Pinned Loading

1. NCCL-Tensor-Parallelism NCCL-Tensor-Parallelism Public

   Orchestration and performance analysis of Tensor Parallelism for hand-optimized SGEMM with NCCL. Implemented in a transformer block with a custom FlashAttention backend in Triton. Validated and ben…

   Python 4

2. FlashAttention FlashAttention Public

   A custom CUDA C++ implementation of FlashAttention achieving O(N) memory complexity, enabling sequence lengths of 262144+ on a single T4 GPU where standard attention fails.

   Jupyter Notebook 5

3. llvm/llvm-project llvm/llvm-project Public

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

   LLVM 41.4k 19.5k

4. C__OOL-Compiler C__OOL-Compiler Public

   Building custom compiler for the COOL object oriented language from scratch (covers lexing, parsing, codegen)

   C++ 3

5. GEMM-Suite GEMM-Suite Public

   GEMM Optimizations from naive GEMM to cuBLAS GEMM performance.

   Cuda 2

6. custom-NN-cuda custom-NN-cuda Public

   Neural network built from scratch, accelerated with CUDA

   Jupyter Notebook 1