Qwen3-0.6B megakernel: 527 tok/s decode on RTX 3090 (3.8x faster than PyTorch)
☆139Feb 10, 2026Updated 7 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☆63Feb 25, 2026Updated 7 months ago
- ☆27Aug 29, 2026Updated last month
- A simple environment for writing and experimenting with hand-written CUDA PTX kernels.☆19Sep 11, 2025Updated last year
- ☆20Apr 26, 2026Updated 5 months ago
- High-performance GPU kernels for LLM inference in OpenAI Triton. Fused RMSNorm, SwiGLU, INT8 GEMM with benchmarks and roofline analysis.☆45Jul 22, 2026Updated 2 months ago
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Writing FLUX in Triton☆42Sep 22, 2024Updated 2 years ago
- GGML bindings that aim to be idiomatic Rust rather than directly corresponding to the C/C++ interface☆20Sep 25, 2023Updated 3 years ago
- ☆24Aug 1, 2026Updated 2 months ago
- Pure Triton kernels for Qwen3.5-27B inference on NVIDIA B200☆124Feb 28, 2026Updated 7 months ago
- LLM Inference via Triton (Flexible & Modular): Focused on Kernel Optimization using CUBIN binaries, Starting from gpt-oss Model☆120Apr 28, 2026Updated 5 months ago
- Kernels, of the mega variety :)☆839May 26, 2026Updated 4 months ago
- MetaAttention: A Unified and Performant Attention Framework Across Hardware Backends(PPoPP'26)☆18Aug 6, 2026Updated last month
- This repositories contains the reference implementation for the Sparse Delta Memory paper.More precisely, it contains the model definitio…☆40Jul 9, 2026Updated 2 months ago
- Humming is a high-performance, lightweight, and highly flexible JIT (Just-In-Time) compiled GEMM kernel library specifically designed for…☆241Updated this week
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- Region-level profiling for CUDA kernels with trace, NVBit, CUPTI, NSys, and an interactive Explorer.☆132Apr 17, 2026Updated 5 months ago
- [TMLR 25] An automated method for explaining complex neuron behaviors in deep vision models using large language models☆12Feb 20, 2025Updated last year
- 存放一些 CUDA 编程相关的博客文件。☆24Oct 16, 2025Updated 11 months ago
- Optimize GEMM with tensorcore step by step☆40Dec 17, 2023Updated 2 years ago
- mKernel: fast multi-node, multi-GPU fused kernels☆287Updated this week
- High-performance LLM operator library built on TileLang.☆191Updated this week
- Companion code for The Physics of LLM Inference book☆29Apr 21, 2026Updated 5 months ago
- MoE training for Me and You and maybe other people☆399Mar 15, 2026Updated 6 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆119Aug 4, 2026Updated 2 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.
- Cuda implemenation of flash-kmeans, 2x faster☆23May 8, 2026Updated 4 months ago
- A Triton-only attention backend for vLLM☆28Jul 14, 2026Updated 2 months ago
- Repository for the CUDA H100 Course☆76Apr 12, 2026Updated 5 months ago
- Mirage Persistent Kernel: Compiling LLMs into a MegaKernel☆2,528Updated this week
- Fastest kernels written from scratch☆625Aug 15, 2026Updated last month
- A PyTorch implementation of the GPT-OSS-20B architecture. All components are coded from scratch: RoPE with YaRN, RMSNorm, SwiGLU with cla…☆237Dec 2, 2025Updated 10 months ago
- ☆57Dec 19, 2025Updated 9 months ago
- ☆485Aug 26, 2026Updated last month
- ☆202Aug 9, 2026Updated last month
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- Benchmark tests supporting the TiledCUDA library.☆19Nov 19, 2024Updated last year
- Github mirror of trition-lang/triton repo.☆197Updated this week
- CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs☆254Updated this week
- A pure-Python implementation of the Nvidia CuTe layout algebra intended to be approachable and easy to learn.☆252Jun 29, 2026Updated 3 months ago
- a fast and customizable CUDA int4 tensor core gemm☆16Aug 2, 2024Updated 2 years ago
- FlashTile is a CUDA Tile IR compiler that is compatible with NVIDIA's tileiras, targeting SM70 through SM121 NVIDIA GPUs.☆60Feb 6, 2026Updated 7 months ago
- My submission for the GPUMODE/AMD fp8 mm challenge☆29Jun 4, 2025Updated last year