jundaf

INT8-Flash-Attention-FMHA-Quantization. Cuda

CUDA-INT8-GEMM. CUDA 8-bit Tensor Core Matrix Multiplication based on m16n16k16 WMMA API

eigenMHA. Forward and backward Attention DNN operators implementationed by LibTorch, cuDNN, and Eigen.

Finite-Element-Domain-Decomposition. Overlapping Schwarz Domain Decomposition Finite Element Algorithm in both Matlab and serial/parallel C++

Time-Domain-CEM. The only known (by 2022) open-source, easy-to-understand basic algorithm implementations in TD-CEM. (Please star and fork this project if you find it useful!)

PDE-Net-FDTD. Python

DNN-2d-FDTD. Python

RNN-1d-FDTD. Python

GPU-Tensor-Permute. permute sequence data on GPU with high bandwidth

dnn-test-framework. DNN unit test framework

regex-gpu. Block-optimized regular expression matching engine on GPU

cutlass-kernel-volta-gemm. volta fp16 gemm kernel

Flash-LightSeq. C++

eigenDNN.

adaptive-filtering-algorithms. Adaptive Algorithms

FETD-MFEM. A simple Finite element time domain example built with MFEM

DNN-Discrete-Hilbert-Transform. Python

cutlass-b2bgemm. an extension to the cutlass half-precision b2b gemm example

Heterogeneous-GPUs. Heterogeneous Nvidia (CUDA) and Intel (OpenCL) GPU Programming

method-of-moments. method of moments (MoM) for conducting wire problem

FelixFu520-README. A pupil in the computer world.(Felix Fu)

FlyDSL. Python