Alternatives and similar repositories for mirage

Users that are interested in mirage are comparing it to the libraries listed below

• HazyResearch / ThunderKittens
  Tile primitives for speedy kernels
  ☆3,017Updated 2 weeks ago
• tspeterkim / flash-attention-minimal
  Flash Attention in ~100 lines of CUDA (forward pass only)
  ☆1,035Updated last year
• ByteDance-Seed / Triton-distributed
  Distributed Compiler based on Triton for Parallel Systems
  ☆1,289Updated last week
• flashinfer-ai / flashinfer
  FlashInfer: Kernel Library for LLM Serving
  ☆4,354Updated this week
• tile-ai / tilelang
  Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels
  ☆4,289Updated this week
• siboehm / SGEMM_CUDA
  Fast CUDA matrix multiplication from scratch
  ☆986Updated 3 months ago
• Dao-AILab / quack
  A Quirky Assortment of CuTe Kernels
  ☆724Updated last week
• bytedance / flux
  A fast communication-overlapping library for tensor/expert parallelism on GPUs.
  ☆1,214Updated 4 months ago
• HazyResearch / Megakernels
  kernels, of the mega variety
  ☆634Updated 3 months ago
• NVIDIA / cutile-python
  cuTile is a programming model for writing parallel kernels for NVIDIA GPUs
  ☆1,686Updated last week
• srush / Triton-Puzzles
  Puzzles for learning Triton
  ☆2,204Updated last year
• efeslab / Nanoflow
  A throughput-oriented high-performance serving framework for LLMs
  ☆926Updated 2 months ago
• ScalingIntelligence / KernelBench
  KernelBench: Can LLMs Write GPU Kernels? - Benchmark + Toolkit with Torch -> CUDA (+ more DSLs)
  ☆728Updated this week
• mit-han-lab / omniserve
  [MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Se…
  ☆800Updated 9 months ago
• thuml / depyf
  depyf is a tool to help you understand and adapt to PyTorch compiler torch.compile.
  ☆772Updated 2 months ago
• microsoft / BitBLAS
  BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.
  ☆735Updated 4 months ago
• flagos-ai / FlagGems
  FlagGems is an operator library for large language models implemented in the Triton Language.
  ☆811Updated this week
• perplexityai / pplx-kernels
  Perplexity GPU Kernels
  ☆542Updated last month
• pytorch / helion
  A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.
  ☆697Updated this week
• pytorch / ao
  PyTorch native quantization and sparsity for training and inference
  ☆2,591Updated this week
• SiriusNEO / Triton-Puzzles-Lite
  Puzzles for learning Triton, play it with minimal environment configuration!
  ☆576Updated 3 weeks ago
• NVIDIA / Model-Optimizer
  A unified library of SOTA model optimization techniques like quantization, pruning, distillation, speculative decoding, etc. It compresse…
  ☆1,720Updated this week
• pranjalssh / fast.cu
  Fastest kernels written from scratch
  ☆501Updated 3 months ago
• IST-DASLab / marlin
  FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.
  ☆962Updated last year
• gpu-mode / resource-stream
  GPU programming related news and material links
  ☆1,879Updated 3 months ago
• sgl-project / mini-sglang
  A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.
  ☆2,381Updated this week
• volcengine / veScale
  Byted PyTorch Distributed for Hyperscale Training of LLMs and RLs
  ☆912Updated last month
• BobMcDear / attorch
  A subset of PyTorch's neural network modules, written in Python using OpenAI's Triton.
  ☆589Updated 4 months ago
• andrewkchan / yalm
  Yet Another Language Model: LLM inference in C++/CUDA, no libraries except for I/O
  ☆542Updated 3 months ago
• NVIDIA / TransformerEngine
  A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on H…
  ☆3,027Updated last week