GeForce GTX 1080 Ti 4枚で同時に bandwidthTest してみました。。。
# ./bandwidthTest --device=all
[CUDA Bandwidth Test] - Starting...
!!!!!Cumulative Bandwidth to be computed from all the devices !!!!!!
Running on...
Device 0: GeForce GTX 1080 Ti
Device 1: GeForce GTX 1080 Ti
Device 2: GeForce GTX 1080 Ti
Device 3: GeForce GTX 1080 Ti
Quick Mode
Host to Device Bandwidth, 4 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 48236.7
Device to Host Bandwidth, 4 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 51457.5
Device to Device Bandwidth, 4 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 1388806.5
Result = PASS
◯計算サーバ
CPU : Intel(R) Xeon(R) Gold 5120 CPU @ 2.20GHz x 2
Memory:256GB
GPU : GeForce GTC 1080 Ti x 4
OS : Linux CentOS 7.5
# ./bandwidthTest --device=all
[CUDA Bandwidth Test] - Starting...
!!!!!Cumulative Bandwidth to be computed from all the devices !!!!!!
Running on...
Device 0: GeForce GTX 1080 Ti
Device 1: GeForce GTX 1080 Ti
Device 2: GeForce GTX 1080 Ti
Device 3: GeForce GTX 1080 Ti
Quick Mode
Host to Device Bandwidth, 4 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 48236.7
Device to Host Bandwidth, 4 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 51457.5
Device to Device Bandwidth, 4 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 1388806.5
Result = PASS
◯計算サーバ
CPU : Intel(R) Xeon(R) Gold 5120 CPU @ 2.20GHz x 2
Memory:256GB
GPU : GeForce GTC 1080 Ti x 4
OS : Linux CentOS 7.5