Aggressive decode optimizations for Qwen3-0.6B on RTX 5090
☆64Feb 25, 2026Updated 8 months ago
Alternatives and similar repositories for qwen_megakernel
Users that are interested in qwen_megakernel are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Qwen3-0.6B megakernel: 527 tok/s decode on RTX 3090 (3.8x faster than PyTorch)☆145Feb 10, 2026Updated 8 months ago
- Speed-of-Light SW efficiency by using ultra low-latency primitives for comms collectives☆13Apr 8, 2026Updated 6 months ago
- ☆13Oct 12, 2025Updated last year
- Region-level profiling for CUDA kernels with trace, NVBit, CUPTI, NSys, and an interactive Explorer.☆133Apr 17, 2026Updated 6 months ago
- Super fast serving stack for LLM on Windows/Linux/Macos☆17Oct 22, 2026Updated last week
- AI Agents on DigitalOcean Gradient AI Platform • AdBuild production-ready AI agents using customizable tools or access multiple LLMs through a single endpoint. Create custom knowledge bases or connect external data.
- writing really fast kernels☆21Oct 11, 2026Updated 2 weeks ago
- MetaAttention: A Unified and Performant Attention Framework Across Hardware Backends(PPoPP'26)☆18Aug 6, 2026Updated 2 months ago
- ☆100Sep 8, 2026Updated last month
- mHC kernels implemented in CUDA☆267Mar 9, 2026Updated 7 months ago
- Unified tracing profiler and visualizer, one timeline from CPU to GPU/HES to in-kernel zones☆149Sep 21, 2026Updated last month
- xllama.cpp — Optimized LLM Inference for NVidia Tesla V100-SXM2-32GB☆17Apr 29, 2026Updated 6 months ago
- (trying to) train llama2 7B fast☆10Oct 14, 2025Updated last year
- ☆18Dec 2, 2024Updated last year
- mKernel: fast multi-node, multi-GPU fused kernels☆306Oct 17, 2026Updated last week
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- A kernel visualization that walks through each step of the code.☆14Aug 1, 2026Updated 2 months ago
- High-performance GPU kernels for LLM inference in OpenAI Triton. Fused RMSNorm, SwiGLU, INT8 GEMM with benchmarks and roofline analysis.☆44Jul 22, 2026Updated 3 months ago
- User space Block device access API☆14Aug 25, 2026Updated 2 months ago
- FlexTensor is a tensor offloading and management library for PyTorch that enables running large models on limited GPU memory by intellige…☆120Oct 14, 2026Updated 2 weeks ago
- MLIR metal dialect☆38Sep 18, 2024Updated 2 years ago
- Implementation of a Light Recurrent Unit in Pytorch☆51Oct 9, 2026Updated 2 weeks ago
- Writing FLUX in Triton☆42Sep 22, 2024Updated 2 years ago
- TVM for Tenstorrent ASICs☆31Oct 21, 2026Updated last week
- FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM Serving☆32Aug 28, 2026Updated 2 months ago
- 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.
- A NCCL extension library, designed to efficiently offload GPU memory allocated by the NCCL communication library.☆122Dec 17, 2025Updated 10 months ago
- Quick and dirty benchmarking tool to measure the performance of RK3588 NPU.☆12Feb 23, 2024Updated 2 years ago
- Efficient Long-context Language Model Training by Core Attention Disaggregation☆106Apr 7, 2026Updated 6 months ago
- A novel temporal fusion framework for propelling autoregressive model inference☆11Updated this week
- 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
- klein.c is a compressed pure C implementation of iris.c for text-to-image generation using FLUX.2 Klein transformer models. Built specifi…☆16Feb 26, 2026Updated 8 months ago
- ☆41Oct 5, 2026Updated 3 weeks ago
- ☆19Jan 4, 2024Updated 2 years ago
- An agent harness that compiles a model into one provably-correct, self-retargeting CUDA megakernel and self-tunes it past cuBLAS at batch…☆157Sep 18, 2026Updated last month
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- An experimental communicating attention kernel based on DeepEP.☆34Jul 29, 2025Updated last year
- ☆11Oct 13, 2023Updated 3 years ago
- Dataset for https://madebyoll.in/posts/game_emulation_via_dnn/☆12Jan 26, 2026Updated 9 months ago
- So, I trained a MoE based a 124M (8x12M) architecture I coded from ground up to build a small instruct model, going through the below-men…☆14Jul 18, 2025Updated last year
- ☆12Mar 16, 2022Updated 4 years ago
- Implemented and benchmarked KV cache compression methods for LLM inference in Triton. Optimized kernels for KIVI and TurboQuant with Flas…☆18Jun 19, 2026Updated 4 months ago
- Fastest kernels written from scratch☆629Aug 15, 2026Updated 2 months ago