Qwen3-0.6B megakernel: 527 tok/s decode on RTX 3090 (3.8x faster than PyTorch)
☆126Feb 10, 2026Updated 5 months ago
Alternatives and similar repositories for MegaQwen
Users that are interested in MegaQwen are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Aggressive decode optimizations for Qwen3-0.6B on RTX 5090☆56Feb 25, 2026Updated 5 months ago
- ☆22Jul 10, 2026Updated 3 weeks ago
- A simple environment for writing and experimenting with hand-written CUDA PTX kernels.☆18Sep 11, 2025Updated 10 months ago
- ☆19Apr 26, 2026Updated 3 months ago
- High-performance GPU kernels for LLM inference in OpenAI Triton. Fused RMSNorm, SwiGLU, INT8 GEMM with benchmarks and roofline analysis.☆34Jul 22, 2026Updated last week
- Wordpress hosting with auto-scaling - Free Trial Offer • AdFully Managed hosting for WordPress and WooCommerce businesses that need reliable, auto-scalable performance. Cloudways SafeUpdates now available.
- Writing FLUX in Triton☆42Sep 22, 2024Updated last year
- GGML bindings that aim to be idiomatic Rust rather than directly corresponding to the C/C++ interface☆20Sep 25, 2023Updated 2 years ago
- Pure Triton kernels for Qwen3.5-27B inference on NVIDIA B200☆122Feb 28, 2026Updated 5 months ago
- Kernels, of the mega variety :)☆790May 26, 2026Updated 2 months ago
- MetaAttention: A Unified and Performant Attention Framework Across Hardware Backends(PPoPP'26)☆17Dec 31, 2025Updated 7 months ago
- 存放一些 CUDA 编程相关的博客文件。☆22Oct 16, 2025Updated 9 months ago
- ☆178Updated this week
- Region-level profiling for CUDA kernels with trace, NVBit, CUPTI, NSys, and an interactive Explorer.☆123Apr 17, 2026Updated 3 months ago
- Optimize GEMM with tensorcore step by step☆40Dec 17, 2023Updated 2 years ago
- Proton VPN Special Offer - Get 70% off • AdSpecial partner offer. Trusted by over 100 million users worldwide. Tested, Approved and Recommended by Experts.
- mKernel: fast multi-node, multi-GPU fused kernels☆260Updated this week
- High-performance LLM operator library built on TileLang.☆167Updated this week
- MoE training for Me and You and maybe other people☆396Mar 15, 2026Updated 4 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆116Jun 28, 2025Updated last year
- A Triton-only attention backend for vLLM☆27Jul 14, 2026Updated 2 weeks ago
- Mirage Persistent Kernel: Compiling LLMs into a MegaKernel☆2,404Jul 28, 2026Updated last week
- Fastest kernels written from scratch☆588Sep 18, 2025Updated 10 months ago
- A PyTorch implementation of the GPT-OSS-20B architecture. All components are coded from scratch: RoPE with YaRN, RMSNorm, SwiGLU with cla…☆238Dec 2, 2025Updated 8 months ago
- ☆52Dec 19, 2025Updated 7 months ago
- Wordpress hosting with auto-scaling - Free Trial Offer • AdFully Managed hosting for WordPress and WooCommerce businesses that need reliable, auto-scalable performance. Cloudways SafeUpdates now available.
- ☆336Jun 9, 2026Updated last month
- ☆193May 7, 2025Updated last year
- Benchmark tests supporting the TiledCUDA library.☆19Nov 19, 2024Updated last year
- Github mirror of trition-lang/triton repo.☆183Updated this week
- CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs☆236Updated this week
- Slow-Fast Inference: Training-free Inference Acceleration via Within-sentence Support Stability☆18Updated this week
- A "standard library" of Triton kernels.☆26Oct 2, 2025Updated 10 months ago
- FlashTile is a CUDA Tile IR compiler that is compatible with NVIDIA's tileiras, targeting SM70 through SM121 NVIDIA GPUs.☆61Feb 6, 2026Updated 5 months ago
- Winner 🏆 (Agent-only) MLSys 2026 - FlashInfer AI Kernel Generation Contest for the DeepSeek Sparse Attention (DSA) track with an average…☆148Jun 10, 2026Updated last month
- 1-Click AI Models by DigitalOcean Gradient • AdDeploy popular AI models on DigitalOcean Gradient GPU virtual machines with just a single click. Zero configuration with optimized deployments.
- My submission for the GPUMODE/AMD fp8 mm challenge☆29Jun 4, 2025Updated last year
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆157May 10, 2025Updated last year
- A series of high-performance GEMM (General Matrix Multiply) implementations Iteratively optimised for H100 GPUs in Pure CUDA.☆81Feb 18, 2026Updated 5 months ago
- Exploring how optimizations for GEMMs work☆37Feb 28, 2026Updated 5 months ago
- An agent harness that compiles a model into one provably-correct, self-retargeting CUDA megakernel and self-tunes it past cuBLAS at batch…☆125Jun 29, 2026Updated last month
- mHC kernels implemented in CUDA☆265Mar 9, 2026Updated 4 months ago
- EleutherAI ML Performance reading group repository (slides, meeting recordings, annotated papers)☆36Mar 20, 2026Updated 4 months ago