batista

deepnccl

Last released May 14, 2024

DEEP-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.

deepncclplugin

Last released Apr 11, 2024

DEEP-NCCL-PLUGIN is a plugin of AI-Accelerator communication framework for NVIDIA-NCCL. It has been optimized to achieve high bandwidth on aliyun machines using InfiniBand Verbs, eRDMA or TCP/IP sockets.

deep-nccl-wrapper

Last released Nov 9, 2023

Deep-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-nccl-plugin

Last released Oct 19, 2023

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.

deep-index

Last released Oct 11, 2023

Deep-Index is DeepGPU/DeepNCCL/DeepyTorch's root_path.

aiacc-nccl

Last released Oct 9, 2023

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.

NCCL-plugin

Last released Oct 9, 2023

NCCL-plugin

aiacc-nccl-cu11

Last released Sep 5, 2023

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.