Menu

#355m

1 post

Feed
1 of 1 post
4 GPUs, 2x speed: why PyTorch DDP stalled on a PCIe-only 4x RTX A6000, and two flags that took it to 3.6x
📰
0Read More

4 GPUs, 2x speed: why PyTorch DDP stalled on a PCIe-only 4x RTX A6000, and two flags that took it to 3.6x

DEV Community: pytorch·Runcrate·about 6 hours ago
#r9P5Piqo
#devtopytorch#devto#dev#gpus#124m

Plain DDP on a rented 4x RTX A6000 gave 2x. Measured all-reduce bandwidth, then bf16 gradients and accumulation took it to 3.6x. Commands inside.

15s