Several optimization methods of half-precision general matrix multiplication (HGEMM) using tensor core with WMMA API and MMA PTX instruction.
-
Updated
Sep 8, 2024 - Cuda
#
tensor-core
Here are 17 public repositories matching this topic...
CUDA编程练习项目-Hands-on CUDA kernels and performance optimization, covering GEMM, FlashAttention, Tensor Cores, CUTLASS, quantization, KV cache, NCCL, and profiling.
parallel-computing cuda high-performance-computing cuda-kernels quantization cutlass gemm performance-optimization nccl gpu-programming roofline-model tensor-core llm-inference flash-attention nsight-compute
-
Updated
Mar 20, 2026 - Cuda
Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.
gpu cuda inference nvidia cutlass mha multi-head-attention llm tensor-core large-language-model flash-attention flash-attention-2
-
Updated
Feb 27, 2025 - C++
Multiple GEMM operators are constructed with cutlass to support LLM inference.
-
Updated
Aug 3, 2025 - C++
FastCuda is a handwritten CUDA operator library featuring progressive GEMM and Reduce kernels, cuBLAS benchmarking, and C/C++/Python interfaces for learning, profiling, and performance optimization.
-
Updated
Mar 18, 2026 - Cuda
Use tensor core to calculate back-to-back HGEMM (half-precision general matrix multiplication) with MMA PTX instruction.
gpu cuda cublas nvidia gemm matrix-multiply tensor-core hgemm back2back-hgemm fused-hgemm back2back-gemm fused-gemm
-
Updated
Nov 3, 2023 - Cuda
Codes for DTC-SpMM (ASPLOS'24)
-
Updated
Jun 2, 2024 - C++
🎓 CUDA HPC Kernel Optimization Lab: Progressive GEMM, FlashAttention, Tensor Core & CUDA 13 Features | 从朴素到 Tensor Core 的 CUDA 高性能算子优化实验室
machine-learning deep-learning hpc cuda nvidia high-performance-computing hopper cuda-kernels gpu-computing ampere gemm cpp20 gpu-programming gpu-performance tensor-cores ai-inference tensor-core flashattention kernel-optimization cuda-optimization
-
Updated
Apr 22, 2026 - Cuda
⚡ LLM-Speed: High-performance CUDA kernels for LLM inference — FlashAttention with O(N) memory, Tensor Core GEMM (95% cuBLAS), and seamless PyTorch integration. Supports Volta to Hopper GPUs.
deep-learning cuda matrix-multiplication gpu-acceleration attention cuda-kernels gpu-computing memory-efficient attention-mechanism gemm pybind11 pytorch-extension inference-optimization cuda-programming llm tensor-core llm-inference flashattention kernel-optimization flashattention-2
-
Updated
Apr 22, 2026 - Python
High-performance GPU kernel library for AI: GEMM, FlashAttention, Conv2D, Normalization & Quantization with progressive optimization from Naive to Tensor Core. Modern C++/CUDA, header-only, production-ready.
deep-learning cpp hpc cuda header-only attention cpp17 cuda-kernels gpu-computing quantization gemm normalization cpp20 gpu-programming tensor-cores tensor-core flash-attention cuda-12 ai-kernels
-
Updated
Apr 22, 2026 - C++
The lab assignments from CS4302 Parallel and Distributed Programming (2022 Fall) with my solutions
-
Updated
Jan 5, 2023 - C++
A reproducible GPU benchmarking lab that compares FP16 vs FP32 training on MNIST using PyTorch, CuPy, and Nsight profiling tools. This project blends performance engineering with cinematic storytelling—featuring NVTX-tagged training loops, fused CuPy kernels, and a profiler-driven README that narrates the GPU’s inner workings frame by frame.
performance-engineering deep-learning reproducible-research cuda pytorch fp16 cupy mixed-precision nsight gpu-benchmark nvtx fp32 tensor-core
-
Updated
Sep 5, 2025 - Python
CUDA Kernel Academy: A systematic learning path for high-performance CUDA kernel development. From SGEMM basics to Tensor Core mastery with 4 progressive modules. | CUDA 高性能算子开发：从 SGEMM 基础到 Tensor Core 精通，4 模块渐进式学习路径
education machine-learning performance tutorial kernel deep-learning cpp hpc optimization cuda inference nvidia gemm sgemm cuda-programming tensor-cores tensor-core flash-attention
-
Updated
Apr 22, 2026 - C++
🎬 Explore GPU training efficiency with FP32 vs FP16 in this modular lab, utilizing Tensor Core acceleration for deep learning insights.
performance-engineering deep-learning reproducible-research cuda pytorch fp16 cupy mixed-precision nsight gpu-benchmark nvtx fp32 tensor-core
-
Updated
Feb 20, 2026 - Python
🚀 High-Performance CUDA C++ LLM Inference Engine | W8A16 Quantization | KV Cache | ~50% Memory Reduction | Zero Dependencies
machine-learning cmake deep-learning cpp high-performance cuda inference nvidia transformer cpp17 cuda-kernels gpu-computing quantization inference-engine kv-cache int8-quantization llm tensor-core llm-inference w8a16
-
Updated
Apr 22, 2026 - C++
Improve this page
Add a description, image, and links to the tensor-core topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the tensor-core topic, visit your repo's landing page and select "manage topics."