I am trying to compile a simpler wmma kernel with FP8 data type on H100. Following is my example code, GPU: H100 CUDA version 12.0 gcc version 11.3 header file #include <cuda.h> #include <mma.h> #include <cuda_fp8.h> #include <cuda_fp8.hpp> using namespace nvcuda; /* Performs a 16x16x16 warp m…

FP8 WMMA kernel compilation error

Robert_Crovella March 26, 2023, 7:05pm 10

You can file a bug. If you don’t provide a complete repro case, you will probably be asked for one.

Topic		Replies	Views
cuBLAS INT8 tensor core mode vs. FP16 mode GPU-Accelerated Libraries cublas	13	5496	December 5, 2022
cublasGemmEx doesn't work with INT8 utilizing __dp4a instruction on NVIDIA 1080TI CUDA Programming and Performance	12	3656	September 25, 2017
How to use WMMA efficiently CUDA Programming and Performance	4	8146	October 23, 2020
CUDA 2.1 Beta Problem/Bugs (Linux) CUDA Programming and Performance	5	1656	January 6, 2009
Error while using different precisions with tensor cores Deep Learning (Training & Inference) mixed-precision	2	1982	July 26, 2019
Struct in CUDA can i use this struct in CUDA CUDA Programming and Performance	15	89487	June 26, 2009
cublasGemmEx failed: 15, ieee_inexact in Fortran and C++ GPU-Accelerated Libraries cuda	4	60	February 28, 2025
cuBLAS INT8 tensor core mode vs. FP16 mode GPU-Accelerated Libraries	0	886	February 15, 2019
cuBLAS GEMM INT8 is much slower than FP16 in T4 GPU-Accelerated Libraries cublas	11	4428	November 2, 2023
Memory Corruption on a Fermi-Class GPU Error only on Fermis, program works on non-Fermis. CUDA Programming and Performance	18	7134	July 22, 2011