There are 2 repositories under gemm-optimization topic.
row-major matmul optimization
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。
Manually optimize the GEMM (GEneral Matrix Multiply) operation. There is a long way to go.
Fast SpMM implementation on GPUs for GNN (IPDPS'23)
Generate optimized MatMul cuda kernel automatically using tvm auto schedule.
Case Studies for using Accera - the open source cross-platform compiler from Microsoft Research - to create high performance deep learning computations (i.e. GEMM, Convolution, etc.)
Implementations of DGEMM algorithm using different tricks to optimize the performance.
My attempt of making a GEMM kernel...
My experiments with convolution
ConvLIB is a library of convolution kernels for multicore processors with ARM (NEON) or RISC-V architecture
Implementations of SGEMM algorithm on Nvidia GPU using different tricks to optimize the performance.