mirror of
https://github.com/NVIDIA/nccl-tests.git
synced 2026-05-03 13:02:36 +00:00
Added explanation about performance numbers
This commit is contained in:
@@ -34,6 +34,10 @@ Run with MPI on 40 processes (potentially on multiple nodes) with 4 GPUs each, d
|
||||
$ mpirun -np 40 ./build/all_reduce_perf -b 8 -e 128M -f 2 -g 4 -c 0
|
||||
```
|
||||
|
||||
### Performance
|
||||
|
||||
See the [doc/PERFORMANCE.md](Performance) page for explanation about numbers, and in particular the "busbw" column.
|
||||
|
||||
### Arguments
|
||||
|
||||
All tests support the same set of arguments :
|
||||
|
||||
Reference in New Issue
Block a user