Pinned Loading

1. TransformerOp TransformerOp Public

   GPT from scratch + custom CUDA optimization study

   Python

2. cornfield cornfield Public

   Op-agnostic CUDA kernel autotuner

   Python

3. SuperOptimizerVerification SuperOptimizerVerification Public

   Verified tensor superoptimizers prove kernel equivalence over exact arithmetic ( finite fields, algebraic axioms, ℝ )

   Python

4. APK APK Public

   Measuring bounded tile-level preemption in persistent CUDA kernels on Ampere

   Cuda

5. satLink satLink Public

   A reliable transport protocol for satellite-grade links | high RTT, real loss, asymmetric bandwidth.

6. customLSM customLSM Public

   Custom LSM-tree key/value store in Rust | MemTable, Bloom filters, etc

   Rust