mirror of
https://github.com/paboyle/Grid.git
synced 2026-06-04 11:14:38 +01:00
42cd9eda71
and also some staggered hdcg type work
2.1 KiB
2.1 KiB
| 1 | Per node summary table |
|---|---|
| 2 | L , Wilson, DWF4, Staggered, NaiveStag |
| 3 | 8 , 90, 933, 38, 23 |
| 4 | 12 , 403, 1688, 178, 113 |
| 5 | 16 , 188, 1647, 449, 295 |
| 6 | 24 , 947, 1574, 674, 553 |
| 7 | 32 , 931, 1371, 718, 643 |
| 8 | Memory Bandwidth |
| 9 | Bytes, GB/s per node |
| 10 | 786432, 40.271620 |
| 11 | 12582912, 433.611792 |
| 12 | 63700992, 905.374321 |
| 13 | 201326592, 1114.979152 |
| 14 | 491520000, 1180.241898 |
| 15 | Communications |
| 16 | Packet bytes, direction, GB/s per node |
| 17 | GEMM |
| 18 | M, N, K, BATCH, GF/s per rank fp64 |
| 19 | 16, 8, 16, 4096, 693.316363 |
| 20 | 16, 12, 16, 4096, 657.277058 |
| 21 | 16, 16, 16, 4096, 711.992616 |
| 22 | 32, 8, 32, 4096, 821.084324 |
| 23 | 32, 12, 32, 4096, 1279.852719 |
| 24 | 32, 16, 32, 4096, 2647.096674 |
| 25 | 64, 8, 64, 4096, 2630.192325 |
| 26 | 64, 12, 64, 4096, 3338.071321 |
| 27 | 64, 16, 64, 4096, 3950.899281 |
| 28 | 16, 8, 256, 4096, 1638.362501 |
| 29 | 16, 12, 256, 4096, 2377.502234 |
| 30 | 16, 16, 256, 4096, 3048.328833 |
| 31 | 32, 8, 256, 4096, 2917.384276 |
| 32 | 32, 12, 256, 4096, 4103.085151 |
| 33 | 32, 16, 256, 4096, 5102.971860 |
| 34 | 64, 8, 256, 4096, 3222.258206 |
| 35 | 64, 12, 256, 4096, 4619.456391 |
| 36 | 64, 16, 256, 4096, 5847.916650 |
| 37 | 8, 256, 16, 4096, 1728.073337 |
| 38 | 12, 256, 16, 4096, 2356.653970 |
| 39 | 16, 256, 16, 4096, 2676.876038 |
| 40 | 8, 256, 32, 4096, 2611.531990 |
| 41 | 12, 256, 32, 4096, 3451.573106 |
| 42 | 16, 256, 32, 4096, 3966.915301 |
| 43 | 8, 256, 64, 4096, 3436.248737 |
| 44 | 12, 256, 64, 4096, 4539.497945 |
| 45 | 16, 256, 64, 4096, 5307.992323 |
| 46 | GEMM |
| 47 | M, N, K, BATCH, GF/s per rank fp32 |
| 48 | 16, 8, 16, 4096, 499.017445 |
| 49 | 16, 12, 16, 4096, 731.543385 |
| 50 | 16, 16, 16, 4096, 958.800786 |
| 51 | 32, 8, 32, 4096, 1549.813550 |
| 52 | 32, 12, 32, 4096, 2147.907502 |
| 53 | 32, 16, 32, 4096, 2601.698596 |
| 54 | 64, 8, 64, 4096, 3785.446233 |
| 55 | 64, 12, 64, 4096, 5116.694843 |
| 56 | 64, 16, 64, 4096, 6109.345016 |
| 57 | 16, 8, 256, 4096, 1206.627737 |
| 58 | 16, 12, 256, 4096, 1809.699599 |
| 59 | 16, 16, 256, 4096, 2412.014053 |
| 60 | 32, 8, 256, 4096, 2406.114488 |
| 61 | 32, 12, 256, 4096, 3605.531907 |
| 62 | 32, 16, 256, 4096, 4798.444037 |
| 63 | 64, 8, 256, 4096, 4688.711196 |
| 64 | 64, 12, 256, 4096, 6990.696301 |
| 65 | 64, 16, 256, 4096, 9214.749925 |
| 66 | 8, 256, 16, 4096, 2596.307289 |
| 67 | 12, 256, 16, 4096, 3439.892562 |
| 68 | 16, 256, 16, 4096, 3907.201036 |
| 69 | 8, 256, 32, 4096, 3012.752067 |
| 70 | 12, 256, 32, 4096, 3904.217583 |
| 71 | 16, 256, 32, 4096, 4599.047092 |
| 72 | 8, 256, 64, 4096, 3721.999042 |
| 73 | 12, 256, 64, 4096, 5098.573927 |
| 74 | 16, 256, 64, 4096, 6159.080872 |