IBM®
Skip to main content
    Country/region [select]      Terms of use
 
 
    
     Home      Products      Services & solutions      Support & downloads      My account     
 
developerworks > My developerWorks >  Dashboard > Linux for Power Architecture > ... > Performance > Performance Insights
developerWorks
Log In   View a printable version of the current page.
Overview Connect Spaces Forums Wikis
Performance Insights
Added by wburos, last edited by billburos on Apr 14, 2009  (view change)
Labels: 
(None)

Performance Insights

This page captures and highlights insights into performance for Linux running on Power systems.


Other Linux on Power performance pages include:



64KB pages with Linux

The Linux kernel is built with 64KB pages as the base page size on Power systems with RHEL 5 and now SLES 11. Older versions of the distros use 4KB pages as the base page size. This option is a kernel build-time option and is not intended to be changed.



Stealing CPU cycles

A common head-scratcher on Linux is the new concept of CPU metric tools reporting that "something" is stealing CPU cycles from schedule'able CPUs. For an initial explanation on that process, see:

Follow-on work is being considered in this space since there are a lot of subtleties of how to measure CPU cycles.



Using IBM's XLC/XLF compilers with openmpi

A common myth is that you need to re-build the openmpi libraries that come with SLES 10 and RHEL 5 in order to use the IBM compilers.



Leveraging OpenMPI in a small p6 575 cluster

Looking for different ways of compiling and linking to generate executable code for Linpack, this article provide methods to transparently back bss/data/text segments and dynamic memory with 16MB pages via Libhugetlbfs. OpenMPI options are provided to bind MPI processes to available processors and make use of the InfiniBand interconnect. The scalability of a cluster having four IBM Power 575 systems is used as a simple example.



An example of using OpenMPI with SPECmpi2007

Here's an example of how to compile and execute the SPEC MPI2007 benchmark using IBM's eHCA InfiniBand interconnect, IBM XL compilers for C/C++ and Fortran, and OpenMPI on IBM POWER6 systems. This example utilizes one shared file system node and two execution nodes.

These steps can be applied to any number of systems to extend the cluster across more execution nodes. In this example, OpenMPI and InfiniBand are used across the execution nodes, and NFS is used to access the shared file system node.



An Assessment of Leadership Performance with POWER6 Processors and Red Hat Enterprise Linux 5.1

This paper highlights the performance on IBM's POWER6 processor-based systems running with the Red Hat Enterprise Linux (RHEL) 5.1, based on published SPEC CPU2006 and SPECjbb2005 results, including single-system LINPACK metrics, on System p 570 4-core, 8-core and 16-core systems. With simultaneous multithreading (SMT) support turned on, Linux is easily able to provide effective scheduling support of the 32 processor threads seen on the 16-core 570 system.



Send questions to http://www.ibm.com/developerworks/forums/forum.jspa?forumID=375


 
    About IBM Privacy Contact