X

Everything you want and need to know about Oracle SPARC systems performance

SPARC T7-4 Delivers 4-Chip World Record for SPEC OMP2012

Brian Whitney
Principal Software Engineer

Oracle's SPARC T7-4 server delivered world record performance on the SPEC OMP2012 benchmark for systems with four chips.  This was accomplished with Oracle Solaris 11.3 and Oracle Solaris Studio 12.4 software.

  • The SPARC T7-4 server delivered world record for systems with four chips of 27.9 SPECompG_peak2012 and 26.4 SPECompG_base2012.

  • The SPARC T7-4 server beat the four chip HP ProLiant DL580 Gen9 with Intel Xeon Processor E7-8890 v3 by 29% on the SPECompG_peak2012 benchmark.

  • This SPEC OMP2012 benchmark result demonstrates that the SPARC M7 processor performs well on floating-point intensive technical computing and modeling workloads.

Performance Landscape

Complete benchmark results are at the SPEC website, SPEC OMP2012 Results.  The table below provides the new Oracle result as well as the previous best four chip results.

SPEC OMP2012 Results
Four Chip Results
System Processor Peak Base
SPARC T7-4 SPARC M7, 4.13 GHz 27.9 26.4
HP ProLiant DL580 Gen9 Intel Xeon E7-8890 v3, 2.5 GHz 21.5 20.4
Cisco UCS C460 M4 Intel Xeon E7-8890 v3, 2.5 GHz -- 20.8

Configuration Summary

Systems Under Test:

SPARC T7-4
4 x 4.13 GHz SPARC M7 processors
2 TB memory (64 x 32 GB dimms)
4 x 600 GB SAS 10,000 RPM HDD (mirrored)
Oracle Solaris 11.3 (11.3.0.30.0)
Oracle Solaris Studio 12.4 with 4/15 Patch Set

Benchmark Description

The following was taken from the SPEC website.

SPEC OMP2012 focuses on compute intensive performance, which means these benchmarks emphasize the performance of:

  • the computer processor (CPU),
  • the memory architecture,
  • the parallel support libraries, and
  • the compilers.

It is important to remember the contribution of the latter three components. SPEC OMP performance intentionally depends on more than just the processor.

SPEC OMP2012 contains a suite that focuses on parallel computing performance using the OpenMP parallelism standard.

The suite can be used to measure along the following vector:

  • Compilation method: Consistent compiler options across all programs of a given language (the base metrics) and, optionally, compiler options tuned to each program (the peak metrics).

SPEC OMP2012 is not intended to stress other computer components such as networking, the operating system, graphics, or the I/O system.  Note that there are many other SPEC benchmarks, including benchmarks that specifically focus on graphics, distributed Java computing, webservers, and network file systems.

Key Points and Best Practices

  • Jobs were bound using the OpenMP environment direction OMP_PLACES.

See Also

Disclosure Statement

SPEC and the benchmark name SPEC OMP are registered trademarks of the Standard Performance Evaluation Corporation.  Results as of November 11, 2015 from www.spec.org.  SPARC T7-4 (4 chips, 128 cores, 1024 threads): 27.9 SPECompG_peak2012, 26.4 SPECompG_base2012; HP ProLiant DL580 Gen9 (4 chips, 72 cores, 144 threads): 21.5 SPECompG_peak2012, 20.4 SPECompG_base2012; Cisco UCS C460 M7 (4 chips, 72 cores, 144 threads): 20.8 SPECompG_base2012.

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.