2×Intel xeon ICX 4316 Vs Epyc 7513 server for quantum simulation

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

Niharendu Barman

New Member
Sep 23, 2021
2
0
1
I am planning to buy a tower server for material simulation code like VASP, Quantum Espresso. These code use intel MKL library intensively. Recent milan processors are faster in every benchmark in openbenchmark sites. But in case of optimized library, i have found in various posts that intel xeon gets benifits from their mkl library, so the overall performance is much higher in xeon. MKL although runs on amd platform, but after mkl version 2020 amd processors performance goes down due to crippling by intel. My questions is
1. Considering all these facts what would be the best processor ( between two intel xeon 4316 and one epyc 7513) for running numerical computational codes like VASP and Quantum Espresso?
 
Last edited:

NateS

Active Member
Apr 19, 2021
159
91
28
Sacramento, CA, US
Have you considered renting time on one of each type of system on a public cloud and trying it out? Because for something like this, which one will be faster will probably depend heavily on the details of your workload. You're right that the optimized math will run faster on the Intel and everything else will run faster on AMD, so which one is faster overall depends on how much time is spent doing optimized math vs everything else in your particular use case. It's probably not something anyone else can answer with high confidence.