1 个仓库
Awesome GitHub RepositoriesKernel Source Transformers
Tools that intercept and modify compiled GPU kernel source code before final output.
Distinct from Workflow-to-Code Generators: Distinct from Workflow-to-Code Generators: focuses on transforming already-compiled kernel code, not generating code from workflows.
Explore 1 awesome GitHub repository matching development tools & productivity · Kernel Source Transformers. Refine with filters or upvote what's useful.
Awesome Kernel Source Transformers GitHub Repositories
用 AI 发现最棒的仓库。我们将通过 AI 为您搜索最匹配的仓库。
- tile-ai/tilelang
tile-ai/tilelang
5,226在 GitHub 上查看TileLang is a Python-embedded domain-specific language compiler that JIT-compiles and autotunes GPU kernels. It uses a tile-based DSL, automatic software pipelining, and parallel autotuning to generate optimized GPU kernels at runtime. It supports tensor core operations with Pythonic syntax, automatic memory management, and thread mapping. The compiler searches over tile sizes, thread counts, and scheduling policies, compiling and benchmarking candidates in parallel to find the fastest kernel. It also caches compiled binaries and tuning results to disk for reuse across sessions. TileLang inc
Allows custom transformations to be applied to compiled GPU kernel source code before final execution.
Python
