Provision NVIDIA GPUs for Generative AI, Traditional AI, HPC and Visualization use cases on the trusted, secure and cost-effective IBM Cloud infrastructure.
Train, validate, tune and deploy Gen AI models for a variety of use cases like customer service, digital labor and application modernization.
Use AI and ML algorithms to forecast and make predictions, based on specific user needs and business requirements.
Improved networking and memory of GPUs to speed up compute intensive applications for Financial, EDA and Life Science workloads.
Use GPUs to support engineering simulation, product design workstations, gaming, 3D modelling, AR/VR.
Long-running job on high end GPU infrastructure
10-500+ high-end GPUs (per job)
Recommended GPUs: H100, L40S
Model tuning with custom data set for downstream tasks
1+ mid high-end GPU (per job)
Recommended GPUs: L40S
Sensitivity to latency & throughput, always cost-sensitive
Single or fraction of GPU per task
Recommended GPUs: L40S, L4
GPU
VCPUs
RAM
Bandwidth
1 x NVIDIA L4 24 GB
16
80 GB
32 Gbps
2 x NVIDIA L4 24 GB
32
160 GB
64 Gbps
4 x NVIDIA L4 24 GB
64
320 GB
128 Gbps
1 x NVIDIA L40S 48 GB
24
120 GB
48 Gbps
2 x NVIDIA L40S 48 GB
48
240 GB
96 Gbps
1 x NVIDIA Tesla V100 16 GB
8
64 GB
16 Gbps
1 x NVIDIA Tesla V100 16 GB
16
128 GB
32 Gbps
2 x NVIDIA Tesla V100 16 GB
16
128 GB
32 Gbps
2 x NVIDIA Tesla V100 16 GB
32
256 GB
64 Gbps
H100 GPUs coming soon
GPU options on Red Hat OpenShift* |
---|
Profile name
GPU
VCPUs
RAM
Bandwidth
gx3.16x80.l4
1 x NVIDIA L4 24 GB
16
80 GB
32 Gbps
gx3.32x160.2l4
2 x NVIDIA L4 24 GB
32
160 GB
32 Gbps
gx3.64x320.4l4
4 x NVIDIA L4 24 GB
64
320 GB
32 Gbps
GPU options on Virtual Server for Classic* |
---|
GPU
VCPUs
RAM
1 x P100
8
60 GB
2 x P100
16
120 GB
GPU options on Bare Metal Server for Classic* |
---|
GPU
CPU
Cores
RAM
NVIDIA T4
Intel Xeon 8474
96
Up to 2048 GB
NVIDIA T4
Intel Xeon 8474
48
Up to 2048 GB
NVIDIA T4
Intel Xeon 8474
64
Up to 2048 GB
Unlock productivity with out of the box AI assistants for customer service, digital labor and application modernization.
Build, scale and govern AI across your business —all on one platform.
Breakthrough data barriers and accelerate AI with global data platform.
Deploy AI models anywhere with hybrid cloud platform.
Cloud: GPUs on IBM Cloud, On-premise: IBM Storage Fusion.
IBM and NASA collaborate on climate change with AI models.
IBM watsonx increased help desk ticket resolutions by 40% per day.
Blendow Group collaborated with IBM ,reduced the time required for discovery and analysis of legal documents by 70%.
20 TB bandwidth included in US, Canada and EU data centers; 5 TB bandwidth included in all other data centers. New prices and offers may not be combined with any other current or future discounts.