Specifications

April 2012 v1 AMD Opteron™ 6200 Linux Tuning Guide
19
3.6 Libraries
Special purpose high-performance libraries are another important contributor to application performance. A
classic example of this is when a program makes extensive use of BLAS (Basic Linear Algebra Subroutines) and
LAPACK (Linear Algebra PACKage) subroutines. In these cases, it is imperative to use an FMA4-enabled floating
point library. The AMD Core Math Library has been tuned to use FMA4 instructions, where possible, and will
provide a significant performance boost when compared to other libraries that have not been tuned for AMD
Opteron™ 4200/6200 Series processors.
• ACML (AMD Core Math Library) provides a free set of thoroughly optimized and threaded math routines
for HPC, scientific, engineering, and related compute-intensive applications.
• ACML consists of the following main components:
- A full implementation of Level 1, 2, and 3 BLAS with key routines optimized for high performance on AMD
Opteron™ processors.
- A full suite of LAPACK routines. As well as taking advantage of the highly tuned BLAS kernels, a key set of
LAPACK routines has been further optimized to achieve considerably higher performance than standard
LAPACK implementations.
- A comprehensive suite of Fast Fourier Transforms (FFTs) in both single-, double-, single-complex, and
double-complex data types.
- Random Number Generators in both single- and double-precision.
- For complete ACML V5.1.0 or later information, data sheet, performance libraries, and download, see
http://developer.amd.com/libraries/acml/pages/default.aspx.
3.7 Other Libraries and Tools
Other modules and tools that may assist with building GCC and OpenMPI include:
• OpenMPI: http://www.open-mpi.org/
• Build the OpenMPI to include:
- Portable Hardware Locality (hwloc): http://www.open-mpi.org/projects/hwloc/
- KNEM module (Core - Core on-board memory single-copy optimization using this KNEM that builds a
kernel module:
http://runtime.bordeaux.inria.fr/knem/
http://runtime.bordeaux.inria.fr/knem/doc/
• libnuma-develop (Module with headers and libraries for developing programs using NUMA policies)
- http://oss.sgi.com/projects/libnuma/
- http://rpm.pbone.net/index.php3?stat=3&search=libnuma-devel&srodzaj=3
- http://rpmfind.net/linux/rpm2html/search.php?query=libnuma-devel
• GMP (GNU Multiple Precision Arithmetic Library)
- http://gmplib.org/
• LIKWID (Lightweight Performance Tools)
- http://code.google.com/p/likwid/
• Linux Performance Tools (Linux profiling with performance counters)
- https://perf.wiki.kernel.org/
• AMD CodeAnalyst (Performance Analyzer for Linux)
- http://developer.amd.com/tools/CodeAnalyst/codeanalystlinux/Pages/default.aspx