aiacc-nccl-cu11 2.17.1

AIACC-NCCL is an AI-Accelerator communication framework for NVIDIA-NCCL. It implements optimized all-reduce, all-gather, reduce, broadcast, reduce-scatter, all-to-all,as well as any send/receive based communication pattern.It has been optimized to achieve high bandwidth on aliyun machines using PCIe, NVLink, NVswitch,as well as networking using InfiniBand Verbs, eRDMA or TCP/IP sockets.

AIACC-2.0 ACSpeed means AIACC communication Compiler Speeding. AIACC-2.0 AGSpeed means AIACC compute Graph SPEEDing. This is a distributed training framework plugin for PyTorch and aiacc nccl communication plugin for many deeplearning framworks including TensorFlow, PyTorch, MXNet and Caffe. This project is to create a uniform distributed training framework plugin tool for major frameworks, and make the distributed training as easy as possible and as fast as possible on Alibaba Cloud.