From 51ee17ce44c565f58e6b4792465a84001798d467 Mon Sep 17 00:00:00 2001 From: Xiaoyu Zhang <35585791+BBuf@users.noreply.github.com> Date: Tue, 3 Mar 2026 02:51:29 +0800 Subject: [PATCH] [diffusion] move skills dir (#19697) --- .../.claude/skills/diffusion-kernel/SKILL.md | 35 +++++++++++++++++++ .../diffusion-kernel}/add-triton-kernel.md | 0 .../diffusion-benchmark-and-profile.md | 0 .../diffusion-kernel}/nsight-profiler.md | 0 .../use-efficient-diffusion-kernels.md | 0 .../diffusion-perf/SKILL.md | 0 6 files changed, 35 insertions(+) create mode 100644 python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/SKILL.md rename {.claude/skills/diffusion => python/sglang/multimodal_gen/.claude/skills/diffusion-kernel}/add-triton-kernel.md (100%) rename {.claude/skills/diffusion => python/sglang/multimodal_gen/.claude/skills/diffusion-kernel}/diffusion-benchmark-and-profile.md (100%) rename {.claude/skills/diffusion => python/sglang/multimodal_gen/.claude/skills/diffusion-kernel}/nsight-profiler.md (100%) rename {.claude/skills/diffusion => python/sglang/multimodal_gen/.claude/skills/diffusion-kernel}/use-efficient-diffusion-kernels.md (100%) rename python/sglang/multimodal_gen/.claude/{.skills => skills}/diffusion-perf/SKILL.md (100%) diff --git a/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/SKILL.md b/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/SKILL.md new file mode 100644 index 000000000..7b696903d --- /dev/null +++ b/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/SKILL.md @@ -0,0 +1,35 @@ +--- +name: diffusion-kernel +description: Index for SGLang Diffusion kernel development skills. +--- + +# Diffusion Kernel Skills + +## Directory Layout + +``` +python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/ +├── SKILL.md +├── add-triton-kernel.md +├── diffusion-benchmark-and-profile.md +├── nsight-profiler.md +└── use-efficient-diffusion-kernels.md +``` + +## Index + +- [add-triton-kernel.md](./add-triton-kernel.md) + + Step-by-step guide for adding a new Triton kernel to SGLang Diffusion's `jit_kernel` module, including authoring, autotune, `torch.compile` compatibility, integration, and tests. + +- [use-efficient-diffusion-kernels.md](./use-efficient-diffusion-kernels.md) + + Practical guidance for using SGLang Diffusion fused kernels and fast CUDA paths, including constraints, fallbacks, and where the fused ops are wired into the runtime. + +- [diffusion-benchmark-and-profile.md](./diffusion-benchmark-and-profile.md) + + End-to-end benchmarking and profiling guide for SGLang Diffusion models, including denoise latency measurement, per-layer breakdown, and regression tracking. + +- [nsight-profiler.md](./nsight-profiler.md) + + Advanced profiling skill for NVIDIA Nsight Systems / Nsight Compute: collecting traces, reading reports, and interpreting kernel-level performance metrics. diff --git a/.claude/skills/diffusion/add-triton-kernel.md b/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/add-triton-kernel.md similarity index 100% rename from .claude/skills/diffusion/add-triton-kernel.md rename to python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/add-triton-kernel.md diff --git a/.claude/skills/diffusion/diffusion-benchmark-and-profile.md b/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/diffusion-benchmark-and-profile.md similarity index 100% rename from .claude/skills/diffusion/diffusion-benchmark-and-profile.md rename to python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/diffusion-benchmark-and-profile.md diff --git a/.claude/skills/diffusion/nsight-profiler.md b/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/nsight-profiler.md similarity index 100% rename from .claude/skills/diffusion/nsight-profiler.md rename to python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/nsight-profiler.md diff --git a/.claude/skills/diffusion/use-efficient-diffusion-kernels.md b/python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/use-efficient-diffusion-kernels.md similarity index 100% rename from .claude/skills/diffusion/use-efficient-diffusion-kernels.md rename to python/sglang/multimodal_gen/.claude/skills/diffusion-kernel/use-efficient-diffusion-kernels.md diff --git a/python/sglang/multimodal_gen/.claude/.skills/diffusion-perf/SKILL.md b/python/sglang/multimodal_gen/.claude/skills/diffusion-perf/SKILL.md similarity index 100% rename from python/sglang/multimodal_gen/.claude/.skills/diffusion-perf/SKILL.md rename to python/sglang/multimodal_gen/.claude/skills/diffusion-perf/SKILL.md