PyTorch 2.9 + HIP/ROCm learning notes: dispatcher, accelerator backend, profiler, TorchInductor, runtime, GEMM, and C++ extensions.
-
Updated
Jul 20, 2026 - Python
#
torchinductor
Here are 2 public repositories matching this topic...
Reproducible study of how tensor layout, reduction length, and pointer alignment drive vendor GEMM dispatch and latency cliffs.
benchmarking gpu cuda pytorch cuda-kernels performance-analysis reproducibility gemm gpu-profiling systems-research torchinductor deep-learning-compilers tensor-layout kernel-dispatch
-
Updated
Sep 2, 2026 - Python
Add this topic to your repo
To associate your repository with the torchinductor topic, visit your repo's landing page and select "manage topics."