Maxwell : Pelegant benchmark

This is simple test run for Pelegant. All files can be found on Maxwell under /beegfs/desy/group/it/Benchmarks/Pelegant/maxwell_bench_1:

• submit.sh: sample to submit single node jobs for different CPUs
• benchmark.pelegant.sh: Pelegant sample script
• da.ele.orig, p4_cell.lte: original input data. da.ele.mod: modified input data to extend the runtime

Running the original simulation on a single node

Note: the runtime of a little more than 3 seconds is too short to provide very good estimates. However, Intel Xeon Gold 6240 is clearly the fastest, followed by AMD EPYC 7402. In general, Pelegant performs best using only physical cores (i.e. half of available cores).

Running an extended simulation on a single node

To get longer runtimes and more solid estimates, simulations were repeated increasing n_steps1→ 4, n_passes 256 → 512 (see da.ele files).

Comparison of fastest execution (shorter is better)	Scaling behavior
	Relative speed compared to AMD EPYC at 40 cores

Note: in contrast to the original run, it appears in most cases favorable to run on all cores. AMD EPYC 7402 is clearly fastest, followed by Intel Xeon Gold 6240 and - surprisingly - Intel Xeon E5-2698 V4.