2H 2015
intel-xeon-phi-sw-ecosystem-guide-2h-2015-public3
intel-xeon-phi-sw-ecosystem-guide-2h-2015-public3
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Comparative Performance<br />
NAMD* 2.10 Pre-Release<br />
STMV<br />
CLUSTER BENCHMARK<br />
32 NODES<br />
APPROVED FOR PUBLIC PRESENTATION<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
1<br />
NAMD* 2.10 (pre-release) Cluster Performance Increase<br />
STMV (~1M atoms)<br />
1.2X<br />
2X<br />
2.1X<br />
6.8X<br />
7.9X<br />
12.2X 13.1X<br />
1 Node 8 Nodes 32 Nodes<br />
Intel® Xeon® processor E5-2697 v2 (Baseline: 1 node, 23 or 47 PPN)<br />
Intel® Xeon® processor E5-2697 v3 (27 or 55 PPN)<br />
Xeon E5-2697 v2 (23 or 47 PPN) + 1 Intel® Xeon Phi coprocessor 7120A (240T)<br />
Xeon E5-2697 v3 (27 or 55 PPN) + 1 Intel® Xeon Phi coprocessor 7110A (240T)<br />
20X<br />
24.2X<br />
“Xeon E5-2697 v2/v3” = Intel® Xeon® processor E5-2697 v2/v3<br />
27.2X<br />
32X<br />
Application: NAMD 2.10 pre-release; STMV<br />
Description: A parallel, object-oriented molecular<br />
dynamics code designed for high-performance<br />
simulation of large biomolecular systems. More at<br />
http://www.ks.uiuc.edu/Research/namd/<br />
Availability:<br />
• Code: Intel® Xeon Phi coprocessor support is<br />
available as a pre-release. Use the nightly build.<br />
• Recipe: Available here.<br />
Usage Model: Single rank on host with 47 threads.<br />
Various computations are offloaded to Intel® Xeon<br />
Phi coprocessor from each thread.<br />
Highlights: Intel® Xeon Phi coprocessor support is<br />
now in the development branch of NAMD 2.10 prerelease.<br />
Results: For the STMV workload, the Intel® Xeon®<br />
processor E5-2697 v3 and the Intel® Xeon Phi<br />
coprocessor (32 nodes, 55 PPN) improved<br />
performance by up to 32X compared to the baseline<br />
processor (1 node, 47 PPN).<br />
For configuration details, go here.<br />
SOURCE: INTEL MEASURED RESULTS AS OF SEPTEMBER, 2014<br />
Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems,<br />
components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated<br />
purchases, including the performance of that product when combined with other products. For more information go to http://www.intel.com/performance *Other names and brands may be claimed as the property of others<br />
20
Change language