AravSrinivas
@ozogysozogysevaq
Using a custom RDMA-based networking library, we've been able to achieve 3200 Gbps GPU memory transfers, bypassing NCCL limits for 97.1% theoretical bandwidth efficiency. Our latest blog shares our journey of building a custom high-performance networking solution on AWS. https://t.co/RVMMDAFGT7
0 reply
0 recast
0 reaction