tensor-core

Several optimization methods of half-precision general matrix multiplication (HGEMM) using tensor core with WMMA API and MMA PTX instruction.

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.

Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.

Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.

One-click Windows installer that unlocks the FP32-accumulate tensor-core path of GeForce RTX 40-series GPUs before the driver loads (UEFI boot program + GUI).

Pre-driver FMLA32 Tensor Core unlocker and PTX benchmark console for GeForce RTX 40-series, RTX 30-series and RTX 20-series on Windows 10/11 (UEFI USB) and Linux (gpu_unlock.ko), with friendly UI for Win/Linux. Restores full 2x FP16/BF16/TF32->FP32 accumulate rate with automatic hardware profiling, pre-flight ROM checks, and zero VBIOS flashing.

Volatile UEFI tensor-rate unlock for RTX 2080 Ti 22GB modded cards (TU102 1e04) on Windows: pre-OS SS0/SS1 unlock, no driver patching, cold-boot revert

Multiple GEMM operators are constructed with cutlass to support LLM inference.

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.

Use tensor core to calculate back-to-back HGEMM (half-precision general matrix multiplication) with MMA PTX instruction.

CUDA GEMM 优化演示，从入门版到超越 cuBLAS，从 SIMT 到 Tensor OP，共22版本，系统的逐知识点演进，含 ncu 性能分析。

可验证的 CUDA 学习主线：SGEMM、通用 GPU 算子、性能优化与轻量推理组件

DeepGEMMPerTensor: clean and efficient FP8 GEMM per tensor kernels without scales

⚡ 3-Million-Gate Blackwell-Class GPU Architecture | 57 SystemVerilog Modules | Yosys Verified | Designed by Adhiraj, India 🇮🇳

Pure C++20 reimplementation of NumPy 2.2 — 760+ routines, header-only, SIMD, lattice/padic/neuromorphic/HBM/tensor/memristor/photonics/quantum/accelerator

GPUUnlockInspector · 显卡算力解锁检测器 — 启动即实测 TFLOPS 与型号解锁后参考值比对，判定 NVIDIA 显卡算力是否已解锁（中英双语 / 单文件便携 / 306 项自检）

Codes for DTC-SpMM (ASPLOS'24)

CUDA kernels of sovereign-kimi, ported to Free Pascal — training compute on CPU with no GPU, no nvcc, no CUDA toolkit.

MX450 (TU117) has no Tensor Core in silicon — a reproducible reverse-engineering case study: Falcon microcode, register-level proofs, and why every HMMA library is 18x slower. 三语文档。

Unify high-performance FP8/FP4/BF16 GEMM, fused MoE, and LLM tensor core kernels in a lightweight runtime-compiled CUDA library.