cassiewilliam / cuda_op_benchmark
方便扩展的Cuda算子理解和优化框架,仅用在学习使用
☆13Updated 9 months ago
Alternatives and similar repositories for cuda_op_benchmark:
Users that are interested in cuda_op_benchmark are comparing it to the libraries listed below
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing. By pro…☆70Updated this week
- GPTQ inference TVM kernel☆38Updated 11 months ago
- A practical way of learning Swizzle☆15Updated last month
- Benchmark tests supporting the TiledCUDA library.☆15Updated 4 months ago
- Quantized Attention on GPU☆45Updated 4 months ago
- ☆19Updated 6 months ago
- Framework to reduce autotune overhead to zero for well known deployments.☆63Updated last week
- DeeperGEMM: crazy optimized version☆61Updated 2 weeks ago
- Transformers components but in Triton☆32Updated last week
- ☆51Updated 2 weeks ago
- FractalTensor is a programming framework that introduces a novel approach to organizing data in deep neural networks (DNNs) as a list of …☆24Updated 3 months ago
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆35Updated 3 weeks ago
- An Attention Superoptimizer☆21Updated 2 months ago
- Implementation from scratch in C of the Multi-head latent attention used in the Deepseek-v3 technical paper.☆15Updated 2 months ago
- A simple calculation for LLM MFU.☆27Updated 3 weeks ago
- Open deep learning compiler stack for cpu, gpu and specialized accelerators☆18Updated last week
- Estimate MFU for DeepSeekV3☆21Updated 2 months ago
- TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators☆32Updated 3 weeks ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆63Updated this week
- PyTorch bindings for CUTLASS grouped GEMM.☆77Updated 4 months ago
- TensorRT LLM Benchmark Configuration☆13Updated 8 months ago
- Implement Flash Attention using Cute.☆74Updated 3 months ago
- TiledKernel is a code generation library based on macro kernels and memory hierarchy graph data structure.☆19Updated 10 months ago
- A Suite for Parallel Inference of Diffusion Transformers (DiTs) on multi-GPU Clusters☆44Updated 8 months ago
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆66Updated this week
- Odysseus: Playground of LLM Sequence Parallelism☆68Updated 9 months ago
- Standalone Flash Attention v2 kernel without libtorch dependency☆108Updated 6 months ago
- ThrillerFlow is a Dataflow Analysis and Codegen Framework written in Rust.☆14Updated 4 months ago
- Multiple GEMM operators are constructed with cutlass to support LLM inference.☆17Updated 6 months ago
- ☆10Updated this week