1 repository
Retrieves optimized convolution kernel configurations tailored to a target hardware architecture.
Distinct from Convolution Kernel Configurators: Distinct from Convolution Kernel Configurators: retrieves hardware-optimized configurations, not just defining the convolution computation.
Explore 1 awesome GitHub repository matching artificial intelligence & ml · Hardware-Aware Convolution Configurators. Refine with filters or upvote what's useful.
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
Retrieves optimized convolution configurations tailored to the target GPU architecture.