This is a series of GPU optimization topics. Here we will introduce how to optimize the CUDA kernel in detail. I will introduce several basic kernel optimizations, including: elementwise, reduce, sgemv, sgemm, etc. The performance of these kernels is basically at or near the theoretical limit.
Multi-Threaded FP32 Matrix Multiplication on x86 CPUs
Accelerated General (FP32) Matrix Multiplication from scratch in CUDA
General Matrix Multiplication using NVIDIA Tensor Cores
maxas Scott Grey's maxas assembler sgemm explaining the (for me) missing parts https://github.com/NervanaSystems/maxas
The repository targets the OpenCL gemm function performance optimization. It compares several libraries clBLAS, clBLAST, MIOpenGemm, Intel MKL(CPU) and cuBLAS(CUDA) on different matrix sizes/vendor's hardwares/OS. Out-of-the-box easy as MSVC, MinGW, Linux(CentOS) x86_64 binary provided. 在不同矩阵大小/硬件/操作系统下比较几个BLAS库的sgemm函数性能，提供binary，开盒即用。
FastCuda is a handwritten CUDA operator library featuring progressive GEMM and Reduce kernels, cuBLAS benchmarking, and C/C++/Python interfaces for learning, profiling, and performance optimization.
A benchmark framework for POWER and x86_64
🚀 Progressive CUDA SGEMM optimization: from naive triple-loop to Tensor Core WMMA. 5 kernel variants demonstrating GPU optimization techniques.
Perform the matrix-matrix operation `C = α*op(A)*op(B) + β*C` where `op(X)` is either `op(X) = X` or `op(X) = X^T`.
NVIDIA GPU validation: PCIe transport, Unified Memory prefetch, SGEMM compute, drift detection.
CUDA Kernel Academy: A systematic learning path for high-performance CUDA kernel development. From SGEMM basics to Tensor Core mastery with 4 progressive modules. | CUDA 高性能算子开发：从 SGEMM 基础到 Tensor Core 精通，4 模块渐进式学习路径
Add a description, image, and links to the sgemm topic page so that developers can more easily learn about it.
To associate your repository with the sgemm topic, visit your repo's landing page and select "manage topics."