I have been using nvcc -arch=sm_xx to build my executable ( GitHub - fangq/mcx: Monte Carlo eXtreme (MCX) - Physically accurate and validated GPU ray-tracer ) for releases/deployment in the past 15 years. When setting the sm_xx to be the lowest CC (such as sm_30 for cuda 10) that the installed nvcc …

[image] FangQ: The compiled binary runs fine on this node for both GPUs, however, when running the same binary (statically linked with libcudart-static) on a different server, which has a 4090, I got the following error again Totally expected, and you could have run a similar experiment 3 yea…

Perhaps -arch=all is what you are looking for. The downside to using your previous method, “setting the sm_xx to be the lowest CC (such as sm_30 for cuda 10) that the installed nvcc supports”, is that the PTX included in the binary is also for sm_30. This means that when this is executed on a later…

I don’t believe there has been a change in the specified behavior. [image] FangQ: the provided PTX was compiled with an unsupported toolchain. The solution is and always has been to update the driver on that machine. Or as already suggested, in some cases it may be sufficient to build an sm-…

What Robert Crovella said. For what it is worth, what the AI chatbot suggested echoes traditional recommendations regarding the correct way to build “universal” binaries: build a fat binary that contains SASS for all GPU architectures one wishes to support plus PTX for the latest GPU architecture s…

thank you both for your prompt rely. I maintain all the GPU servers used by my group. To avoid the kernel/driver version mismatch, which forces to reboot the servers, I had to lock the driver version (currently, all locked at 535) using apt-mark hold, because of that, I was not able to test this un…

[image] njuffa: For what it is worth, what the AI chatbot suggested echoes traditional recommendations regarding the correct way to build “universal” binaries: build a fat binary that contains SASS for all GPU architectures one wishes to support plus PTX for the latest GPU architecture supported…

thank you again. I think I got it. sorry that I did not read your previous reply more carefully. this is what I learned each new CUDA version requires a min driver version that increases over CUDA releases the PTX built by each CUDA version is also associated with a PTX version number, which also…

CUDA 12/13 `-arch` flag no longer produces "universal" binaries

Accelerated Computing CUDA CUDA Programming and Performance

Topic		Replies	Views
NVIDIA Blackwell and NVIDIA CUDA 12.9 Introduce Family-Specific Architecture Features Technical Blog	0	138	May 1, 2025
Unified binary creation with -tp and NVIDIA HPC SDK 21.2 Legacy PGI Compilers	13	1080	April 23, 2021
Can no longer create backward compatible CUDA binary with Titan V and CUDA 9 CUDA Setup and Installation	4	1133	August 2, 2018
sm_1x vs. compute_1x? CUDA Programming and Performance	9	13071	November 9, 2008
CUDA is cross-nvidia device? CUDA Programming and Performance	1	720	September 1, 2011
future-proof binaries -- nvcc -code and -arch options how to select the best combination of -code an CUDA Programming and Performance	7	8877	November 11, 2009
Keep PTX with CUDA_SEPARABLE_COMPILATION CUDA NVCC Compiler cuda	4	350	October 13, 2025
How to check the Version of GPU to dynamically set '-gencode=arch=compute_?'? CUDA Programming and Performance	19	5021	August 14, 2023
Does the driver JIT support PTX 1.3/2.0 binaries on older cards? If so, how? CUDA Programming and Performance	5	2899	April 24, 2010
Reducing Application Build Times Using CUDA C++ Compilation Aids Technical Blog	1	714	October 31, 2021

CUDA 12/13 `-arch` flag no longer produces "universal" binaries

Related topics