carlushuang

cpu_gemm_opt. how to design cpu gemm on x86 with avx256, that can beat openblas.

gcnasm. amdgpu example code in hip/asm

FFT_implement. fft/ifft, r2c/c2r, 2d_r2c/2d_c2r, convolve, correlation, tiling fft, srfft, pfa, radix-2/3/5

avx_flops. Benchmark cpu flops using avx instructions

miopen_cudnn_ops. C++

gemm_implementations. C

deepcore_source_code. Subpart source code of of deepcore v0.7

ogl_cube. observe a cube with basic arcball camera in c++

project-orcus. project orcus

kernel-launcher-amdgpu. C++