[Reserved for scores]
| CPU | RAM | RAM Channels | Operating System | CFD-only | CFD-EM | EM-only | Notes |
|---|---|---|---|---|---|---|---|
| Dual Intel Xeon 9480 | 128GB | 8192-bit wide HBM2E @ 3200MHz | openSUSE Leap 15.6 | 16m 19s | 2h 17m 40s | � | ran with numactl -C 0-55 -m 0-7 |
| Dual EPYC 9374F | 1536GB | 24 (1536-bit) DDR5 | Ubuntu | 13m 20s | 2h 47m 50s | � | |
| Intel Xeon W5-3495X | � | 8 (512-bit) DDR5 | Ubuntu | � | 2h 56m 21s | � | |
| AMD 9970X | 192GB | 4 (256-bit) DDR5 @ 7200MHz | Windows 11 | 18m 31s | 3h 4m 6s | � | run with numasets -4; ~500 watt power usage |
| Intel Xeon W5-3435X | 512GB | 8 (512-bit) DDR5 @ 5880MHz | Centos Stream 9 | 23m 12s | 3h 11m 52s | 16m 38s | ~500 watt power usage |
| Intel Xeon W5-3495X | 512GB | 8 (512-bit) DDR5 @ 4800MHz | Linux | 19m 23s | 3h 12m 1s | 20m 30s | ~375 watt peak power usage |
| Xeon 658x | 128GB | 512-bit wide DDR5 | � | 20m 56s | 3h 27m 31s | � | There might have been a memory overflow on CFD-EM bench |
| Apple M1 Ultra | 64GB | 1024-bit wide LPDDR5-6400 | MacOS 14.4.1 | 31m 33s | 3h 36m 10s | � | FGMRES run on CFD-EM using 16 p-cores; ~112 watt power usage |
| Threadripper 7960X | 128GB | 4 (256-bit) DDR5 @ 6600MHz | Windows 11 | 20m 34s | 3h 50m 6s | � | -numasets 4 |
| Ryzen 7950X3D | 64GB | 2 (128-bit) @ 6200MHz | Windows 11 | 28m 18s | 5h 13m 36s | � | -numasets 2 |
| Apple M1 Max | 64GB | 512-bit wide LPDDR5-6400 | � | � | 5h 27m 34s | � | FGMRES run on CFD-EM using 8 p-cores; ~30 watt power usage |
| Dual EPYC 9575F | 1536GB | 24 (1536-bit) DDR5 | � | 11m 57s | 5h 30m 32s | 7m 9s | Something is wrong with CFD-EM run |
| Ryzen 7950X | 96GB | 2 (128-bit) @ 6200MHz | fedora 39 | 32m 25s | 5h 52m 56s | � | |
| Ryzen 9950X | 96GB | 2 (128-bit) @ 6400MHz | Arch Linux | 25m 40s | � | � | Memory Tuned + PBO |
| Ryzen AI Max+ 395 | 128GB | 256-bit wide LPDDR5x-8000 | Windows 11 Pro 24H2 | 35m 56s | � | � | -np 16 -blas aocl; 45 watt steady state power |