Pinned Loading

1. CUDA-Kernel-Library CUDA-Kernel-Library Public

   Hand-written CUDA linear algebra library. A GEMM driven from a naive baseline to a compute-bound tensor-core kernel at ~90% of cuBLAS on an RTX 5070, profiled with Nsight Compute, plus GEMV, TRSM, …

   C++ 1

2. ARM-Neon-Conv3x3 ARM-Neon-Conv3x3 Public

   ARMv8-A NEON 3×3 convolution in C++17. Scalar, NEON naive, and cache/register-blocked variants with runtime dispatch and perf/PMU analysis.

   C++

3. LLVM-Optimization-Pass-Framework LLVM-Optimization-Pass-Framework Public

   Out of tree LLVM new pass manager framework: custom analysis and transformation passes with lit/FileCheck tests, differential execution validation, and a measured before/after benchmark report agai…

   C++

4. GPU-Roofline-Profiler GPU-Roofline-Profiler Public

   CUDA roofline profiling suite for the RTX 5070: measured compute and memory ceilings, a hand-written kernel ladder from SAXPY to a register-blocked GEMM, Nsight-backed bottleneck analysis, and a fu…

   Python

5. MLIR-Backend-for-a-Simulated-Edge-NPU MLIR-Backend-for-a-Simulated-Edge-NPU Public

   A from-scratch deep learning compiler: it lowers trained ONNX models through two custom MLIR dialects and a full optimization pipeline to a binary instruction stream for a simulated edge NPU, then …

   C++

6. CPU-Microbenchmark-Suite CPU-Microbenchmark-Suite Public

   Measures cache latency, STREAM memory bandwidth, and peak AVX2/NEON FLOPS on real hardware, then assembles them into a measured roofline and a BLIS-style GEMM tile prediction validated empirically.…

   C++