sglang-diffusion-ako4all-kernel

Name: sglang-diffusion-ako4all-kernel
Availability: InStock
Author: annealing-inversion

Automate AKO4ALL-based optimization of SGLang diffusion kernels with benchmarking and validation.

Updated Jun 4, 2026

Search Tags:#benchmark #ncu #harness #diffusion #sglang #kernel-optimization #ako4all

Ready to run

npx skills add https://github.com/annealing-inversion/sglang-kimi-deferral --skill sglang-diffusion-ako4all-kernel-annealing-inversion

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: sglang-diffusion-ako4all-kernel
Source: https://github.com/annealing-inversion/sglang-kimi-deferral/tree/main/python/sglang/multimodal_gen/.claude/skills/sglang-diffusion-ako4all-kernel
Command: npx skills add https://github.com/annealing-inversion/sglang-kimi-deferral --skill sglang-diffusion-ako4all-kernel-annealing-inversion

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires git, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Enables end-to-end AKO4ALL-based optimization of an existing SGLang diffusion kernel, including setup, benchmarking, and model-level validation.

Core Features & Use Cases

Custom AKO4ALL harness for a diffusion kernel, with mirror of input and reference, microbench setup, and iterative profiling.
Preflight checks and baseline establishment to ensure a clean, synced AKO4ALL environment.
End-to-end validation via targeted denoise benchmarks and artifact preparation for PRs.

Quick Start

Bootstrap a clean AKO4ALL workspace and begin the kernel optimization loop using the AKO4ALL harness.

Frequently Asked Questions about sglang-diffusion-ako4all-kernel

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I optimize a diffusion kernel in an SGLang project?▼

Optimize an SGLang diffusion kernel by bootstrapping a clean AKO4ALL workspace, establishing a baseline, running microbenchmarks, and validating results with targeted denoise benchmarks.

What is AKO4ALL harness-based kernel optimization?▼

AKO4ALL harness-based kernel optimization automates iterative profiling and microbenchmarking for existing kernels, enabling end-to-end tuning from preflight checks to model-level validation.

Do I need a clean AKO4ALL repository to run SGLang kernel benchmarks?▼

Yes, you need a clean AKO4ALL repository to ensure preflight checks pass and the harness can properly bootstrap the microbench setup for SGLang kernel benchmarking.

How do I profile a diffusion kernel using ncu in an AKO4ALL harness?▼

Profile a diffusion kernel using ncu by running the provided AKO4ALL harness profilers during the iterative optimization loop to capture microbenchmark metrics.

Can I validate SGLang diffusion kernel changes with denoise benchmarks?▼

Yes, you can validate SGLang diffusion kernel changes by running targeted denoise benchmarks within the AKO4ALL harness to ensure model-level performance and correctness.

Does this AKO4ALL harness require specific hardware to run microbenchmarks?▼

Yes, the AKO4ALL harness requires access to relevant hardware to execute microbenchmarks and run profilers like ncu for accurate kernel optimization measurements.