mirror of
https://github.com/microsoft/mscclpp.git
synced 2026-05-13 09:46:00 +00:00
20eca28942
* Add a FIFO test code that reproduced a correctness issue * Fix the correctness issue by using pinned memory instead of cudaMemcpy --------- Co-authored-by: Binyang Li <binyli@microsoft.com>
20 lines
2.8 KiB
JSON
20 lines
2.8 KiB
JSON
{"name":"allgather", "kernel":1, "ranks":8, "ranksPerNode":8, "algBw":291.52, "busBw":255.08, "size":1073741824, "time":3683.13, "target":"throughput"}
|
|
{"name":"allgather", "kernel":2, "ranks":16,"ranksPerNode":8, "algBw":244.61, "busBw":229.33, "size":3221225472, "time":13168.31,"target":"throughput"}
|
|
{"name":"allgather", "kernel":3, "ranks":8, "ranksPerNode":8, "algBw":0.1012, "busBw":0.090, "size":8192, "time":83.41, "target":"latency"}
|
|
{"name":"allreduce", "kernel":1, "ranks":8, "ranksPerNode":8, "algBw":139.41, "busBw":243.96, "size":1073741824, "time":7701.98, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":2, "ranks":8, "ranksPerNode":8, "algBw":1.25, "busBw":2.19, "size":8192, "time":6.51, "target":"latency"}
|
|
{"name":"allreduce", "kernel":2, "ranks":16,"ranksPerNode":8, "algBw":0.44, "busBw":0.83, "size":8192, "time":18.42, "target":"latency"}
|
|
{"name":"allreduce", "kernel":3, "ranks":8, "ranksPerNode":8, "algBw":139.08, "busBw":243.40, "size":1073741824, "time":7719.85, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":4, "ranks":8, "ranksPerNode":8, "algBw":106.98, "busBw":187.22, "size":16777216, "time":156.81, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":4, "ranks":8, "ranksPerNode":8, "algBw":116.24, "busBw":203.42, "size":33554432, "time":288.65, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":5, "ranks":8, "ranksPerNode":8, "algBw":126.52,"busBw":221.418,"size":50331648, "time":397.79, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":6, "ranks":8, "ranksPerNode":8, "algBw":3.3919,"busBw":5.9359, "size":24576, "time":7.24, "target":"latency"}
|
|
{"name":"allreduce", "kernel":6, "ranks":8, "ranksPerNode":8, "algBw":6.21, "busBw":10.87, "size":49152, "time":7.91, "target":"latency"}
|
|
{"name":"allreduce", "kernel":6, "ranks":8, "ranksPerNode":8, "algBw":8.90, "busBw":15.57, "size":73728, "time":8.28, "target":"latency"}
|
|
{"name":"allreduce", "kernel":4, "ranks":16,"ranksPerNode":8, "algBw":80.84, "busBw":151.58, "size":25165824, "time":311.28, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":4, "ranks":16,"ranksPerNode":8, "algBw":97.27, "busBw":182.38, "size":50331648, "time":517.43, "target":"throughput"}
|
|
{"name":"allreduce", "kernel":4, "ranks":16,"ranksPerNode":8, "algBw":125.99, "busBw":236.24, "size":3221225472, "time":25565.46,"target":"throughput"}
|
|
{"name":"allreduce", "kernel":3, "ranks":16,"ranksPerNode":8, "algBw":119.5, "busBw":224.06, "size":3221225472, "time":26955.85,"target":"throughput"}
|
|
{"name":"alltoall", "kernel":0, "ranks":16,"ranksPerNode":8, "algBw":46.53, "busBw":43.63, "size":1073741824, "time":23071.5, "target":"throughput"}
|
|
{"name":"alltoall", "kernel":1, "ranks":8, "ranksPerNode":8, "algBw":276.17, "busBw":241.65, "size":1073741824, "time":3887.87, "target":"throughput"}
|