Same Performance as Group-Wise Weight-Only Quantization But with Better Accuracy

Calling cuBLAS GEMM API Correctly

Playing Videos with GPU Acceleration in SMPlayer

CuTe Shared Memory Swizzling Abstractions

Matrix Transpose CUDA Kernel Implementation Using CuTe

Employing CUTLASS for Accelerated Computing

Mathematical Fundamentals to CUTLASS Computing

CUDA Reduction Using Cooperative Groups As An Example

Parallel Reduction CUDA Implementations

Dealing With CUDA Shared Memory Bank Conflicts Using Swizzling

Portable TensorRT

TensorRT Custom Plugin Implementation and Integration