SPEChpc®2002
The SPEChpc® 2002 benchmark is derived
from real HPC applications and application practices, and measure the
overall performance of high-end computer systems, including the computer's
processors (CPUs), the interconnection system (shared or distributed
memory), the compilers, the MPI and/or OpenMP parallel library implementation,
and the input/output system.
Benchmark retirement - Benchmark results are no longer being accepted
for publication on SPEC's site. Sales of the software were discontinued
in February 2007, and technical support is no longer provided.
[ Results | Information | HPG | Future | Join ]
The SPEChpc®2002 Benchmarks
The benchmarks which make up the suite are:
-
The SPECCHEM® 2002 benchmark, which is based
on a quantum chemistry application called GAMESS and includes the SPECchemM 2002 performance metric and the
SPECchemS 2002 performance metrric.
-
-
SPECENV® 2002 benchmark, which is based on
a weather research and forecasting model called WRF. It includes the SPECenvM 2002 performance metric and the SPECenvS 2002 performance metric, one for each dataset size.
-
-
The SPECSEIS® 2002 benchmark, which represents
an industrial application that performs time and depth migrations used to locate gas
and oil deposits and includes the SPECseisM 2002 performance metric and the SPECseisS 2002 performance metric.
Each of the three benchmarks has both Small and Medium workloads available.
Benchmark Features:
-
Code is selected from full HPC applications
-
Parallelism supported: serial, OpenMP, MPI, or combined MPI-OpenMP
-
Architectures supported: shared memory, distributed memory, clusters
-
IO and communication included in the benchmarks
-
Run rules tailored to HPC application environments
-
Implemented under SPEC tools for building applications, running with
different data sets, verifying results and measuring runtime.
Results
-
Submitted Results
-
Text, HTML, PDF, PostScript file outputs for the SPEChpc® 2002
benchmark metrics; includes all of the results submitted to SPEC
from the SPEC member companies and other licensees of the benchmark.
-
Flag and Source Disclosures
-
Explanations from the testers elaborating on flags in the
results' notes sections, and any source code modifications provided
for peak performance results.
-
Approved Alternate Source Codes
-
Some of the results submitted to SPEC and published on www.spec.org use
alternate benchmark source code and/or third-party alternate software
approved by SPEC. Licensees of the HPC benchmark suite may obtain the
alternate source codes by sending an email request to info@spec.org
Information
-
Press Release
-
Benchmark announcement day press release materials:
-
Benchmark Documentation
-
Support and technical documents
-
-
Issues Building/Running the Benchmark
-
Installation, build, and runtime issues raised by users of the benchmark.
-
-
