Standalone kernel benchmarking skill for cuda-cpp, cutlass, cute-dsl, and triton implementations. Use when the user wants to compare a custom CUDA/CUTLASS .cu kernel or CuTe DSL/Triton .py kernel against selectable PyTorch eager, torch.compile, or FlashInfer baselines, validate correctness, measure execution time with KernelBench-style CUDA event timing, or generate benchmark.md for kernel optimization results.
Corpus-backed GPU kernel knowledge base for CUDA, Triton, CuTe, CUTLASS, and Ampere/Hopper/Blackwell kernel research. Use when the user needs to search merged kernel PR pages, inspect PR diff/provenance artifacts, find KernelWiki synthesis pages, query blog/doc/contest notes, or retrieve evidence-backed implementation patterns by hardware feature, technique, repo, language, or kernel type. Do not use for environment checks, correctness checks, Nsight Compute profiling, benchmarking, or iterative optimization bookkeeping.
Iterative GPU kernel optimization orchestrator for CUDA/CUTLASS/CuTe DSL/Triton kernels. Use for measured, one-change-at-a-time optimization loops with correctness, NCU profiling, KBS evidence, hypothesis discipline, hard iteration gates, final benchmarking, and a traceable report.
Standalone kernel profiling skill for cuda-cpp, cute-dsl, cutlass, and triton implementations. Checks CUDA/PyTorch/Triton/CuTe DSL/CUTLASS/NCU/nsight-python readiness, optionally locks GPU clocks, validates correctness, collects Nsight Compute metrics with nsight-python, produces env_check.md, correctness.md, ncu_summary.md and ncu_details.md, and classifies GPU bottlenecks from NCU evidence. Use when the user wants to profile a CUDA/CUTLASS .cu kernel or CuTe DSL/Triton .py kernel, compare against a Python reference, inspect occupancy, memory, compute, scheduler, stall, or branch metrics, or diagnose Memory-Bound, Compute-Bound, Latency-Bound, Occupancy-Bound, or Mixed behavior.