5-inch SATA/SSD Disk Drive

Feature availability matrix

The following feature availability matrix for MT 8335 uses the letter "A" to indicate features that are available and orderable on the specified models. "S" indicates a feature that is supported on the new model during a model conversion; these features will work on the new model, but additional quantities of these features cannot be ordered on the new model; they can only be removed. "N" indicates that the feature is not supported on the new model and must be removed during the model conversion. As additional features are announced, supported, or withdrawn, this list will be updated. Please check with your Marketing Representative for additional information.

        |G| A = AVAILABLE  S = SUPPORTED
        |T| N = NOT SUPPORTED, MUST BE REMOVED
        |X|
        | |
FEAT/PN | |        DESCRIPTION
--------|-|----------------------------------------------------------
0010    |A| One CSC Billing Unit
0011    |A| Ten CSC Billing Units

2147    |A| Primary OS - Linux
4256    |A| Extender Cable - USB Keyboards, 1.8M
One and only one rack indicator feature is required on
all orders (#4650 to #4666).
4650    |A| Rack Indicator- Not Factory Integrated
4651    |A| Rack Indicator, Rack #1
4652    |A| Rack Indicator, Rack #2
4653    |A| Rack Indicator, Rack #3
4654    |A| Rack Indicator, Rack #4
4655    |A| Rack Indicator, Rack #5
4656    |A| Rack Indicator, Rack #6
4657    |A| Rack Indicator, Rack #7
4658    |A| Rack Indicator, Rack #8
4659    |A| Rack Indicator, Rack #9
4660    |A| Rack Indicator, Rack #10
4661    |A| Rack Indicator, Rack #11
4662    |A| Rack Indicator, Rack #12
4663    |A| Rack Indicator, Rack #13
4664    |A| Rack Indicator, Rack #14
4665    |A| Rack Indicator, Rack #15
4666    |A| Rack Indicator, Rack #16
5000    |A| Software Preload Required
8143    |A| Linux Software Pre-install
8144    |A| Linux Software Pre-install (Business Partners)
9300    |A| Language Group Specify - US English
9442    |A| New Red Hat License Core Counter
9445    |A| Other Linux License Core Counter
9446    |A| 3rd Party Linux License Core Counter
9449    |A| Other License Core Counter
9461    |A| Month Indicator
9462    |A| Day Indicator
9463    |A| Hour Indicator
9464    |A| Minute Indicator
9465    |A| Qty Indicator
9466    |A| Countable Member Indicator
9700    |A| Language Group Specify - Dutch
9703    |A| Language Group Specify - French
9704    |A| Language Group Specify - German
9705    |A| Language Group Specify - Polish
9706    |A| Language Group Specify - Norwegian
9707    |A| Language Group Specify - Portuguese
9708    |A| Language Group Specify - Spanish
9711    |A| Language Group Specify - Italian
9712    |A| Language Group Specify - Canadian French
9714    |A| Language Group Specify - Japanese
9715    |A| Language Group Specify - Traditional Chinese (Taiwan)
9716    |A| Language Group Specify - Korean
9718    |A| Language Group Specify - Turkish
9719    |A| Language Group Specify - Hungarian
9720    |A| Language Group Specify - Slovakian
9721    |A| Language Group Specify - Russian
9722    |A| Language Group Specify - Simplified Chinese (PRC)
9724    |A| Language Group Specify - Czech
9725    |A| Language Group Specify - Romanian
9726    |A| Language Group Specify - Croatian
9727    |A| Language Group Specify - Slovenian
9728    |A| Language Group Specify - Brazilian Portuguese
9729    |A| Language Group Specify - Thai
EB27    |S| QSFP+ 40GbE Base-SR Transceiver
EB2B    |S| 1m (3.3-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)
EB2H    |S| 3m (9.8-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)
EB2J    |A| 10m (30.3-ft), IBM Passive QSFP+ MTP Optical Cable
EB2K    |A| 30m (90.3-ft), IBM Passive QSFP+ MTP Optical Cable
EB2X    |A| AC Power Supply - 2200 WATT (200-240V/277V)
EB46    |A| 10GbE Optical Transceiver SFP+ SR
EB47    |A| 25GbE Optical Transceiver SFP28
EB4J    |A| 0.5m SFP28/25GbE copper Cable
EB4K    |A| 1.0m SFP28/25GbE copper Cable
EB4L    |S| 1.5m SFP28/25GbE copper Cable
EB4M    |A| 2.0m SFP28/25GbE copper Cable
EB4P    |A| 2.0m QSFP28/100GbE copper split Cable to SFP28 4x25GbE
EB50    |S| 0.5m EDR IB Copper Cable QSFP28
EB51    |A| 1.0m EDR IB Copper Cable QSFP28
EB52    |A| 2.0M EDR IB Copper Cable QSFP28
EB54    |A| 1.5M EDR IB Copper Cable QSFP28
EB57    |A| QSFP+ 40GbE Base-SR4 Transceiver
EB59    |A| 100GbE Optical Transceiver QSFP28
EB5A    |A| 3M EDR IB Optical Cable QSFP28
EB5B    |A| 5M EDR IB Optical Cable QSFP28
EB5C    |A| 10M EDR IB Optical Cable QSFP28
EB5D    |A| 15M EDR IB Optical Cable QSFP28
EB5E    |A| 20M EDR IB Optical Cable QSFP28
EB5F    |A| 30M EDR IB Optical Cable QSFP28
EB5G    |A| 50M EDR IB Optical Cable QSFP28
EB5H    |A| 100M EDR IB Optical Cable QSFP28
EB5J    |S| 0.5M 100GbE Copper Cable QSFP28
EB5K    |A| 1.0M 100GbE Copper Cable QSFP28
EB5L    |A| 1.5M 100GbE Copper Cable QSFP28
EB5M    |A| 2.0M 100GbE Copper Cable QSFP28
EB5N    |S| 25M EDR IB Optical Cable QSFP28
EB5R    |A| 3M 100GbE Optical Cable QSFP28 (AOC)
EB5S    |A| 5M 100GbE Optical Cable QSFP28 (AOC)
EB5T    |A| 10M 100GbE Optical Cable QSFP28 (AOC)
EB5U    |A| 15M 100GbE Optical Cable QSFP28 (AOC)
EB5V    |A| 20M 100GbE Optical Cable QSFP28 (AOC)
EB5W    |A| 30M 100GbE Optical Cable QSFP28 (AOC)
EB5X    |A| 50M 100GbE Optical Cable QSFP28 (AOC)
EB5Y    |S| 100M 100GbE Optical Cable QSFP28 (AOC)
EBB1    |S| 1m Passive Copper cable IB HDR QSFP56
EBB2    |S| 1.5m Passive Copper cable IB HDR, QSFP/QSFP
EBB3    |S| 2m Passive Copper cable IB HDR, QSFP/QSFP
EBBA    |S| 3m Active Optical cable IB HDR QSFP56
EBBB    |S| 5m Active Optical cable IB HDR, QSFPP/QSFP
EBBC    |S| 10m Active Optical cable IB HDR, QSFP/QSFP
EBBE    |S| 20m Active Optical cable IB HDR, QSFP/QSFP
EBBF    |S| 30m Active Optical cable IB HDR, QSFP/QSFP
EBBK    |S| 1m Passive Copper Splitter cable, IB HDR 200Gb/s to
        | |2x100Gb/s, QSFP56 to 2xQSFP56
EBBL    |S| 1.5m Passivee Copper Split cable, IB HDR 200GB/s to
        | |2X100GB/s, QSFP56 to 2xQSFP56
EBBM    |S| 2m Passive Copper Split cable, IB HDR 200Gb/s to 2x100Gb/
        | |s, QSFP56 to 2xQSFP56
EBBQ    |S| 3m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/
        | |s, QSFP56 to 2xQSFP56
EBBR    |S| 5m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/
        | |s, QSFP56 to 2xQSFP56
EBBS    |S| 10m Active Optical Split cable, IB HDR 200Gb/s to
        | |2x100Gb/s, QSFP56 to 2xQSFP56
EBBU    |S| 20m Active Optical Split cable, 200Gb/s to 2x100Gb/s,
        | |QSFP56 to 2xQSFP56
EBBV    |S| 30m Active Optical Split cable, IB HDR 200Gb/s to
        | |2x100Gb/s, QSFP56 to 2xQSFP56
EC16    |A| Open Power non-virtualized configuration
EC2R    |S| PCIe3 LP 2-Port 10Gb NIC&ROCE SR/Cu Adapter
EC2T    |A| PCIe3 LP 2-Port 25/10Gb NIC&ROCE SR/Cu Adapter
EC3L    |S| PCIe3 LP 2-port 100GbE (NIC& RoCE) QSFP28 Adapter x16
EC4H    |S| NVIDIA Tesla V100 GPU with NVLink Water-Cooled (16 GB)
EC4K    |A| NVIDIA Tesla V100 GPU with NVLink Water-Cooled (32 GB)
EC5A    |S| PCIe3 LP 1.6 TB SSD NVMe adapter
EC5C    |A| PCIe3 LP 3.2 TB SSD NVMe adapter
EC5E    |A| PCIe3 LP 6.4 TB SSD NVMe adapter
EC5G    |A| PCIe3 LP 1.6 TB SSD NVMe adapter
EC62    |A| PCIe4 LP 1-port 100Gb EDR IB CAPI adapter
EC64    |A| PCIe4 LP 2-port 100Gb EDR IB CAPI adapter
EC67    |A| PCIe4 LP 2-port 100Gb ROCE EN LP adapter
EC6E    |S| PCIe4 LP 1-port HDR100 IB Adapter x16
EC6G    |S| PCIe4 LP 2-port HDR100 IB Adapter x16
ECBN    |S| 5m (16.4-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)
ECP0    |S| Cloud Private Solution
ECSF    |S| Custom Service Specify, Montpellier, France
ECW0    |A| Optical Wrap Plug
EHL9    |A| Marketing Specify - NVIDIA GPU
EHLU    |S| IBM Power Systems for SAS Viya (Linux)
EJ31    |A| Four Water-Cooled Planar Assembly
EJ34    |A| Six Water-Cooled Planar Assembly
EJTX    |A| Rack-mount Fixed Rail Kit
EL3Z    |A| PCIe2 LP 2-port 10/1GbE BaseT RJ45 Adapter
EL43    |S| PCIe3 LP 2-port 16Gb Fibre Channel Adapter
EL4M    |S| PCIe2 LP 4-port 1GbE Adapter
EL5V    |A| PCIe3 LP 2-port 32Gb Fibre Channel Adapter
ELD0    |A| 1 TB 7.2k RPM 5xx SATA SFF-4 Disk Drive
ELU4    |A| 960 GB 2.5-inch SATA/SSD Disk Drive
ELU5    |A| 1.92 TB 2.5-inch SATA/SSD Disk Drive
ELU6    |A| 3.84 TB 2.5-inch SATA/SSD Disk Drive
EM60    |S| 8 GB DDR4 Memory
EM61    |A| 16 GB DDR4 Memory
EM63    |A| 32 GB DDR4 Memory
EM64    |S| 64 GB DDR4 Memory
EM65    |S| 128 GB DDR4 Memory
EM7B    |A| 64 GB DDR4 Memory Dimm
EM7C    |A| 128 GB DDR4 Memory Dimm
EN01    |A| 1m (3.3-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper
EN02    |A| 3m (9.8-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper
EN03    |A| 5m (16.4-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper
EN0T    |A| PCIe2 LP 4-Port (10Gb+1GbE) SR+RJ45 Adapter
EN0V    |A| PCIe2 LP 4-port (10Gb+1GbE) Copper SFP+RJ45 Adapter
EP0Q    |A| 18-core 3.15 GHz (3.45 GHz Turbo) POWER9
EP0S    |A| 22-core 2.80 GHz (3.10 GHz Turbo) POWER9
EP72    |A| 18-core 3.15 GHz (3.45 GHz Turbo) POWER9 Processor
EP74    |A| 22-core 2.80 GHz (3.10 GHz Turbo) POWER9 Processor
EPAJ    |A| Power Cord 2m (6.56-foot), Drawer to IBM PDU, 250V/16A
EPAL    |A| Power Cord 2.8m (9.18-foot), Drawer to IBM PDU, 250V/16A
EPAM    |A| Power Cord 4.3m (14.10-foot), Drawer to IBM PDU, 250V/16A
ERB0    |A| Bulk Packaging Request ID
ERB1    |A| Bulk Packaging ID #1
ERB2    |A| Bulk Packaging ID #2
ERB3    |A| Bulk Packaging ID #3
ERB4    |A| Bulk Packaging ID #4
ERB5    |A| Bulk Packaging ID #5
ERB6    |A| Bulk Packaging ID #6
ERB7    |A| Bulk Packaging ID #7
ERB8    |A| Bulk Packaging ID #8
ERB9    |A| Bulk Packaging ID #9
ERBA    |A| Bulk Packaging ID #10
ERBB    |A| Bulk Packaging ID #11
ERBC    |A| Bulk Packaging ID #12
ERBD    |A| Bulk Packaging ID #13
ERBE    |A| Bulk Packaging ID #14
ERBF    |A| Bulk Packaging ID #15
ERBG    |A| Bulk Packaging ID #16
ERBH    |A| Bulk Packaging ID #17
ERBJ    |A| Bulk Packaging ID #18
ERBK    |A| Bulk Packaging ID #19
ERBL    |A| Bulk Packaging ID #20
ERBZ    |A| No Bulk Packaging Specify
ERF1    |S| RFID Tags for Servers, Compute Nodes, Chassis, Racks, and
        | |HMCs
ES6A    |A| 2 TB 7.2k RPM 5xx SATA SFF-4 Disk Drive
ESC0    |A| S&H - No Charge
ESC5    |A| S&H-a
EUA5    |A| Stand-alone USB DVD drive w/cable

Feature descriptions


Note: Not all of the following features are available in all countries. Check with your country representative for specific feature availability.

The following is a list of all feature codes in numeric order for the IBM Power Systems 8335 machine type.

Attributes, as defined in the following feature descriptions, state the interaction of requirements among features.

Minimums and maximums are the absolute limits for a single feature without regard to interaction with other features. The maximum valid quantity for MES orders may be different than for initial orders. The maximums listed below refer to the largest quantity of these two possibilities.

The order type defines if a feature is orderable only on initial orders, only on MES orders, on both initial and MES orders, or if a feature is supported on a model due to a model conversion. Supported features cannot be ordered on the converted model, only left on or removed from the converted model.

(#0010) - One CSC Billing Unit

One Billing Unit used by the Customer Solution Center.

  • Attributes provided: One CSC Billing Unit
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: Does not apply
(#0011) - Ten CSC Billing Units

Ten Billing Units used by the Customer Solutions Center.

  • Attributes provided: Ten CSC Billing Units
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 100 (Initial order maximum: 100)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: Does not apply
(#2147) - Primary OS - Linux

Indicates clients intend to use the Linux operating system on the primary system partition. This feature is used as a Manufacturing Routing indicator and does not deliver parts, software or services.

  • Attributes provided: None
  • Attributes required: Indicates clients intend to use the Linux operating system on the primary system partition.
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#4256) - Extender Cable - USB Keyboards, 1.8M

This feature provides a 1.8M extension cable for use with USB keyboards.

  • Attributes provided: 1.8M Extension Cable
  • Attributes required: USB Keyboard
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No

One and only one rack indicator feature is required onall orders (#4650 to #4666).

(#4650) - Rack Indicator- Not Factory Integrated

This indicator is used to specify that the rack mountable device in this initial order should not be merged into a rack within IBM Manufacturing. If a device with 4650 is ordered with a rack, the device will not be factory integrated in the ordered rack and will ship uninstalled in the rack.
Note: This "no additional charge" feature will be placed on an initial order for a rack mountable device by the Configuration Tool when the order does not ship from IBM Manufacturing in a Rack.

A rack integration indicator is required on all 19" Rack mountable device initial orders. One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed.

  • Attributes provided: System will not be shipped in a rack.
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4651) - Rack Indicator, Rack #1

When added to an initial rack order, this indicator is used to specify the first rack for a multi rack order, or the only rack for a single rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #1.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack Integration/ Rack Specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4652) - Rack Indicator, Rack #2

When added to an initial rack order, this indicator is used to specify the second rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #2 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack Integration/Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4653) - Rack Indicator, Rack #3

When added to an initial rack order, this indicator is used to specify the third rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #3 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4654) - Rack Indicator, Rack #4

When added to an initial rack order, this indicator is used to specify the fourth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #4 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4655) - Rack Indicator, Rack #5

When added to an initial rack order, this indicator is used to specify the fifth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #5 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4656) - Rack Indicator, Rack #6

When added to an initial rack order, this indicator is used to specify the sixth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #6 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4657) - Rack Indicator, Rack #7

When added to an initial rack order, this indicator is used to specify the seventh rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #7 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4658) - Rack Indicator, Rack #8

When added to an initial rack order, this indicator is used to specify the eighth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #8 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4659) - Rack Indicator, Rack #9

When added to an initial rack order, this indicator is used to specify the ninth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #9 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4660) - Rack Indicator, Rack #10

When added to an initial rack order, this indicator is used to specify the tenth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #10 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4661) - Rack Indicator, Rack #11

When added to an initial rack order, this indicator is used to specify the eleventh rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #11 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4662) - Rack Indicator, Rack #12

When added to an initial rack order, this indicator is used to specify the twelfth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #12 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4663) - Rack Indicator, Rack #13

When added to an initial rack order, this indicator is used to specify the thirteenth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #13 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4664) - Rack Indicator, Rack #14

When added to an initial rack order, this indicator is used to specify the fourteenth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #14 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4665) - Rack Indicator, Rack #15

When added to an initial rack order, this indicator is used to specify the fifteenth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #15 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#4666) - Rack Indicator, Rack #16

When added to an initial rack order, this indicator is used to specify the sixteenth rack for a multi rack order.

When added to an initial rack mountable device order, this indicator is used to specify that the rack mountable device (such as a system or I/O drawer) is to be mounted in rack #16 of a multi rack order.
Note: For 19" rack mountable device orders: One feature code from the group 4650 to 4666 must be listed on the order. More than one feature code from this group is not allowed. For 19" rack orders: If IBM Mfg. is to assemble a rack mountable device into the rack, one feature code selection from the group 4651 to 4666 must be listed on the order. More than one feature code selection from this group is not allowed. The quantity of this selected feature code on the 19" rack order must equal the number of rack mountable devices to be installed in the rack by IBM Mfg.

  • Attributes provided: Rack specify
  • Attributes required: Rack
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#5000) - Software Preload Required

Indicates that preloaded software and/or consolidated I/O is shipped with the initial order. A maximum of one (#5000) is supported. This feature has country-specific usage.
Note: Not supported in Brazil.

  • Attributes provided: Software Pre-load
  • Attributes required: N/A
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#8143) - Linux Software Pre-install

This feature indicates that the Linux operating system is to be pre-installed on the system. Requires feature number 5000.

  • Attributes provided: Linux pre-install
  • Attributes required: Feature number 5000.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
    • Red Hat Enterprise Linux
    • Refer to Software Requirements for specific O/S levels supported
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#8144) - Linux Software Pre-install (Business Partners)

This feature indicates that the Linux operating system is to be pre-installed on the system. Requires feature number 5000 or 7305. This feature is only available to IBM Business Partner - Solution Providers and IBM Business Partner - Systems Integrators.

  • Attributes provided: Linux pre-install
  • Attributes required: Feature number 5000 or 7305.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
    • Red Hat Enterprise Linux
    • Refer to Software Requirements for specific O/S levels supported
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9300) - Language Group Specify - US English

English language group for nomenclature and standard publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9442) - New Red Hat License Core Counter

This feature is used to count the number of cores licensed to run Red Hat Linux.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 44 (Initial order maximum: 44)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9445) - Other Linux License Core Counter

This feature is used to count the number of existing Linux licenses transferred from another server.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 44 (Initial order maximum: 44)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9446) - 3rd Party Linux License Core Counter

This feature is used to count the number of cores licensed to run 3rd party Linux.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 44 (Initial order maximum: 44)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9449) - Other License Core Counter

This feature is used to count the number of other cores licensed.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 44 (Initial order maximum: 44)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9461) - Month Indicator

This month indicator is used to create a date stamp to enable CFR splitting and rejoining in order to circumvent the WTAAS maximum limitation of 30 systems entered on any one order. The quantity ordered for this feature is generated by eConfig.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 12 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9462) - Day Indicator

This day indicator is used to create a date stamp to enable CFR splitting and rejoining in order to circumvent the WTAAS maximum limitation of 30 systems entered on any one order. The quantity ordered for this feature is generated by eConfig.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 31 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9463) - Hour Indicator

This hour indicator is used to create a time stamp to enable CFR splitting and rejoining in order to circumvent the WTAAS maximum limitation of 30 systems entered on any one order. The quantity ordered for this feature is generated by eConfig.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 24 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9464) - Minute Indicator

This hour indicator is used to create a time stamp to enable CFR splitting and rejoining in order to circumvent the WTAAS maximum limitation of 30 systems entered on any one order. The quantity ordered for this feature is generated by eConfig.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 60 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9465) - Qty Indicator

This quantity indicator is used to specify the remaining, or N-1 quantity of CFR entities that need to be accumulated for rejoining. The quantity ordered for this feature is generated by eConfig. and is equal to N-1, where 'N' equals the total quantity of CFRs being rejoined.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9466) - Countable Member Indicator

This administrative indicator used to identify each CFR associated with a date/time stamp that is eligible for splitting and rejoining. The quantity ordered for this feature is generated by eConfig.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9700) - Language Group Specify - Dutch

Dutch language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9703) - Language Group Specify - French

French language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9704) - Language Group Specify - German

German language group for Nomenclature and Standard Publications.

  • Attributes provided: Language specify
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9705) - Language Group Specify - Polish

Polish language group for Nomenclature and Standard Publications.

  • Attributes provided: Language specify
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9706) - Language Group Specify - Norwegian

Norwegian language group for Nomenclature and Standard Publications.

  • Attributes provided: Language specify
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9707) - Language Group Specify - Portuguese

Portuguese language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9708) - Language Group Specify - Spanish

Spanish language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9711) - Language Group Specify - Italian

Italian language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9712) - Language Group Specify - Canadian French

Canadian French language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9714) - Language Group Specify - Japanese

Japanese language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9715) - Language Group Specify - Traditional Chinese (Taiwan)

Traditional Chinese language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9716) - Language Group Specify - Korean

Korean language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9718) - Language Group Specify - Turkish

Turkish language group for nomenclature and publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9719) - Language Group Specify - Hungarian

Hungarian language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9720) - Language Group Specify - Slovakian

Slovakian language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9721) - Language Group Specify - Russian

Russian language group for nomenclature and standard publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9722) - Language Group Specify - Simplified Chinese (PRC)

Simplified Chinese language group for nomenclature and standard publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9724) - Language Group Specify - Czech

Czech language group for nomenclature and standard publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9725) - Language Group Specify - Romanian

Romanian language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9726) - Language Group Specify - Croatian

Croatian language group for Nomenclature and Standard Publications.

  • Attributes provided: Language specify
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9727) - Language Group Specify - Slovenian

Slovenian language group for Nomenclature and Standard Publications.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9728) - Language Group Specify - Brazilian Portuguese

Brazilian Portuguese language group for Nomenclature and Standard Publications.

  • Attributes provided: Language specify
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#9729) - Language Group Specify - Thai

Thai language group for Nomenclature and Standard Publications.

  • Attributes provided: Language specify
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#EB27) - QSFP+ 40GbE Base-SR Transceiver

(No longer available as of September 30, 2021)

IBM QSFP+ optical transceiver required for 40 Gbs ports which are not using copper QSFP+ transceiver.

  • Attributes provided: QSFP+ transceiver for 40 Gbs ports.
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB2B) - 1m (3.3-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)

(No longer available as of March 5, 2021)

QSFP+ cable is used for 40Gb-to-40Gb Ethernet connectivity. Clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This copper cable has QSFP+ transceivers already attached to each end.

  • Attributes provided: 1m QSFP+ to QSFP+ Cable
  • Attributes required: QSFP/QSFP+ ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB2H) - 3m (9.8-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)

(No longer available as of March 5, 2021)

QSFP+ cable is used for 40Gb-to-40Gb Ethernet connectivity. Clients can use this QSFP+ Direct Attach Cable for Ethenet connectivity. This copper cable has QSFP+ transceivers already attached to each end.
Note: Do not use this cable between switches.

  • Attributes provided: 3m QSFP+ to QSFP+ Cable
  • Attributes required: QSFP/QSFP+ ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB2J) - 10m (30.3-ft), IBM Passive QSFP+ MTP Optical Cable

QSFP+ cable is used for 40Gb-to-40Gb Ethernet connectivity. Clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity.

  • Attributes provided: 10m QSFP+ to QSFP+ Cable
  • Attributes required: QSFP/QSFP+ ports with optical transceivers
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB2K) - 30m (90.3-ft), IBM Passive QSFP+ MTP Optical Cable

QSFP+ cable is used for 40Gb-to-40Gb Ethernet connectivity. Clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity.

  • Attributes provided: 30m QSFP+ to QSFP+ Cable
  • Attributes required: QSFP/QSFP+ ports with optical transceivers
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB2X) - AC Power Supply - 2200 WATT (200-240V/277V)

One 200-240V/277V, 2200 watt AC power supply.

  • Attributes provided:

    AC Power Supply

  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: No
(#EB46) - 10GbE Optical Transceiver SFP+ SR

One optical transceiver for 10Gb Ethernet adapter such as #EC2S or #EC2R and also #EC2U or #EC2T using SFP+ SP. Does not include cable.

The 10 Gb optical transceiver is capable up to 300 M through the OM3 cable or 82 M through OM2 cable. Either one or both of the adapter's two SFP+ ports can be populated.

  • Attributes provided: Optical Transceiver SFP+ SR 10Gb
  • Attributes required: SFP+ socket
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB47) - 25GbE Optical Transceiver SFP28

One optical transceiver for 25Gb Ethernet adapter such as #EC2U or #EC2T using SFP28. Does not include cable.

The 25 Gb optical transceiver is capable up to 100 m through the OM4 cable or 70 M through OM3 cable. Either one or both of the adapters two SFP28 ports can be populated.
Note: The SFP28 25GbE transceiver only supports 25GbE speeds.

  • Attributes provided: SFP28 optical transceiver
  • Attributes required: SFP28 socket
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB4J) - 0.5m SFP28/25GbE copper Cable

Feature EB4J is a passive 0,5 metre copper cable that contains a single high-speed copper pair, operating at data rates of up to 25 Gb/s. This cable can be used for either 25Gb Ethernet adapters or switches. Built onto each end of the cable is a passive SFP28 copper cable transceiver. Cables are available in various lengths: 0.5M - #EB4J, 1.0M=#EB4K, 1.5M=#EB4L, 2.0M=#EB4M

  • Attributes provided: Copper cable with SFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB4K) - 1.0m SFP28/25GbE copper Cable

Feature EB4K is a passive 1.0 metre copper cable that contains a single high-speed copper pair, operating at data rates of up to 25 Gb/s. This cable can be used for either 25Gb Ethernet adapters or switches. Built onto each end of the cable is a passive SFP28 copper cable transceiver. Cables are available in various lengths: 0.5M - #EB4J, 1.0M=#EB4K, 1.5M=#EB4L, 2.0M=#EB4M

  • Attributes provided: Copper cable with SFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB4L) - 1.5m SFP28/25GbE copper Cable

(No longer available as of December 31, 2020)

Feature EB4L is a passive 1.5 metre copper cable that contains a single high-speed copper pair, operating at data rates of up to 25 Gb/s. This cable can be used for either 25Gb Ethernet adapters or switches. Built onto each end of the cable is a passive SFP28 copper cable transceiver. Cables are available in various lengths: 0.5M - #EB4J, 1.0M=#EB4K, 1.5M=#EB4L, 2.0M=#EB4M

  • Attributes provided: Copper cable with SFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB4M) - 2.0m SFP28/25GbE copper Cable

Feature EB4M is a passive 2.0 metre copper cable that contains a single high-speed copper pair, operating at data rates of up to 25 Gb/s. This cable can be used for either 25Gb Ethernet adapters or switches. Built onto each end of the cable is a passive SFP28 copper cable transceiver. Cables are available in various lengths: 0.5M - #EB4J, 1.0M=#EB4K, 1.5M=#EB4L, 2.0M=#EB4M

  • Attributes provided: Copper cable with SFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB4P) - 2.0m QSFP28/100GbE copper split Cable to SFP28 4x25GbE

Feature EB4P is a 2.0 metre, active optical 100Gb E to 4x25Gb E splitter cable. It provides connectivity between system units with a QSFP28 port on one side and up to four different SFP28 ports on the other side, such as a switch and four servers.

  • Attributes provided: Copper splitter cable with QSFP28 and 4x SFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB50) - 0.5m EDR IB Copper Cable QSFP28

(No longer available as of January 18, 2019)

0.5 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a passive Quad (4-channel) Small Form-Factor Pluggable (QSFP28) copper cable transceiver. Cable can also be used for FDR IB (56Gb). Cables are available in various lengths: 0.5m = #EB50, 1m = #EB51, 2m = #EB52, 1.5m = #EB54 See also optical fibre cables for longer lengths such as #EB5A through #EB5H.

  • Attributes provided: Copper twinax cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB51) - 1.0m EDR IB Copper Cable QSFP28

1.0 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a passive Quad (4-channel) Small Form-Factor Pluggable (QSFP28) copper cable transceiver. Cable can also be used for FDR IB (56Gb). Cables are available in various lengths: 0.5m = #EB50, 1m = #EB51, 2m = #EB52, 1.5m = #EB54 See also optical fibre cables for longer lengths such as #EB5A through #EB5H.

  • Attributes provided: Copper twinax cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB52) - 2.0M EDR IB Copper Cable QSFP28

2.0 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a passive Quad (4-channel) Small Form-Factor Pluggable (QSFP28) copper cable transceiver. Cable can also be used for FDR IB (56Gb). Cables are available in various lengths: 0.5M = #EB50, 1M = #EB51, 2M = #EB52, 1.5M = #EB54 See also optical fibre cables for longer lengths such as #EB5A through #EB5H.

  • Attributes provided: Copper twinax cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB54) - 1.5M EDR IB Copper Cable QSFP28

1.5 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a passive Quad (4-channel) Small Form-Factor Pluggable (QSFP28) copper cable transceiver. Cable can also be used for FDR IB (56Gb). Cables are available in various lengths: 0.5M = #EB50, 1M = #EB51, 2M = #EB52, 1.5M = #EB54 See also optical fibre cables for longer lengths such as #EB5A through #EB5H.

  • Attributes provided: Copper twinax cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB57) - QSFP+ 40GbE Base-SR4 Transceiver

IBM QSFP+ optical base-SR4 transceiver required for 40 Gbps ports which are not using copper QSFP+ transceiver.

The QSFP+ 40GbE Base-SR4 Transceiver is designed to deliver the following functions:

  • Hot-swappable input/output device that plugs into a 40 Gigabit Ethernet QSFP port
  • Interoperable with other IEEE-compliant 40GBASE interfaces where applicable
  • High-speed electrical interface compliant to the IEEE 802.3ba standard
  • QSFP Form factor, 2-wire I2C communication interface and other low-speed electrical interface compliant to SFF 8436 and QSFP Multisource Agreement (MSA)
  • Attributes provided: QSFP+ transceiver for 40 Gbs ports
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB59) - 100GbE Optical Transceiver QSFP28

One optical transceiver for 100Gb Ethernet adapter such as #EC3L or #EC3M or #EC66 or #EC67 using QSFP28. Does not include cable.

See also AOC fibre cables which include QSFP28 transceivers EB5R - EB5Y.

  • Attributes provided: Optical Transceiver QSFP28 100Gb.
  • Attributes required: Port on adapter with QSFP28 socket.
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5A) - 3M EDR IB Optical Cable QSFP28

3 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5B) - 5M EDR IB Optical Cable QSFP28

5 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5C) - 10M EDR IB Optical Cable QSFP28

10 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5D) - 15M EDR IB Optical Cable QSFP28

15 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5E) - 20M EDR IB Optical Cable QSFP28

20 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5F) - 30M EDR IB Optical Cable QSFP28

30 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5G) - 50M EDR IB Optical Cable QSFP28

50 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5H) - 100M EDR IB Optical Cable QSFP28

100 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver. Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54. Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5J) - 0.5M 100GbE Copper Cable QSFP28

(No longer available as of March 5, 2021)

0.5 metre length passive copper cable with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches. Copper cable is also called "copper twinax" or "DAC" (Direct Attach Copper).

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EB5M (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Copper cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5K) - 1.0M 100GbE Copper Cable QSFP28

1.0 metre length passive copper cable with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches. Copper cable is also called "copper twinax" or "DAC" (Direct Attach Copper).

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EB5M (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Copper cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5L) - 1.5M 100GbE Copper Cable QSFP28

1.5 metre length passive copper cable with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches. Copper cable is also called "copper twinax" or "DAC" (Direct Attach Copper).

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EB5M (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Copper cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5M) - 2.0M 100GbE Copper Cable QSFP28

2.0 metre length passive copper cable with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches. Copper cable is also called "copper twinax" or "DAC" (Direct Attach Copper).

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EB5M (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Copper cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5N) - 25M EDR IB Optical Cable QSFP28

(No longer available as of December 31, 2020)

25 metre length optical fibre cable, also called an AOC (Active Optical Cable). The cable can be used for either 100Gb EDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a active Quad (4-channel) Small Form-Factor Pluggable (QSFP28) optical cable transceiver.

Cables are available in various lengths: 3m = #EB5A, 5M = #EB5B, 10M = #EB5C, 15M = #EB5D, 20M = #EB5E, 25M = #EB5N, 30M = #EB5F, and 50M = #EB5G, 100M = #EB5H. See also copper twinax cables for shorter lengths such as #EB50 through #EB54.

Limitation: Adapter and switch must have been manufactured by Mellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5R) - 3M 100GbE Optical Cable QSFP28 (AOC)

3 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5S) - 5M 100GbE Optical Cable QSFP28 (AOC)

5 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5T) - 10M 100GbE Optical Cable QSFP28 (AOC)

10 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5U) - 15M 100GbE Optical Cable QSFP28 (AOC)

15 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5V) - 20M 100GbE Optical Cable QSFP28 (AOC)

20 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5W) - 30M 100GbE Optical Cable QSFP28 (AOC)

30 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5X) - 50M 100GbE Optical Cable QSFP28 (AOC)

50 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EB5Y) - 100M 100GbE Optical Cable QSFP28 (AOC)

(No longer available as of March 5, 2021)

100 metre length Active Optical fibre Cable (AOC) with QSFP28 transceivers on each end. The cable used for 100Gb Ethernet adapters or switches.

Cables are available in various lengths. See shorter passive copper cables #EB5J - #EJM (0.5M - 2.0M) or see active optical fibre cables #EJ5R - #EJ5Y (3M - 100M).

  • Attributes provided: Optical fibre cable with QSFP28 transceivers
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBB1) - 1m Passive Copper cable IB HDR QSFP56

(No longer available as of December 31, 2020)

1 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. This passive copper cable can provide up to 200Gb/s connectivity between devices using QSFP56 ports.Cables are available in various lengths: 1M - #EBB1, 1.5M=#EBB2, 2M=#EBB3.

  • Attributes provided: 1m QSFP56 Cable
  • Attributes required: QSFP56 ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBB2) - 1.5m Passive Copper cable IB HDR, QSFP/QSFP

(No longer available as of December 31, 2020)

1.5 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. The cable can be used for either 100Gb HDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a passive Quad (4-channel) Small Form-Factor Pluggable (QSFP28) copper cable transceiver. Built onto each end of the cable is a passive QSFP copper cable transceiver. Cables are available in various lengths: 1M - #EBB1, 1.5M=#EBB2, 2M=#EBB3.

  • Attributes provided: 1.5m QSFP/QSFP Cable
  • Attributes required: QSFP/QSFP ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBB3) - 2m Passive Copper cable IB HDR, QSFP/QSFP

(No longer available as of December 31, 2020)

2 metre length copper twinax cable, also called a DAC (Direct Attached Copper) cable. The cable can be used for either 100Gb HDR InfiniBand (IB) adapters or IB switches. Built onto each end of the cable is a passive Quad (4-channel) Small Form-Factor Pluggable (QSFP28) copper cable transceiver. Built onto each end of the cable is a passive QSFP copper cable transceiver. Cables are available in various lengths: 1M - #EBB1, 1.5M=#EBB2, 2M=#EBB3.

  • Attributes provided: 2m QSFP/QSFP Cable
  • Attributes required: QSFP/QSFP ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBA) - 3m Active Optical cable IB HDR QSFP56

(No longer available as of August 9, 2020)

3 metre length Active Optical cable for InfiniBand HDR, also called an AOC (Active Optical Cable). The QSFP56 cable is an assembly of 4full-duplex lanes, each lane is capable of transmitting data at rates up to 50Gb/s per direction, providing up to 200Gb/s connectivity between devices using QSFP56 ports.Cables are available in various lengths: 3m = #EBBA, 5M = #EBBB, 10M= #EBBC, 15M = #EBBD, 20M = #EBBE, and 30M = #EBBF. See also coppertwinax cables for shorter lengths such as #EBB1 through #EBB3.

  • Attributes provided: 3m active optical cable QSFP56
  • Attributes required: QSFP56 ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBB) - 5m Active Optical cable IB HDR, QSFPP/QSFP

(No longer available as of August 9, 2020)

This 5 metre HDR 100 Gb/s InfinBand optical cable is used to connect the QDR Host Channel Adapter to QDR InfiniBand Switches. It can alsobe used for QDR InfiniBand switch-to-switch applications.Cables are available in various lengths: 3m = #EBBA, 5M = #EBBB, 10M= #EBBC, 15M = #EBBD, 20M = #EBBE, and 30M = #EBBF. See also coppertwinax cables for shorter lengths such as #EBB1 through #EBB3.Limitation: Adapter and switch must have been manufactured byMellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: 5m HDR 100 Gb/s InfiniBand Optical Cable, QSFP/QSFP
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBC) - 10m Active Optical cable IB HDR, QSFP/QSFP

(No longer available as of August 9, 2020)

This 10 metre HDR 100 Gb/s InfinBand optical cable is used to connect the QDR Host Channel Adapter to QDR InfiniBand Switches. It can also be used for QDR InfiniBand switch-to-switch applications.Cables are available in various lengths: 3m = #EBBA, 5M = #EBBB, 10M= #EBBC, 15M = #EBBD, 20M = #EBBE, and 30M = #EBBF. See also coppertwinax cables for shorter lengths such as #EBB1 through #EBB3.Limitation: Adapter and switch must have been manufactured byMellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: 10m HDR 100 Gb/s InfiniBand Optical Cable, QSFP/QSFP
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBE) - 20m Active Optical cable IB HDR, QSFP/QSFP

(No longer available as of August 9, 2020)

This 20 metre HDR 100 Gb/s InfinBand optical cable is used to connect the QDR Host Channel Adapter to QDR InfiniBand Switches. It can also be used for QDR InfiniBand switch-to-switch applications.Cables are available in various lengths: 3m = #EBBA, 5M = #EBBB, 10M= #EBBC, 15M = #EBBD, 20M = #EBBE, and 30M = #EBBF. See also coppertwinax cables for shorter lengths such as #EBB1 through #EBB3.Limitation: Adapter and switch must have been manufactured byMellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: 20m HDR 100 Gb/s InfiniBand Optical Cable, QSFP/QSFP
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBF) - 30m Active Optical cable IB HDR, QSFP/QSFP

(No longer available as of August 9, 2020)

This 30 metre HDR 100 Gb/s InfinBand optical cable is used to connect the QDR Host Channel Adapter to QDR InfiniBand Switches. It can also be used for QDR InfiniBand switch-to-switch applications.Cables are available in various lengths: 3m = #EBBA, 5M = #EBBB, 10M= #EBBC, 15M = #EBBD, 20M = #EBBE, and 30M = #EBBF. See also coppertwinax cables for shorter lengths such as #EBB1 through #EBB3.Limitation: Adapter and switch must have been manufactured byMellanox to use this cable. Cable is not supported for FDR IB (56Gb).

  • Attributes provided: 30m HDR 100 Gb/s InfiniBand Optical Cable, QSFP/QSFP
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBK) - 1m Passive Copper Splitter cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

1m length DAC splitter cable can provide connectivity between system units with a 200Gb/s HDR QSFP56 port on one side and two 100Gb/s HDR100 QSFP56 ports on the other side. The cable connects the data signals from each of the 2 dual copper pairs on the single QSFP56 (pair 1&2, 3&4) end to the dual copper pair of each of the QSFP56 (pair 1&2) ends on the multiport side. Cables are available in various lengths: 1m = #EBBK, 1.5M = #EBBL, and 2M = #EBBM.

  • Attributes provided: 1m QSFP56 passive copper splitter cable
  • Attributes required: QSFP56 ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBL) - 1.5m Passive Copper Split cable, IB HDR 200GB/s to 2X100GB/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

This 1.5 metre Passive Copper Splitter cable, IB HDR 200 Gb/s to 2x100 Gb/s is used to get 80 x HDR 100Gb/s ports. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This copper cable has QSFP56 transceivers already attached to each end. Cables are available in various lengths: 1m = #EBBK, 1.5M = #EBBL, and 2M = #EBBM.

  • Attributes provided: 1.5m QSFP56 passive copper splitter cable
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBM) - 2m Passive Copper Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

This 2 metre Passive Copper Splitter cable, IB HDR 200 Gb/s to 2x100 Gb/s is used to get 80 x HDR 100Gb/s ports. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This copper cable has QSFP56 transceivers already attached to each end. Cables are available in various lengths: 1m = #EBBK, 1.5M = #EBBL, and 2M = #EBBM.

  • Attributes provided: 1.5m QSFP56 passive copper splitter cable
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBQ) - 3m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

3 length active optical splitter cables for InfiniBand HDR provides QSFP56 to 2xQSFP56 cable. The cable contains 4 full-duplex lanes, each lane is capable of transmitting data at rates up to 50Gb/s per direction, using 50Gb/s PAM4 modulation, providing an aggregated rate of 200Gb/s. The 200G HDR end is configured as 4-channels x 50G PAM4 in a QSFP56 shell. The HDR100 side splits off 2 channels x 50G PAM4 into two separate QSFP56 ends operating at 100Gb/s. Cables are available in various lengths. 3m = #EBBQ, 5M = #EBBR, 10M = #EBBS, 15M = #EBBT, 20M = #EBBU, and 30M = #EBBV.

  • Attributes provided: 3m QSFP56 active optical splitter cable
  • Attributes required: QSFP56 ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBR) - 5m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

5 metre length Active Optical fibre Cable (AOC), IB HDR 200 Gb/s to 2x100 Gb/s is used to get 80 x HDR 100Gb/s ports. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This active optical cable has QSFP56 transceivers already attached to each end. Cables are available in various lengths. 3m = #EBBQ, 5M = #EBBR, 10M = #EBBS, 15M = #EBBT, 20M = #EBBU, and 30M = #EBBV.

  • Attributes provided: Optical fibre cable with QSFP56 transceivers
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBS) - 10m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

10 metre length Active Optical fibre Cable (AOC), IB HDR 200 Gb/s to 2x100 Gb/s is used to get 80 x HDR 100Gb/s ports. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This active optical cable has QSFP56 transceivers already attached to each end. Cables are available in various lengths. 3m = #EBBQ, 5M = #EBBR, 10M = #EBBS, 15M = #EBBT, 20M = #EBBU, and 30M = #EBBV.

  • Attributes provided: Optical fibre cable with QSFP56 transceivers
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBU) - 20m Active Optical Split cable, 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

20 metre length Active Optical fibre Cable (AOC), IB HDR 200 Gb/s to 2x100 Gb/s is used to get 80 x HDR 100Gb/s ports. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This active optical cable has QSFP56 transceivers already attached to each end. Cables are available in various lengths. 3m = #EBBQ, 5M = #EBBR, 10M = #EBBS, 15M = #EBBT, 20M = #EBBU, and 30M = #EBBV.

  • Attributes provided: Optical fibre cable with QSFP56 transceivers
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EBBV) - 30m Active Optical Split cable, IB HDR 200Gb/s to 2x100Gb/s, QSFP56 to 2xQSFP56

(No longer available as of August 9, 2020)

30 metre length Active Optical fibre Cable (AOC), IB HDR 200 Gb/s to 2x100 Gb/s is used to get 80 x HDR 100Gb/s ports. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This active optical cable has QSFP56 transceivers already attached to each end. Cables are available in various lengths. 3m = #EBBQ, 5M = #EBBR, 10M = #EBBS, 15M = #EBBT, 20M = #EBBU, and 30M = #EBBV.

  • Attributes provided: Optical fibre cable with QSFP56 transceivers
  • Attributes required: IB switch
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC16) - Open Power non-virtualized configuration

Open Power firmware for Power non-virtualized configuration.

  • Attributes provided: Allows Linux to be used on bare metal without any virtualization
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#EC2R) - PCIe3 LP 2-Port 10Gb NIC&ROCE SR/Cu Adapter

(No Longer Available as of August 31, 2021)

This PCIe Gen3 Ethernet adapter provides two 10 Gb SFP+ ports. The adapter supports both NIC and IBTA RoCE standards. RoCE is Remote Direct Memory Access (RDMA) over Converged Ethernet. Using RoCE, the adapter can support significantly greater bandwidth with low latency and minimise CPU overhead by more efficiently using memory access. This offloads the CPU from I/O networking tasks, improving performance and scalability.

Cables: For 10GbE, IBM offers Direct Attach (DAC) cables up to 5m. SFP-based transceivers are included on each end of the cables. See features EN01, EN02, EN03.

Transceivers: IBM qualifies and supports SFP+ optical transceiver (FC EB46) to install into the adapter. Customers can also use their own optical cabling and SFP+ optical transceiver for the other end. The 10 Gb optical transceiver is capable up to 300 M through the OM3 cable or 82 M through OM2 cable. Either one or both of the adapter's two SFP+ ports can be populated.

Feature code #EC2R and #EC2S have identical electronics and function and CCIN (58FA), but have different tail stock brackets. #EC2R is low profile and #EC2S is full high. The adapter is based on a Mellanox ConnectX-4 adapter which uses a ConnectX-4 Lx EN Network Controller.

Attributes:

  • PCI Express 3.0 (up to 8GT/s) x8
  • PCIe Gen 3.0 compliant, 1.1 and 2.0 compatible
  • RDMA over Converged Ethernet (RoCE)
  • NIC and RoCE are concurrently supported
  • RoCE supported on Linux and AIX (7.2 and later)
  • NIC supported on all OSes
  • IEEE 802.3ae (10Gb Ethernet), IEEE 802.3ad (Link Aggregation & Failover), IEEE 802.3az (Energy Efficient Ethernet), IEEE 802.1Q/P (VLAN Tagging), IEEE 802.10au (Congestion Notification), IEEE 802.1Qbg, IEEE 802.3Qaz D0.2 (ETS), IEEE 802.1Qbb D1.0 (PFC), IEEE 1588v2 (PTP)
  • Jumbo frame support up to 9.6KB
  • VXLAN and NVGRE Overlay Network offload support
  • TCP/UDP/IP stateless offload
  • TCP checksum offload
  • TCP segmentation offload
  • UDP checksum offload
  • MSI-X, MSI and support of legacy pin interrupt
  • NIM boot support
  • PowerVM SR-IOV support

    For SR-IOV FAQs visit https://community.ibm.com/community/user/power/viewdocument/sr-iov-faq s-t est?CommunityKey=71e6bb8a-5b34-44da-be8b-277834a183b0&tab=librarydocum ent s#Q3

  • Attributes provided: 2-port 10Gb Ethernet
  • Attributes required: Low profile PCIe Gen3 slot
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat supported with Mellanox OFED 4.3, or later
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC2T) - PCIe3 LP 2-Port 25/10Gb NIC&ROCE SR/Cu Adapter

This PCIe Gen3 Ethernet adapter provides two 25/10 Gb SFP28 ports. The adapter supports both NIC and IBTA RoCE standards. RoCE is Remote Direct Memory Access (RDMA) over Converged Ethernet. Using RoCE, the adapter can support significantly greater bandwidth with low latency and minimise CPU overhead by more efficiently using memory access. This offloads the CPU from I/O networking tasks, improving performance and scalability.

Cables: For 25GbE, IBM offers SFP28 Passive Copper 25Gb Ethernet cables up to 2m. SFP28 based transceivers are included on each end of these cables. See features EB4J, EB4K, EB4L, and EB4M for a 0.5M, 1.0M, 1.5M and 2.0M copper cable.

For 10GbE, IBM offers Direct Attach (DAC) cables up to 5m. SFP-based transceivers are included on each end of the cables. See features EN01, EN02, EN03.

Transceivers: For 25 GbE, IBM qualifies and supports SFP28 optical transceiver (FC EB47) to install into the adapter. Customers can also use their own optical cabling and SFP28 optical transceiver for the other end. The 25 Gb optical transceiver is capable up to 100 m through the OM4 cable or 70 M through OM3 cable. Either one or both of the adapter's two SFP28 ports can be populated.
Note: The (FC EB47) SFP28 25GbE transceiver only supports 25GbE speeds.

For 10 GbE, IBM qualifies and supports SFP+ optical transceiver (FC EB46) to install into the adapter. Customers can also use their own optical cabling and SFP+ optical transceiver for the other end. The 10 Gb optical transceiver is capable up to 300 M through the OM3 cable or 82 m through OM2 cable. Either one or both of the adapter's two SFP28 ports can be populated.

Feature code #EC2T and #EC2U have identical electronics and function and CCIN (58FB), but have different tail stock brackets. #EC2T is low profile and #EC2U is full high. The adapter is based on a Mellanox ConnectX-4 adapter which uses a ConnectX-4 Lx EN Network Controller.

Attributes:

  • PCI Express 3.0 (up to 8GT/s) x8
  • PCIe Gen 3.0 compliant, 1.1 and 2.0 compatible
  • RDMA over Converged Ethernet (RoCE)
  • NIC and RoCE are concurrently supported
  • RoCE supported on Linux and AIX (7.2 and later)
  • NIC supported on all OSes
  • IEEE 802.3ae (25Gb or 10Gb Ethernet), IEEE 802.3ad (Link Aggregation & Failover), IEEE 802.3az (Energy Efficient Ethernet), IEEE 802.1Q/P (VLAN Tagging), IEEE 802.10au (Congestion Notification), IEEE 802.1Qbg, IEEE 802.3Qaz D0.2 (ETS), IEEE 802.1Qbb D1.0 (PFC), IEEE 1588v2 (PTP)
  • Jumbo frame support up to 9.6KB
  • VXLAN and NVGRE Overlay Network offload support
  • TCP/UDP/IP stateless offload
  • TCP checksum offload
  • TCP segmentation offload
  • UDP checksum offload
  • MSI-X, MSI and support of legacy pin interrupt
  • NIM boot support
  • PowerVM SR-IOV support

    For SR-IOV FAQs visit https://community.ibm.com/community/user/power/viewdocument/sr-iov-faq s-t est?CommunityKey=71e6bb8a-5b34-44da-be8b-277834a183b0&tab=librarydocum ent s#Q3

  • Attributes provided: 2-port 25/10Gb Ethernet
  • Attributes required: Low profile PCIe Gen3 slot
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat supported with Mellanox OFED 4.3, or later
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC3L) - PCIe3 LP 2-port 100GbE (NIC& RoCE) QSFP28 Adapter x16

(No longer available as of June 25, 2021)

This PCIe Gen3 Ethernet x16 adapter provides two 100 Gb QSFP28 ports. The adapter supports both NIC and IBTA RoCE standards. RoCE is Remote Direct Memory Access (RDMA) over Converged Ethernet. Using RoCE, the adapter can support significantly greater bandwidth with low latency and minimise CPU overhead by more efficiently using memory access. This offloads the CPU from I/O networking tasks, improving performance and scalability.

IBM offers either passive copper twinax cables up to 2 metre in length or active optical cables up to 100 meters in length. See features #EB5J - #EB5M for a 0.5M, 1.0M, 1.5M and 2.0M copper cable. See features #EB5R - #EB5Y for a 3M, 5M, 10M, 15M, 20M, 30M, 50M or 100M active optical cable. Transceivers are included on each end of these QSFP28 cables. Alternatively to the above supported cables, you may chose to order an IBM qualified and supported QSFP28 optical transceiver (feature #EB59) to put into the adapter and provide your own 100GE optical cabling with your own QSP28 optical transceiver for the other end.

Either one or both of the adapter's two QSP28 ports can be populated. When two ports are filled, both can have copper cables, both can have optical cables, or one can be copper and one can be optical.

Feature code #EC3L and #EC3M have identical electronics and function and CCIN (2CEC), but have different tail stock brackets. #EC3L is low profile and #EC3M is full high. The adapter is based on a Mellanox ConnectX-4 adapter which uses a ConnectX-4 EN Network Controller.

Attributes:

  • PCI Express 3.0 (up to 8GT/s) x16
  • PCIe Gen 3.0 compliant, 1.1 and 2.0 compatible
  • RDMA over Converged Ethernet (RoCE)
  • NIC and RoCE are concurrently supported
  • RoCE supported on Linux and AIX (7.2 and later)
  • NIC supported on all OSes
  • TCP/UDP/IP stateless offload
  • LSO, LRO, checksum offload
  • NIM boot support
  • Backward compatible with 40Gb Ethernet when using compatible cables/transceivers.
  • PowerVM SR-IOV support

    For SR-IOV FAQs visit https://community.ibm.com/community/user/power/viewdocument/sr-iov-faq s-t est?CommunityKey=71e6bb8a-5b34-44da-be8b-277834a183b0&tab=librarydocum ent s#Q3

  • Attributes provided: 2-port 100Gb Ethernet
  • Attributes required: x16 PCIe Gen3 slot
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
    • Red Hat supported with Mellanox OFED 4.3, or later
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC4H) - NVIDIA Tesla V100 GPU with NVLink Water-Cooled (16 GB)

(No longer available as of February 23, 2021)

  • Minimum required: 0
  • Maximum allowed: 6 (Initial order maximum: 6)
  • OS level required:
    • Red Hat Enterprise Linux 7.6
    • Ubuntu 18.04
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: No
(#EC4K) - NVIDIA Tesla V100 GPU with NVLink Water-Cooled (32 GB)
  • Minimum required: 0
  • Maximum allowed: 6 (Initial order maximum: 6)
  • OS level required:
    • Red Hat Enterprise Linux 7.6
    • Ubuntu 18.04
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: No
(#EC5A) - PCIe3 LP 1.6 TB SSD NVMe adapter

(No longer available as of April 30, 2019)

The PCIe3 x8 NVMe 1.6 TB SSD NVMe Adapter is a Peripheral Component Interconnect Express (PCIe) generation 3 (Gen3) x8 adapter. The adapter can be used in either a x8 or x16 PCIe (Gen3) slot in the system and uses Non-Volatile Memory Express (NVMe).

NVMe is a high performance software interface that can read/write flash memory. Compared to a SAS or SATA SSD, the NVMe Flash adapter provides more read/write input/output operations per second (IOPS) and larger throughput (GB/sec). The NVMe Flash Adapter is read intensive and is not designed for high write intensive workloads. At about 8,760 to 17,000 TB of writes to the adapter, it will be at its maximum projected write capability. The nature of the workload has a great impact on the maximum write capacity. If a high percentage of more sequentially oriented writes is used instead of random writes, the maximum write capacity will be closer to the larger value in the range.

In the case of a high percentage of random writes, the maximum will be closer to the smaller value in the range. Writes past the adapter's maximum write capacity will continue to work for some period of time, but much more slowly. A Predictive Failure Analysis message will indicate that it is time to replace the adapter if enabled by the system administrator. After the warranty period, if the maximum write capability is achieved, the adapter's replacement is not covered under IBM maintenance.

This adapter has protection against single Flash channel failures. To prevent the entire adapter from failing, software RAID should be used. For high value applications where the content in the adapter must be protected, additional NVMe Flash adapters with OS mirroring or software Redundant Array of Independent Disks (RAID) are recommended. This adapter is not supported in the PCIe Gen3 I/O drawer.

This PCIe NVMe device can have 32 namespaces per device.

  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat Enterprise Linux
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC5C) - PCIe3 x8 LP 3.2 TB NVMe Flash adapter for AIX/Linux

The PCIe3 x8 3.2 TB NVMe Adapter is a Peripheral Component Interconnect Express (PCIe) generation 3 (Gen3) x8 adapter. The adapter can be used in either a x8 or x16 PCIe slot in the system and uses Non-Volatile Memory Express (NVMe). NVMe is a high performance software interface that can read/write flash memory. Compared to a SAS or SATA SSD, the NVMe Flash adapter provides more read/write input/output operations per second (IOPS) and larger throughput (GB/sec).

CCIN is 58FD.

Feature #EC5C and #EC5D are identical cards except that the tailstock bracket is different. #EC5C fits a low profile PCIe slot. #EC5D fits a full high PCIe slot. See also EC5E for a card with more memory.

The nature of the workload has a great impact on the maximum write capacity. If a high percentage of more sequentially oriented writes are used instead of random writes, the maximum write capacity will be larger. To extend the life of the device, the application that is using the device must convert small random writes to larger sequential writes. Writes past the adapter's maximum write capacity will continue to work for some period of time, but much more slowly. Whether the application uses sequential or random reads from the device does not affect the life of the device. A Predictive Failure Analysis message will indicate that it is time to replace the adapter if enabled by the system administrator. Customers are recommended to monitor the smart log via their operating system where fuel gauge shows the percentage used.

IBM NVMe adapter failures will be replaced during the standard warranty and maintenance period for adapters that have not reached the maximum number of write cycles. Adapters that reach this limit may fail to operate according to specifications and must be replaced at the client's expense. Data protection is not implemented in the card; protection is provided by OS mirroring or software RAID wherever applicable.

This adapter is not supported in the PCIe I/O drawer.

This PCIe NVMe device can have 32 namespaces per device.

  • Attributes provided: 3.2 TB of low latency flash memory with boot capability
  • Attributes required: PCIe slot in system unit
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat Enterprise Linux
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC5E) - PCIe3 x8 LP 6.4 TB NVMe Flash adapter for AIX/Linux

The PCIe3 x8 6.4 TB NVMe Adapter is a Peripheral Component Interconnect Express (PCIe) generation 3 (Gen3) x8 adapter. The adapter can be used in either a x8 or x16 PCIe slot in the system and uses Non-Volatile Memory Express (NVMe). NVMe is a high performance software interface that can read/write flash memory. Compared to a SAS or SATA SSD, the NVMe Flash adapter provides more read/write input/output operations per second (IOPS) and larger throughput (GB/sec).

CCIN is 58FE.

Feature #EC5E and #EC5F are identical cards except that the tailstock bracket is different. #EC5E fits a low profile PCIe slot. #EC5F fits a full high PCIe slot. See also EC5C for a card with less memory.

The nature of the workload has a great impact on the maximum write capacity. If a high percentage of more sequentially oriented writes are used instead of random writes, the maximum write capacity will be larger. To extend the life of the device, the application that is using the device must convert small random writes to larger sequential writes. Writes past the adapter's maximum write capacity will continue to work for some period of time, but much more slowly. Whether the application uses sequential or random reads from the device does not affect the life of the device. A Predictive Failure Analysis message will indicate that it is time to replace the adapter if enabled by the system administrator. Customers are recommended to monitor the smart log via their operating system where fuel gauge shows the percentage used.

IBM NVMe adapter failures will be replaced during the standard warranty and maintenance period for adapters that have not reached the maximum number of write cycles. Adapters that reach this limit may fail to operate according to specifications and must be replaced at the client's expense. Data protection is not implemented in the card; protection is provided by OS mirroring or software RAID wherever applicable.

This adapter is not supported in the PCIe I/O drawer.

This PCIe NVMe device can have 32 namespaces per device.

  • Attributes provided: 6.4 TB of low latency flash memory with boot capability
  • Attributes required: PCIe slot in system unit
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat Enterprise Linux
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC5G) - PCIe3 x8 LP 1.6 TB NVMe Flash Adapter for AIX/Linux

The PCIe3 x8 1.6 TB NVMe Adapter is a Peripheral Component Interconnect Express (PCIe) generation 3 (Gen3) x8 adapter. The adapter can be used in either a x8 or x16 PCIe slot in the system and uses Non-Volatile Memory Express (NVMe). NVMe is a high performance software interface that can read/write flash memory. Compared to a SAS or SATA SSD, the NVMe Flash adapter provides more read/write input/output operations per second (IOPS) and larger throughput (GB/sec).

CCIN is 58FC.

Feature #EC5G and #EC5B are identical cards except that the tailstock bracket is different. #EC5G fits a low profile PCIe slot. #EC5B fits a full high PCIe slot. See also #EC5D/EC5F for a card with more memory.

The nature of the workload has a great impact on the maximum write capacity. If a high percentage of more sequentially oriented writes are used instead of random writes, the maximum write capacity will be larger. To extend the life of the device, the application that is using the device must convert small random writes to larger sequential writes. Writes past the adapter's maximum write capacity will continue to work for some period of time, but much more slowly. Whether the application uses sequential or random reads from the device does not affect the life of the device. A Predictive Failure Analysis message will indicate that it is time to replace the adapter if enabled by the system administrator. Customers are recommended to monitor the smart log via their operating system where fuel gauge shows the percentage used.

IBM NVMe adapter failures will be replaced during the standard warranty and maintenance period for adapters that have not reached the maximum number of write cycles. Adapters that reach this limit may fail to operate according to specifications and must be replaced at the client's expense. Data protection is not implemented in the card; protection is provided by OS mirroring or software RAID wherever applicable.

This adapter is not supported in the PCIe I/O drawer.

This PCIe NVMe device can have 32 namespaces per device.

  • Attributes provided: 1.6 TB of low latency flash memory with boot capability
  • Attributes required: PCIe slot in system unit
  • Limitations: Not supported in PCIe Gen3 I/O drawer. Data protection not implemented in the card and protection provided by OS mirroring or software RAID.
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Linux supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC62) - PCIe4 LP 1-port 100Gb EDR IB CAPI adapter

The PCIe Gen4 x16 1-port InfiniBand EDR adapter provides high speed connectivity with other servers or IB switches. The port's maximum of 100Gb assumes no other system and/or switch bottlenecks are present. A PCIe Gen4 x16 PCIe slot is required. This adapter is sourced from Mellanox Corporation. based on ConnectX-5 technology. The adapter supports the InfiniBand Trade Association (IBTA) specification version 2.

The 100Gb port has a QSFP+ connection which supports EDR cables, either EDR DAC or EDR optical. Transceivers are included in the cables. IBM cable features EB50-EB54 (copper shorter distance) and #EB5A-EB5H (optical longer distance) are supported or their copper or optical Mellanox equivalents are supported. Other cables are not supported.

#EC62 and #EC63 adapters are electronically and functionally identical with the same CCIN of 2CF1. #EC62 is low profile and #EC63 is full high.

Limitation: Adapter does not fit in x8 PCIe slot. Adapter is not supported in a PCIe Gen3 I/O drawer.

  • Attributes provided: EDR InfiniBand PCIe Adapter
  • Attributes required: available x16 PCIe Gen4 slot
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat supported with Mellanox OFED 4.3, or later
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC64) - PCIe4 LP 2-port 100Gb EDR IB CAPI adapter

The PCIe Gen4 x16 2-port InfiniBand EDR adapter provides high speed connectivity with other servers or IB switches. Each port maximum of 100Gb assumes no other system and/or switch bottlenecks are present. A PCIe Gen4 x16 PCIe slot is required. This adapter is sourced from Mellanox Corporation. based on ConnectX-5 technology. The adapter supports the InfiniBand Trade Association (IBTA) specification version 2.

The two 100Gb ports have QSFP+ connections which support EDR cables, either EDR DAC or EDR optical. One adapter can support either or both types of cable. The user can choose to cable up just one port if they desire. Transceivers are included in the cables. IBM cable features EB50-EB54 (copper shorter distance) and #EB5A-EB5H (optical longer distance) are supported or their copper or optical Mellanox equivalents are supported. Other cables are not supported.

#EC64 and #EC65 adapters are electronically and functionally identical with the same CCIN of 2CF2. #EC64 is low profile and #EC65 is full high.

Limitation: Adapter does not fit in x8 PCIe slot. Adapter is not supported in a PCIe Gen3 I/O drawer.

  • Attributes provided: EDR InfiniBand PCIe Adapter
  • Attributes required: available x16 PCIe Gen4 slot
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat supported with Mellanox OFED 4.3, or later
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC67) - PCIe4 LP 2-port 100Gb ROCE EN LP adapter

This PCIe Gen4 Ethernet x16 adapter provides two 100 GbE QSFP28 ports. The adapter supports both NIC and IBTA RoCE standards. RoCE is Remote Direct Memory Access (RDMA) over Converged Ethernet. Using RoCE, the adapter can support significantly greater bandwidth with low latency and minimise CPU overhead by more efficiently using memory access. This offloads the CPU from I/O networking tasks, improving performance and scalability.

For 100GbE network connectivity, IBM offers either passive copper twinax cables up to 2 meters in length or active optical cables up to 100 meters in length. See features #EB5J - #EB5M for a 0.5M, 1.0M, 1.5M and 2.0M copper cable. See features #EB5R - #EB5Y for a 3M, 5M, 10M, 15M, 20M, 30M, 50M or 100M active optical cable. Transceivers are included on each end of these QSFP28 cables. Alternatively to the above supported cables, you may choose to order an IBM qualified and supported QSFP28 optical transceiver (feature #EB59) to put into the adapter ports and MTP/MPO cable 10M or 30M in length (feature #EB2J or #EB2K) optical cabling to use with your own QSP28 optical transceiver for the other end.

For 40GbE network connectivity, IBM offers either passive copper twinax cables up to 3 meters in length. See features #EB2B and EB2H for a 1.0M and 3.0M copper cable. Transceivers are included on each end of these QSFP+ cables. Alternatively to the above supported cables, you may choose to order an IBM qualified and supported QSFP+ 40G BASE-SR optical transceiver (feature #EB27) to put into the adapter ports and MTP/MPO cable 10M or 30M in length (feature #EB2J or #EB2K) optical cabling to use with your own QSP28 optical transceiver for the other end.

Either one or both of the adapter's two QSP28 ports can be populated. When two ports are filled, both can have copper cables, both can have optical cables, or one can be copper and one can be optical.

Feature code #EC66 and #EC67 have identical electronics and function and CCIN (2CF3), but have different tail stock brackets. #EC66 is high profile and #EC67 is low profile. The adapter is based on a Mellanox ConnectX-5 adapter which uses a ConnectX-5 EN Network Controller.

Attributes:

  • PCI Express 4.0 (up to 16GT/s) x16
  • PCIe Gen 4.0 compliant, 1.1, 2.0 and 3.0 compatible
  • RDMA over Converged Ethernet (RoCE)
  • NIC and RoCE are concurrently supported
  • RoCE supported on Linux and AIX
  • NIC supported on all OSes
  • TCP/UDP/IP stateless offload
  • LSO, LRO, checksum offload
  • NIM boot support
  • Backward compatible with 40Gb Ethernet when using compatible cables/transceivers
  • PowerVM SR-IOV support

For SR-IOV FAQs visit https://community.ibm.com/community/user/power/viewdocument/sr-iov-faq s-test?CommunityKey=71e6bb8a-5b34-44da-be8b-277834a183b0&tab=library document s#Q3

  • Attributes provided: Dual-port 100 GbE Adapter with RoCE capability
  • Attributes required: Available Gen4 PCIe Slot
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
    • Red Hat Enterprise Linux 7 for POWER9, version 7.6, or later, with Mellanox OFED 4.6, or later
    • Red Hat Enterprise Linux 8, for POWER, or later, with Mellanox OFED 4.6, or later
    • Ubuntu Server 18.04.1, with Mellanox OFED 4.6, or later
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC6E) - PCIe4 LP 1-port HDR100 IB Adapter x16

(No longer available as of August 9, 2020)

The PCIe Gen4 1-port InfiniBand HDR100 x16 adapter provides high speed connectivity with other servers or IB switches. Each port maximum of 100Gb assumes no other system and/or switch bottlenecks are present. A PCIe Gen4 PCIe x16 slot is required. This adapter is sourced from Mellanox Corporation. based on ConnectX-6 technology. The adapter supports the InfiniBand Trade Association (IBTA) specification version 1.3.

The 100Gb port have QSFP56 connections which support HDR100 via splitter cables, either HDR DAC or HDR optical. Splitter cables allow a single HDR (200Gb) port to connect to 2 HDR100 (100Gb) ports thus doubling a typical HDR switches capacity from 40 ports to 80 ports. One adapter can support either or both types of cable. Transceivers are included in the cables. IBM cable features EBBK-EBBM (copper shorter distance) and EBBQ-EBBV (optical longer distance) are supported or their copper or optical Mellanox equivalents are supported. The adapter is also backward compatible to slower speeds EDR, FDR, QDR, DDR, and SDR using appropriate cables.

CCIN is 590D.

Limitation: Adapter does not fit in x8 PCIe slot. Adapter is not supported in a PCIe Gen3 I/O drawer.
Note: This product is being withdrawn and it not available for sales. Contact Douglas Gibbs (douglasg@ca.ibm.com) for further details

  • Attributes provided: HDR InfiniBand PCIe Adapter
  • Attributes required: available x16 PCIe Gen4 slot
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EC6G) - PCIe4 LP 2-port HDR100 IB Adapter x16

(No longer available as of August 9, 2020)

The PCIe Gen4 2-port InfiniBand HDR100 x16 adapter provides high speed connectivity with other servers or IB switches. Each port maximum of 100Gb assumes no other system and/or switch bottlenecks are present. A PCIe Gen4 PCIe x16 slot is required. This adapter is sourced from Mellanox Corporation. based on ConnectX-6 technology. The adapter supports the InfiniBand Trade Association (IBTA) specification version 1.3.

The two 100Gb ports have QSFP56 connections which support HDR100 via splitter cables, either HDR DAC or HDR optical. Splitter cables allow a single HDR (200Gb) port to connect to 2 HDR100 (100Gb) ports thus doubling a typical HDR switches capacity from 40 ports to 80 ports. One adapter can support either or both types of cable. The user can choose to cable up just one port if they desire. Transceivers are included in the cables. IBM cable features EBBK-EBBM (copper shorter distance) and EBBQ-EBBV (optical longer distance) are supported or their copper or optical Mellanox equivalents are supported. The adapter is also backward compatible to slower speeds EDR, FDR, QDR, DDR, and SDR using appropriate cables.

CCIN is 590E.

Limitation: Adapter does not fit in x8 PCIe slot. Adapter is not supported in a PCIe Gen3 I/O drawer.
Note: This product is being withdrawn and it is not available for sales. Contact Douglas Gibbs (douglasg@ca.ibm.com) for further details

  • Attributes provided: HDR InfiniBand PCIe Adapter
  • Attributes required: available x16 PCIe Gen4 slots
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ECBN) - 5m (16.4-ft), IBM Passive QSFP+ to QSFP+ Cable (DAC)

(No longer available as of January 18, 2019)

QSFP+ cable is used for 40Gb-to-40Gb Ethernet connectivity. clients can use this QSFP+ Direct Attach Cable for Ethernet connectivity. This copper cable has QSFP+ transceivers already attached to each end.

  • Attributes provided: 5m QSFP+ to QSFP+ Cable
  • Attributes required: QSFP/QSFP+ ports
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ECP0) - Cloud Private Solution

(No longer available as of March 5, 2021)

This feature indicates that it is a Cloud Private Solution.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: No
(#ECSF) - Custom Service Specify, Montpellier, France

(No longer available as of March 5, 2021)

Having #ECSF on the order, will cause the order to be routed to France and the machine to be internally routed to the CSC build area.

  • Attributes provided: Customization
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: N/A
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ECW0) - Optical Wrap Plug

A wrap plug is a small connector designed to perform a diagnostic test called a loopback test. This wrap plug is inserted into a SR optical port on a PCIe Fibre Channel adapter or a SR or LR optical port on a PCIe Ethernet adapter

This is a multi-mode LC fibre optic wrap plug with an inside/ outside optics diameter of 50/125. Its IBM part number as of early 2016 is 12R9314. An earlier equivalent function IBM part number which is no longer shipped is 11P3847.

It is strongly recommended that Fibre Channel adapters (HBAs) fill any empty adapter ports with a wrap plug. There is no technical issue leaving a port empty. However, filling all ports with a cable to a device/switch or with a wrap plug can speed the booting/IPLing of a partition and can avoid error messages uselessly pointing to a planned empty port.

There is no technical issue leaving an Ethernet port empty. Whether an Ethernet port is empty or contains a wrap plug should not impact boot/IPL time or impact empty-port messages.

  • Attributes provided: Wrap plug
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EHL9) - Marketing Specify - NVIDIA GPU
  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EHLU) - IBM Power Systems for SAS Viya (Linux)

(No Longer Available as of March 23, 2021)

This feature is applicable to IBM Power Systems for SAS Viya deals only. This is a no-cost feature for tracking purposes of the offering.

  • Attributes provided: None
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 1)
  • OS level required:
    • Red Hat Enterprise Linux 7.4
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EJ31) - Four Water-Cooled Planar Assembly
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
    • Red Hat Enterprise Linux
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: No
(#EJ34) - Six Water-Cooled Planar Assembly
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
    • Red Hat Enterprise Linux
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: No
(#EJTX) - Rack-mount Fixed Rail Kit

Feature provides a rack fixed rail kit used to install a rack-mount 2U server such as 8335-CGA or 8335-GTA in an IBM or OEM 19-inch rack. Rails adjust from approximately 585 mm (23 inches) to 755 mm (29.5 inches) in depth. See also feature #EJTY for alternative rails which allow he server to be slid in and out of the rack.

  • Attributes provided: Fixed rails
  • Attributes required: 2U server using these rails
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EL3Z) - PCIe2 LP 2-port 10/1GbE BaseT RJ45 Adapter

PCIe Gen2 short x8 adapter which provides two 10G-BaseT ports. The ports are RJ45. The ports default to auto negotiate the highest speed either 10Gb (10GBaseT), 1Gb (1000BaseT) or 100Mb (100BaseT) full duplex. Each RJ45 port's configuration is independent of the other. The adapter supports Ethernet NIC (Network Interface Card) traffic. CCIN is 2CC4.

The RJ45 ports use 4-pair CAT-6A cabling for distances of up to 100 meters or CAT-6 cabling for distances up to 37 meters. CAT5 cabling is not tested and is not supported.

#EN0W and #EN0X are electronically identical with the same CCIN of 2CC4. #EN0W has a full high tail stock and #EN0X has a low profile tail stock.

Details for the ports include:

  • LINUX NETWORK INSTALL is supported.
  • IEEE 802.3an (10GBASE-T), IEEE 802.3ab (1000BASE-T GbE), IEEEu (100BASE-T), 802.1p priority, 802.1Q VLAN tagging, 802.3x flow control, 802.3ad load-balancing and failover
  • Link aggregation, IEEE 802.3ad 802.3
  • Multiple MAC addresses / promiscuous mode (for PowerVM/VIOS) per interface
  • Message Signalling Interrupt (MSI-X, MSI) support of legacy pin interrupts
  • Ether II and IEEE 802.3 encapsulated frames
  • Jumbo frames up to 9.6 Kbytes
  • TCP checksum offload for IPv4 and IPv6
  • TCP segmentation Offload (TSO)
  • RSS (Receive Side Scaling) support for IPv4, IPv6 and UDP.
  • UDP checksum offload for IPv4 and IPv6
  • Linux provides software iSCSI support through the adapter.
  • Attributes provided: Two 10G-BaseT ports
  • Attributes required: PCIe Gen2 or Gen3 slot - low profile
  • Minimum required: 0
  • Maximum allowed: 4 (Initial order maximum: 4)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EL43) - PCIe3 LP 2-port 16Gb Fibre Channel Adapter

(No Longer Available as of June 08, 2021)

PCIe Gen3 16 Gigabit dual-port Fibre Channel (FC) Adapter is a high-performance 8x short form adapter based on the Emulex LPe16002B PCIe Host Bus Adapter (HBA). The adapter provides two ports of 16Gb Fibre Channel capability using SR optics. Each port can provide up to 16Gb Fibre Channel functions simultaneously.

Each port provides single initiator capability over a fibre link or with NPIV, multiple initiator capability is provided. The ports are SFP+ and include an optical SR transceiver. The ports have LC type connectors and utilize shortwave laser optics. The adapter operates at link speeds of 4, 8 and 16 Gbps and will automatically negotiate to the highest speed possible. LEDs on each port provide information on the status and link speed of the port.

The adapter connects to a Fibre Channel switch at 4Gb, 8Gb or 16Gb. It can directly attach to a device without a switch at 16Gb under Linux. Attachment without a switch is not supported at 4Gb or 8Gb

Feature #EL43 and #EL5B are electronically identical. They differ physically only that EL5B has a tail stock for full high PCIe slots and #EL43 has a tail stock for low profile PCIe slots. CCIN is 577F for both features.

Cables are the responsibility of the customer. Use multimode fibre optic cables with short-wave lasers that adhere to the following specifications:

  • OM4 - multimode 50/125 micron fibre, 4700 MHz*km bandwidth
  • OM3 - multimode 50/125 micron fibre, 2000 MHz*km bandwidth
  • OM2 - multimode 50/125 micron fibre, 500 MHz*km bandwidth
  • OM1 - multimode 62.5/125 micron fibre, 200 MHz*km bandwidth

The maximum cable lengths at the three different link speeds are:

Cable | 4 Gbps | 8 Gbps | 16 Gbps
------|--------|--------|-------------
OM4   | 400m   | 190m   | 125m
OM3   | 380m   | 150m   | 100m
OM2   | 150m   | 50m    | 35m
OM1   | 70m    | 21m    | 15m
  • Attributes provided: Two 16Gb FC ports (with LC connectors)
  • Attributes required: PCIe Gen3 or Gen2 slot in supported server
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EL4M) - PCIe2 LP 4-port 1GbE Adapter

(No Longer Available as of April 13, 2021)

This short PCIe Gen2 adapter provides four 1Gb Ethernet ports that can be configured to run at 1000, 100 or 10 Mbps. 4-pair CAT-5 Unshielded Twisted Pair (UTP) cables up to 100 meters in length are attached to the copper RJ45 connectors. Each port is independent of one another and supports full-duplex or half-duplex. 1000 Mbps speed is not supported in Half Duplex (HDX) mode. Feature #EL4M and #EL4L are electronically identical and have the same CCIN of 576F. #EL4M indicates a low profile tail stock while #EL4L indicates a full high tail stock. (See also equivalent features #5260 and #5899).

Details for the ports include:

  • Linux install support
  • IEEE 802.3ab (1 GbE), 802.1p priority, 802.1Q VLAN tagging, 802.3x flow control, 802.3ad load-balancing and failover,
  • Link aggregation, IEEE 802.3ad 802.3
  • Multiple MAC addresses per interface
  • MSI-X, MSI and support of legacy pin interrupts
  • Ether II and IEEE 802.3 encapsulated frames
  • Jumbo frames up to 9.6 Kbytes
  • TCP checksum offload for IPv4 and IPv6
  • TCP segmentation Offload (TSO) for IPv4 and IPv6
  • UDP checksum offload for IPv4 and IPv6
  • Linux provide software iSCSI support through the adapter. Linux can also leverage adapter hardware support including initiator and header & data digest (CRC) generation and checking
  • Attributes provided: Four-port 1 Gb Ethernet
  • Attributes required: 1 Low Profile (LP) PCIe slot
  • Minimum required: 0
  • Maximum allowed: 4 (Initial order maximum: 4)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EL5V) - PCIe3 LP 2-port 32Gb Fibre Channel Adapter

PCIe Gen3 32 Gigabit dual-port Optical Fibre Channel (FC) Adapter is a high-performance 8x short form adapter based on the Broadcom LPe32000-series PCIe Host Bus Adapter (HBA). The adapter provides two ports of 32Gb Fibre Channel capability using SR optics. Each port can provide up to 32Gb Fibre Channel functions simultaneously.

Each port provides single initiator capability over a fibre link or with NPIV, multiple initiator capability is provided. The ports are SFP+ and include an optical SR transceiver. The ports have LC type connectors and utilize shortwave laser optics. The adapter operates at link speeds of 4, 8, 16 and 32Gbps and will automatically negotiate to the highest speed possible.

Feature #EL5U and #EL5V are electronically identical. They differ physically only that EL5U has a tail stock for full high PCIe slots and #EL5V has a short tail stock for low profile PCIe slots.

CCIN is 578F for both features.

Each port has two LED indicators located on the bracket next to each connector. These LEDs communicate boot status and give a visual indication of the operating state. The LEDs have five defined states; solid on, solid off, slow blink, fast blink, and flashing. The slow blink rate is 1Hz, the fast blink is 4Hz, and the flashing refers to an irregular on/off transition that reflects test progress. The operator should observe the LED sequence for several seconds to ensure that the operating state is correctly identified.

Cables:

Cables are the responsibility of the customer. Use multimode fibre optic cables with short-wave lasers that adhere to the following specifications:

  • OM4 - multimode 50/125 micron fibre, 4700 MHz*km bandwidth with LC connectors o 4GFC: 0.5m - 400m o 8GFC: 0.5m - 190m o 16GFC: 0.5m - 125m o 32GFC: 0.5m - 100m
  • OM3 - multimode 50/125 micron fibre, 2000 MHz*km bandwidth with LC connectors o 4GFC: 0.5m - 380m o 8GFC: 0.5m - 150m o 16GFC: 0.5m - 100m o 32GFC: 0.5m - 70m
  • OM2 - multimode 50/125 micron fibre, 500 MHz*km bandwidth with LC connectors o 4GFC: 0.5m - 150m o 8GFC: 0.5m - 50m o 16GFC: 0.5m - 35m o 32GFC: 0.5m - 20m
  • OM1 - multimode 62.5/125 micron fibre, 200 MHz*km bandwidth with LC connectors o 4GFC: 0.5m - 70m o 8GFC: 0.5m - 21m o 16GFC: 0.5m - 15m o 32GFC: N/A

Note: The H/W cannot detect what length and type of cable is installed. The link will auto-negotiate to the speed reported during negotiation by the Target. The user must manually set the maximum negotiation speed. If too high of speed is selected, bit errors may occur.

See also optional wrap plug feature #ECW0 which is: a) Required to run some diagnostic procedures and b) In some cases may speed system boot when placed in empty ports as well as avoid useless messages pointing to a planned empty port.

  • Attributes provided: 2-port 32Gb Optical FC
  • Attributes required: Low profile PCIe Gen3 slot
  • Minimum required: 0
  • Maximum allowed: 3 (Initial order maximum: 3)
  • OS level required:
    • Red Hat Enterprise Linux
    • Refer to Software Requirements for specific O/S levels supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ELD0) - 1 TB 7.2k RPM 5xx SATA SFF-4 Disk Drive

1 TB 7200 rpm SATA 2.5-inch (SFF) disk drive (HDD) formatted in 512 byte sectors (5xx) mounted on a 2.5-inch SFF-4 carrier/tray. Drive is rated at 6Gb/s though actual performance is dependent upon multiple factors. Limitations: Drive does not fit in 2.5-inch (SFF-1, SFF-2, SFF-3) bays due to physical differences in carriers/trays. Drive does not support the SAS interfaces. 5xx drives can not be reformatted to 4k byte sectors (or vice versa).

  • Attributes provided: 1TB of disk storage
  • Attributes required: One SFF-4 bay
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ELU4) - 960 GB 2.5-inch SATA/SSD Disk Drive
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ELU5) - 1.92 TB 2.5-inch SATA/SSD Disk Drive
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ELU6) - 3.84 TB 2.5-inch SATA/SSD Disk Drive
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EM60) - 8 GB DDR4 Memory

(No longer available as of August 7, 2020)

Provides 8 GB 2666 MHz DDR4 of system memory.

  • Attributes provided: 8 GB RDIMM memory
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EM61) - 16 GB DDR4 Memory

Provides 16 GB 2666 MHz DDR4 of system memory.

  • Attributes provided: 16 GB RDIMM memory
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EM63) - 32 GB DDR4 Memory

Provides 32 GB of DDR4 system memory.

  • Attributes provided: 32 GB DDR4 memory dimm
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EM64) - 64 GB DDR4 Memory

(No longer available as of March 5, 2021)

Provides 64 GB of DDR4 system memory.

  • Attributes provided: 64 GB DDR4 memory dimm
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
  • Initial Order/MES/Both/Supported: MES
  • CSU: Yes
  • Return parts MES: No
(#EM65) - 128 GB DDR4 Memory

(No longer available as of March 5, 2021)

Provides 128 GB of DDR4 system memory.

  • Attributes provided: 128 GB DDR4 memory dimm
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
  • Initial Order/MES/Both/Supported: MES
  • CSU: Yes
  • Return parts MES: No
(#EM7B) - 64 GB DDR4 Memory Dimm

Provides 64 GB running at 2666Mhz 16Gbit DDR4 system memory.

  • Attributes provided: 64GB Memory Dimm
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
    • Linux supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No

    Editor Note: FW 950, or later is required

(#EM7C) - 128 GB DDR4 Memory Dimm

Provides 128 GB running at 2666Mhz 16Gbit DDR4 system memory.

  • Attributes provided: 128 GB Memory Dimm
  • Attributes required: Available memory slot
  • Minimum required: 0
  • Maximum allowed: 16 (Initial order maximum: 16)
  • OS level required:
    • Linux supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No

    Editor Note: FW 950, or later is required

(#EN01) - 1m (3.3-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper

1m (3.3-ft) copper active twinax Ethernet cable which supports Ethernet data transfer rates up to 10 Gb/s. The cable has a copper twinax transceiver on each end which is placed in an SFP+ port of an adapter and/or a switch. This cabling option can be a cost effective alternative to optical cable for short reach link high-speed connection.

  • Attributes provided: 10Gb/s copper active twinax Ethernet cable
  • Attributes required: One available SFP+ 10Gb/s Ethernet Port
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
    • Red Hat Enterprise Linux
    • Refer to Software Requirements for specific O/S levels supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EN02) - 3m (9.8-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper

3m (9.8-ft) copper active twinax Ethernet cable which supports Ethernet data transfer rates up to 10 Gb/s. The cable has a copper twinax transceiver on each end which is placed in an SFP+ port of an adapter and/or a switch. This cabling option can be a cost effective alternative to optical cable for short reach link high-speed connection.

  • Attributes provided: 10Gb/s copper active twinax Ethernet cable
  • Attributes required: One available SFP+ 10Gb/s Ethernet Port
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
    • Red Hat Enterprise Linux
    • Refer to Software Requirements for specific O/S levels supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EN03) - 5m (16.4-ft), 10Gb E'Net Cable SFP+ Act Twinax Copper

5m (16.4-ft) copper active twinax Ethernet cable which supports Ethernet data transfer rates up to 10 Gb/s. The cable has a copper twinax transceiver on each end which is placed in an SFP+ port of an adapter and/or a switch. This cabling option can be a cost effective alternative to optical cable for short reach link high-speed connection.

  • Attributes provided: 10Gb/s copper active twinax Ethernet cable
  • Attributes required: One available SFP+ 10Gb/s Ethernet Port
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
    • Red Hat Enterprise Linux
    • Refer to Software Requirements for specific O/S levels supported
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EN0T) - PCIe2 LP 4-Port (10Gb+1GbE) SR+RJ45 Adapter

PCIe Gen2 x8 short Ethernet adapter supports Ethernet NIC (Network Interface Card) traffic. The adapter provides two 10 Gb SR optical ports and two 1 Gb RJ45 ports in a PCIe 8x short form adapter.

The 10Gb ports are SFP+ and include optical SR transceivers. The ports have LC Duplex type connectors and utilize shortwave laser optics and MMF-850nm fibre cabling. With 62.5 micron OM1, up to 33 metre length fibre cables are supported. With 50 micron OM2, up to 82 metre fibre cable lengths are supported. With 50 micron OM3 or OM4, up to 300 metre fibre cable lengths are supported.

For the 1Gb RJ45 ports, 4-pair CAT-5 Unshielded Twisted Pair (UTP) cable or higher is supported for distances of up to 100 meters. In addition to 1Gb (1000 MB) networks, 100 MB networks are also supported, but 10Mb networks are not supported.

#EN0S and #EN0T adapters are electronically identical. They are physically identical except #EN0S has a tail stock for full high PCIe slots and #EN0T has a tail stock for low profile slots. The CCIN is 2CC3 for both features.

Details for the ports include:

  • LINUX NETWORK INSTALL are supported.
  • IEEE 802.3ae (10GBASE-SR), IEEE 802.3ab (1000BASE-T GbE), IEEEu 802.3u (100BASE-T), 802.1p priority, 802.1Q VLAN tagging, 802.3x flow control, 802.3ad load-balancing and failover,
  • Link aggregation, IEEE 802.3ad 802.3
  • Message Signalling Interrupt MSI-X, MSI and support of legacy pin interrupts
  • Ether II and IEEE 802.3 encapsulated frames
  • Jumbo frames up to 9.6 Kbytes
  • TCP checksum offload for IPv4 and IPv6
  • TCP segmentation Offload (TSO) RSS (Receive Side Scaling) support for IPv4, IPv6 and UDP for IPv4 and IPv6
  • UDP checksum offload for IPv4 and IPv6
  • Linux provide software iSCSI support through the adapter.
  • Attributes provided: Four ports - two 10Gb and two 1Gb E
  • Attributes required: PCIe Gen2 or Gen3 slot - low profile
  • Minimum required: 0
  • Maximum allowed: 4 (Initial order maximum: 4)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EN0V) - PCIe2 LP 4-port (10Gb+1GbE) Copper SFP+RJ45 Adapter

PCIe Gen2 x8 short Ethernet adapter supports Ethernet NIC (Network Interface Card) traffic. The adapter provides two 10 Gb twinax copperl ports and two 1 Gb RJ45 ports in a PCIe 8x short form adapter.

The 10Gb ports are SFP+ and do not include a transceiver. Active Copper twinax cables up to 5 metre in length are supported such as provided by feature #EN01, #EN02 or #EN03. A transceiver is included with these cables. Note that SFP+ twinax copper is NOT AS/400 5250 twinax or CX4 or 10 GBase-T. Active cables differ from passive cables.

For the 1Gb RJ45 ports, 4-pair CAT-5 Unshielded Twisted Pair (UTP) cable or higher is supported for distances of up to 100 meters. In addition to 1Gb (1000 MB) networks, 100 MB networks are also supported, but 10Mb networks are not supported.

#EN0U and #EN0V adapters are electronically identical. They are physically identical except #EN0U has a tail stock for full high PCIe slots and #EN0V has a tail stock for low profile slots. The CCIN is 2CC3 for both features.

Details for the ports include:

  • LINUX NETWORK INSTALL are supported.
  • IEEE 802.3ae (10 GbE),IEEE 802.3ab (1000BASE-T GbE), 100BASE-T IEEEu, 802.1p priority, 802.1Q VLAN tagging, 802.3x flow control, 802.3ad load-balancing and failover,
  • Link aggregation, IEEE 802.3ad 802.3
  • Message Signalling Interrupt MSI-X, MSI and support of legacy pin interrupts
  • Ether II and IEEE 802.3 encapsulated frames
  • Jumbo frames up to 9.6 Kbytes
  • TCP checksum offload for IPv4 and IPv6
  • TCP segmentation Offload (TSO) for IPv4 and IPv6
  • UDP checksum offload for IPv4 and IPv6
  • Linux provide software iSCSI support through the adapter.
  • Attributes provided: Four ports - two 10Gb and two 1Gb E
  • Attributes required: PCIe Gen2 or Gen3 slot - low profile
  • Minimum required: 0
  • Maximum allowed: 4 (Initial order maximum: 4)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EP0Q) - 18-core 3.15 GHz (3.45 GHz Turbo) POWER9
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: N/A
  • Return parts MES: No
(#EP0S) - 22-core 2.80 GHz (3.10 GHz Turbo) POWER9
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: N/A
  • Return parts MES: No
(#EP72) - 18-core 3.15 GHz (3.45 GHz Turbo) POWER9 Processor
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
    • Red Hat Enterprise Linux 8 for POWER, or later, with all available maintenance updates
    • Red Hat Enterprise Linux 7 for POWER9, version 7.6, or later, with all available maintenance updates
    • Ubuntu Server 18.04.1, with all available maintenance updates
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EP74) - 22-core 2.80 GHz (3.10 GHz Turbo) POWER9 Processor
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
    • Red Hat Enterprise Linux 8 for POWER, or later, with all available maintenance updates
    • Red Hat Enterprise Linux 7 for POWER9, version 7.6, or later, with all available maintenance updates
    • Ubuntu Server 18.04.1, with all available maintenance updates
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EPAJ) - Power Cord 2m (6.56-foot), Drawer to IBM PDU, 250V/16A
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EPAL) - Power Cord 2.8m (9.18-foot), Drawer to IBM PDU, 250V/16A
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#EPAM) - Power Cord 4.3m (14.10-foot), Drawer to IBM PDU, 250V/16A
  • Minimum required: 0
  • Maximum allowed: 9999 (Initial order maximum: 250)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ERB0) - Bulk Packaging Request ID

This indicator is used to optimize package density of 2U server shipments through the use of bulk packaging.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Bulk packaging from the IBM factory to the customer.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB1) - Bulk Packaging ID #1

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #1.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB2) - Bulk Packaging ID #2

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #2.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB3) - Bulk Packaging ID #3

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #3.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB4) - Bulk Packaging ID #4

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #4.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB5) - Bulk Packaging ID #5

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #5.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB6) - Bulk Packaging ID #6

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #6.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB7) - Bulk Packaging ID #7

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #7.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB8) - Bulk Packaging ID #8

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #8.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERB9) - Bulk Packaging ID #9

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #9.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBA) - Bulk Packaging ID #10

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #10.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBB) - Bulk Packaging ID #11

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #11.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBC) - Bulk Packaging ID #12

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #12.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBD) - Bulk Packaging ID #13

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #13.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBE) - Bulk Packaging ID #14

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #14.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBF) - Bulk Packaging ID #15

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #15.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBG) - Bulk Packaging ID #16

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #16.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBH) - Bulk Packaging ID #17

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #17.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBJ) - Bulk Packaging ID #18

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #18.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBK) - Bulk Packaging ID #19

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #19.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBL) - Bulk Packaging ID #20

This indicator is used to specify that servers will be bulk packaged for shipment in package identifier #20.
Note: Bulk packaging would be defaulted via a specify code on the order for any customer order that contains 2 or more servers.

Limitations: There will be no mixing of MTMs with in the bulk packaging.

  • Attributes provided: Multiple servers in 1 package.
  • Attributes required: Order that contains 2 or more servers.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERBZ) - No Bulk Packaging Specify

This indicator is used to indicate single packaging will be used and not allow for bulk packaging.

  • Attributes provided: Single packages
  • Attributes required: Single customer orders.
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None.
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ERF1) - RFID Tags for Servers, Compute Nodes, Chassis, Racks, and HMCs

(No Longer Available as of September 30, 2021)

The specify feature attaches one Radio Frequency Identification Device (RFID) tag to a Server CEC, rack, HMC, compute node, or chassis enclosure. This can be used with MTM (machine type model) rack such as a 7953-94X or 7014-T42, not a feature code rack such as a #0553. It applies to newly shipped MTM servers, racks, HMCs, compute nodes, and chassis enclosures, not MES orders with one exception. POWER5 CECs being upgraded to a POWER6 CEC or POWER6 CECs being upgraded to a POWER 7 CEC can order this feature. The RFID tag meets the Financial Services Technology Consortium (FSTC) specifications for IT Data Center Asset Tracking.

  • Attributes provided: RFIDs
  • Attributes required: Server CEC, Compute Node, Chassis, MTM Rack, or HMC
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#ES6A) - 2 TB 7.2k RPM 5xx SATA SFF-4 Disk Drive

2 TB 7200 rpm SATA 2.5-inch (SFF) disk drive (HDD) formatted in 512 byte sectors (5xx) mounted on a 2.5-inch SFF-4 carrier/tray. Drive is rated at 6Gb/s though actual performance is dependent upon multiple factors.

Limitations: Drive does not fit in 2.5-inch (SFF-1, SFF-2, SFF-3) bays due to physical differences in carriers/trays. Drive does not support the SAS interfaces. 5xx drives can not be reformatted to 4k byte sectors (or vice versa).

  • Attributes provided: 2 TB of disk storage
  • Attributes required: One SFF-4 bay
  • Minimum required: 0
  • Maximum allowed: 2 (Initial order maximum: 2)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No
(#ESC0) - S&H - No Charge

No charge shipping and handling

  • Attributes provided: None
  • Attributes required: Sales Preapproval Required
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: Does not apply
(#ESC5) - S&H-a

Shipping and handling

  • Attributes provided: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required: None
  • Initial Order/MES/Both/Supported: Initial
  • CSU: N/A
  • Return parts MES: Does not apply
(#EUA5) - Stand-alone USB DVD drive w/cable

The Stand-alone USB DVD drive (FC EUA5) is an optional, stand-alone external USB-DVD device. It requires high current at 5V and must use the front USB 3.0 port on the 9008-22L, 9009-22A, 9009-41A, 9009-42A, 9223-22H, and 9223-42H systems.

  • Custom card identification number (CCIN): 6331 model 005
  • Media: Reads CD-ROM, CD-R, CD-RW, DVD-R, DVD+R, DVD-RW, DVD+RW, DVD-ROM, and DVD-RAM discs; Writes 4.7 GB DVD-RAM discs; CD read speed at 24X; DVD-ROM read speed at 8X; DVD-RAM at 5X; DVD-RAM has a write speed of 5X; The buffer size is 0.75 MB and cannot be disabled.
  • Interface: USB
  • Connector: USB 2.0
  • Loading tray: supports 12 cm and 8 cm discs floor) Note: A USB extension cable is included (P/N 32N1311). The USB extension cable is to be used when there are no safe, flat spots available in the rack. This cable allows the drive to reach the floor. Alternate or additional extension cables are not supported as the total USB cable length can be no longer than 3 meters.
  • Form factor: stand-alone USB DVD drive
  • DVD video: not supported
  • Attributes provided: USB DVD drive
  • Attributes required: None
  • Minimum required: 0
  • Maximum allowed: 1 (Initial order maximum: 1)
  • OS level required:
  • Initial Order/MES/Both/Supported: Both
  • CSU: Yes
  • Return parts MES: No


Back to topBack to top

Accessories

Top rule

None.

Customer replacement parts

The following parts and feature(s) have been designated as Tier 1 CRUs:

  • Hard disk drive
  • Optical drive
  • Fan
  • Fan Cage
  • PCI adapters
  • RAID card cable
  • Memory DIMMs
  • Native USB Serial Card
  • I/O adapters
  • Operator panel
  • Operator panel cable
  • Power cord
  • Power supply
  • Processor power regulator
  • Service processor
  • Keyboard
  • Mouse
  • External cables
  • GPU Power Cable
  • GPU Adapter
  • Display
  • Sheet metal trays and dividers
  • Planar
  • Air duct
  • Processors
  • Heat sinks
  • Rails


Back to topBack to top

Supplies

Top rule

None.

Supplemental media

None.



Back to topBack to top

Trademarks

Top rule

(R), (TM), * Trade-mark or registered trade-mark of International Business Machines Corporation.

** Company, product, or service name may be a trade-mark or service mark of others.

Terms of use

IBM products and services which are announced and available in your country can be ordered under the applicable standard agreements, terms, conditions, and prices in effect at the time. IBM reserves the right to modify or withdraw this Sales Manual at any time without notice. This Sales Manual is provided for your information only. Additional terms of use are located at

Terms of use


© IBM Corporation 2021.