Benchmark Submit Results Form
Solver Information
|
Application: *
|
|
|
Version: *
|
|
Benchmark
| Model (Name): *
|
|
Platform Configuration
|
Please describe the configuration of the computer system on which you ran these benchmarks by filling in the fields below. Please submit separate forms for each unique system configuration.
|
|
Hardware (Manufacturer): *
|
|
|
System (Name): *
|
|
|
Processor (Type): *
|
|
| Please cite the processor name (code names are acceptable for pre-released processors)
and also the operating frequency for the processor |
|
CPU-cores per (processor): *
|
|
|
Processors per Node: *
|
|
|
Memory (GB/CPU-core): * |
|
|
Interconnect: *
|
|
|
Please first identify the technology class (IB=Infiniband, GiGE=Goigabit Ethernet, 10GiGE=10GB Ethernet, if proprietary list vendor's name), then you may also list the interconnect supplier (optional entry, examples: Mellenox, Voltaire, Cisco) and the model number of the interconnect switch (optional) and finally please list the software protocol (optional) running on the interconnect (examples: TCP, IPoIB & upapl)
|
|
Operating System: *
|
|
|
(Vendor: OS name_revision level)
Please use the following naming conventions- MSFT for Microsoft, RH for Red Hat, Novell. Please also provide the name of the operating systems and the version level.
|
|
MPI library: *
|
|
| (name_version level)
Please identify the name and version level of the MPI libriries utilized for these benchmarks
|
|
Comment: *
|
|
Benchmark Performance Results
|
Note: Please insert the wall clock times (total elapsed times) for all individual benchmark
jobs run across a range of CPU-core job-widths. Please insert integers in row 4 measured in seconds.
Hardware partners and customers who submit their results to be published on Altair’s website must run
at least 3 different benchmark jobs for each input model for each hardware platform.
Each specific benchmark job should deploy a different number of CPU-cores.
Altair will accept benchmark job runs across: 1, 2, 4, 8, 16, 32, 64, 128, 256 & 512 CPU-cores.
Altair will publish two dimension plots (scalability charts) on our website which illustrate the wall
clock time required to complete each benchmark (vertical axis) against the quantity of CPU-cores
deployed for each benchmark job. We need at least 3 data points to draw a reasonable curve.
Hardware partners are encouraged to run additional benchmarks jobs (more than the minimum of 3) to
demonstrate the exceptional scalability of Radioss on their hardware platforms.
|
| Summary of Benchmark Runs |
|
|
|
|