2 repository-uri
Uses a block-based programming model where threads cooperate on data tiles to leverage Tensor Cores for matrix multiplication and Fourier transforms.
Distinct from Tile-Based Kernel Authoring: Distinct from Tile-Based Kernel Authoring: focuses on the programming model for tile-based GPU kernels, not just memory-tiling strategies.
Explore 2 awesome GitHub repositories matching artificial intelligence & ml · Tile-Based GPU Kernel Programming. Refine with filters or upvote what's useful.
Uses a block-based programming model where threads cooperate on data tiles to leverage Tensor Cores for matrix multiplication and Fourier transforms.
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
Provides a tile-level programming model that abstracts individual thread management.