When I run the following code, using nvcc harta.cu -gencode arch=compute_90a,code=sm_90a -o harta I get the output of: tid: 0, bid: 0, p0[2] = 0.000000 tid: 0, bid: 0, p0[3] = 0.000000 tid: 0, bid: 0, p1[0] = 0.000000 tid: 0, bid: 0, p1[1] = 0.000000 This means that the cuda conversions between fl…

Bug in fp8x2 and fp8x4 converts

Robert_Crovella March 11, 2025, 1:57pm 2

Suggestions:

Topic		Replies	Views
ptx miscompile bug report in cuda8.0 CUDA Programming and Performance	1	441	May 23, 2018
Use_fast_math Difference CUDA Programming and Performance	1	3425	August 8, 2007
bug report for ptx miscompile in cuda8.0 CUDA Programming and Performance	0	307	May 18, 2018
CUDA 11.6.0 with gcc 11.2.1 fails to process system headers included by <functional> CUDA NVCC Compiler	4	6347	May 24, 2022
CUDA Bug report CUDA Programming and Performance	5	5974	October 7, 2009
CUDA Image Processing on TX2, converting NV12 to RGB [TX2, OpenCV] CUDA Programming and Performance	1	2426	August 28, 2018
Nvcc lower version than CUDA causes compiled code runtime error 300 CUDA NVCC Compiler	4	93	September 24, 2024
Compilation broken sign-change-detection code CUDA Programming and Performance	4	4911	January 28, 2011
bug report for nvcc version 10 CUDA Programming and Performance	3	1018	November 5, 2018
nvcc -O3 error Ubuntu 8.10 AMD64/64-bit -O3 error CUDA Programming and Performance	0	4008	January 27, 2009