Pinned Loading

1. nn_c nn_c Public

   High-performance CNN framework in C/CUDA: AVX-512 im2col convolution, warp-level shuffle reductions for softmax, fused optimizer kernels, and a 256-byte aligned workspace allocator for coalesced gl…

   C

2. nvim-syncer nvim-syncer Public

   A lightweight Neovim plugin to sync files across hosts using Rsync.

   Lua 2

3. custom-mpi-impl custom-mpi-impl Public

   MPI library implemented from scratch in C over Unix sockets: point-to-point send/recv, gather, broadcast, and barrier primitives built to spec without using any MPI library code.

   C

4. cfd-lake cfd-lake Public

   CUDA/MPI parallel simulation of 2D wave propagation using centralized finite difference. Implementations across CUDA, multi-GPU+MPI, OpenMP, OpenACC, and Mojo. 20x speedup on large grids.

   Cuda

5. LER-IR LER-IR Public

   MLIR compiler for loop redundancy elimination: implements the GLORE algorithm (OOPSLA '17) with a custom dialect, hand-written lexer/parser, and lowering pipeline through affine/scf/arith/memref to…

   Java

6. llvm/llvm-project llvm/llvm-project Public

   The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.

   LLVM 38.7k 17.4k