Server Gurus Discussions

[EH]

Hello,

I hope somebody can help me with the following issue:

We have bought a new workstation for FEA simulations (vehicle crash etc.). Unfortunately we are not getting the expected performance.

Our System:

• AMD EPYC 7542 32-Core Processor
• Sockets = 2
• 64 physical cores (128 Threads)
• 265 GB Ram (8 x 32 GB DDR4 3200) -> so on every second slot
• OS Win 10

Accroding to the following Benchmark for an older CPU the simulation time should be under 2000 seconds:

https://www.amd.com/system/files/documents/amd-epyc-with-altair-radioss-powering-hpc.pdf

With our current configuration the same simulation model (Neon1M11) needs 10000 seconds (5 times longer).

For the calculation we are using Intel MPI (-mpi -i -np 64). From altair we got following recommendations for the envoirnment variables which already helped that the programm actually using all cores:

KMP_AFFINITY=disabled

I_MPI_DOMAIN=auto

And we turned off hyperthreading in the BIOS.

But still the simulations are 5 times slower tahn expected. 

Other Benchmark like Cinebench are performing as expected. 

Are there any further settings we have missed?

Thanks in advance

Best regards 

Edward

[EPYCperformance]

Sorry to hear and not sure if you found a resolution or not, but our internal testing on EPYC 7542 using Radioss 2018 neon benchmark completed in 175 seconds. However, this performance was measured on RHEL 7.7 (3.10.0-1062.el7.x86_64) running Altair RADIOSS™ 2018 neon with a single-node, 2x AMD EPYC™ 7F52 (16C) processor-based System, 16x Micron™ 64GB DDR4-3200 DR 1DPC, Mellanox™ CX-6 HDR 200 Gb/s IB x16 PCIe® Gen 4, 1x Micron 1100 256 GB SATA (OS), 1x1 TB NVMe (Data), BIOS settings: Defaults, plus NPS=NPS4 (7F52), SMT = Off, Boost = On, APBDIS=1, Fixed SOC P state=P0, DLWM=off, X2APIC = On, Determinism Slider = Performance, Preferred IO=Enabled. 

The Workload Tuning Guide has some guidance https://developer.amd.com/wp-content/resources/56745_0.80.pdf along with the Windows Tuning Guide (although targeted towards Server versions - I am not sure if it all applies to Win10) https://developer.amd.com/wp-content/resources/56782_1.0.pdf. 

[kevin2023]

https://forum.ansys.com/forums/topic/fluent-2022-r1-slow-bandwitdth/

Please see this link. I also found windows's speed is slower than linux. and the link in this post does not exist.