thanks for the kernel nomenclatures
CUDA Ontology
11–20 of 44 posts
Re: CUDA Ontology
#12Wondering why a $4T company can't afford a smart installation assistant that can auto-detect problems and apply fixes as needed. I wasted too many days chasing driver and torch versions. It's probably the worst part of working in ML. Combine this with Python's horrible package management and you got a perfect combo - like the cough and the stitch.
They provide containers to cater to those needs: https://catalog.ngc.nvidia.com/search
Containers don't include drivers which is the primary reason for issues.
Re: CUDA Ontology
#13Earlier quoted context omitted.
They provide containers to cater to those needs: https://catalog.ngc.nvidia.com/search
Containers don't include drivers which is the primary reason for issues.
Containers afair rely on the exact driver version matching between the host system and the container itself.
We were on AWS when we used this so setting up seemed easy enough - AWS gave you the driver, and a matching docker image was easy enough to find.
Re: CUDA Ontology
#14This is a good resource. But for the computer vision and machine learning practitioner most of the fun can start where this article ends. nvcc from the CUDA toolkit has a compatibility range with the underlying host compilers like gcc. If you install a newer CUDA toolkit on an older machine, likely you'll need to upgrade your compiler toolchain as well, and fix the paths. While orchestration in many (research) projec…
Sounds like most of these problems come from using Python.
Re: CUDA Ontology
#15For reference: CUDA means "Compute Unified Device Architecture".
Re: CUDA Ontology
#16This is a good resource. But for the computer vision and machine learning practitioner most of the fun can start where this article ends. nvcc from the CUDA toolkit has a compatibility range with the underlying host compilers like gcc. If you install a newer CUDA toolkit on an older machine, likely you'll need to upgrade your compiler toolchain as well, and fix the paths. While orchestration in many (research) projec…
Sounds like most of these problems come from using Python.
You imply these problems would go away (or wouldn't be replaced by new ones) with another language.
Re: CUDA Ontology
#17Wondering why a $4T company can't afford a smart installation assistant that can auto-detect problems and apply fixes as needed. I wasted too many days chasing driver and torch versions. It's probably the worst part of working in ML. Combine this with Python's horrible package management and you got a perfect combo - like the cough and the stitch.
Just have claude code fix it
Re: CUDA Ontology
#18this is fantastic
Re: CUDA Ontology
#19This is a good resource. But for the computer vision and machine learning practitioner most of the fun can start where this article ends. nvcc from the CUDA toolkit has a compatibility range with the underlying host compilers like gcc. If you install a newer CUDA toolkit on an older machine, likely you'll need to upgrade your compiler toolchain as well, and fix the paths. While orchestration in many (research) projec…
> nvcc from the CUDA toolkit has a compatibility range with the underlying host compilers like gcc. If you install a newer CUDA toolkit on an older machine, likely you'll need to upgrade your compiler toolchain as well, and fix the paths.
Conversely, nvcc often stops working with major upgrades of gcc/clang. Fun times, indeed.
This is why a lot of people just use NVIDIA's containers even for local solo dev. It's a hassle to set up initially (docker/podman hell) but all the tools are there and they work fine.
Re: CUDA Ontology
#20Earlier quoted context omitted.
Containers don't include drivers which is the primary reason for issues.
Containers afair rely on the exact driver version matching between the host system and the container itself. We were on AWS when we used this so setting up seemed easy enough - AWS gave you the driver, and a matching docker image was easy enough to find.
That's not the case, CUDA containers user space does not have to match the host drivers CUDA capability. The container needs to be the same major version or lower. So a system with a CUDA 13 capable driver should be able to run all previous versions.
For some versions there's even sometimes compat layers built into the container to allow forward version compatibility.