muriloboratto / NCCLLinks
Sample examples of how to call collective operation functions on multi-GPU environments. A simple example of using broadcast, reduce, allGather, reduceScatter and sendRecv operations.
☆33Updated last year
Alternatives and similar repositories for NCCL
Users that are interested in NCCL are comparing it to the libraries listed below
Sorting:
- Matrix Multiply-Accumulate with CUDA and WMMA( Tensor Core)☆134Updated 4 years ago
- ☆110Updated last month
- An extension library of WMMA API (Tensor Core API)☆97Updated 10 months ago
- ☆88Updated 5 months ago
- ☆73Updated 3 weeks ago
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆62Updated 9 months ago
- ☆121Updated 6 months ago
- ☆96Updated 8 months ago
- ☆97Updated last year
- A lightweight design for computation-communication overlap.☆136Updated this week
- Benchmark code for the "Online normalizer calculation for softmax" paper☆94Updated 6 years ago
- ☆51Updated 5 years ago
- NCCL Examples from Official NVIDIA NCCL Developer Guide.☆16Updated 7 years ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆79Updated 3 weeks ago
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆182Updated 4 months ago
- Examples of CUDA implementations by Cutlass CuTe☆190Updated 4 months ago
- ☆79Updated 2 years ago
- Optimize GEMM with tensorcore step by step☆26Updated last year
- ☆37Updated 5 months ago
- ☆148Updated 4 months ago
- GEMM by WMMA (tensor core)☆12Updated 2 years ago
- Dissecting NVIDIA GPU Architecture☆95Updated 2 years ago
- A hierarchical collective communications library with portable optimizations☆35Updated 5 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆88Updated this week
- CUDA 8-bit Tensor Core Matrix Multiplication based on m16n16k16 WMMA API☆30Updated last year
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆69Updated 9 months ago
- A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.☆81Updated 3 weeks ago
- ☆25Updated 3 months ago
- Magicube is a high-performance library for quantized sparse matrix operations (SpMM and SDDMM) of deep learning on Tensor Cores.☆88Updated 2 years ago
- Implementation of TSM2L and TSM2R -- High-Performance Tall-and-Skinny Matrix-Matrix Multiplication Algorithms for CUDA☆32Updated 4 years ago