Skip to content

Add PTX helpers #686

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions cuda_bindings/cuda/bindings/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -2,3 +2,4 @@
# SPDX-License-Identifier: LicenseRef-NVIDIA-SOFTWARE-LICENSE

from cuda.bindings._version import __version__
from cuda.bindings import utils
131 changes: 131 additions & 0 deletions cuda_bindings/cuda/bindings/utils/__init__.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,131 @@
# Copyright 2025 NVIDIA Corporation. All rights reserved.
# SPDX-License-Identifier: LicenseRef-NVIDIA-SOFTWARE-LICENSE

import re


# Mapping based on the official PTX ISA <-> CUDA Release table
# https://docs.nvidia.com/cuda/parallel-thread-execution/#release-notes-ptx-release-history
_ptx_to_cuda = {
"1.0": (1, 0),
"1.1": (1, 1),
"1.2": (2, 0),
"1.3": (2, 1),
"1.4": (2, 2),
"2.0": (3, 0),
"2.1": (3, 1),
"2.2": (3, 2),
"2.3": (4, 0),
"3.0": (4, 1),
"3.1": (5, 0),
"3.2": (5, 5),
"4.0": (6, 0),
"4.1": (6, 5),
"4.2": (7, 0),
"4.3": (7, 5),
"5.0": (8, 0),
"6.0": (9, 0),
"6.1": (9, 1),
"6.2": (9, 2),
"6.3": (10, 0),
"6.4": (10, 1),
"6.5": (10, 2),
"7.0": (11, 0),
"7.1": (11, 1),
"7.2": (11, 2),
"7.3": (11, 3),
"7.4": (11, 4),
"7.5": (11, 5),
"7.6": (11, 6),
"7.7": (11, 7),
"7.8": (11, 8),
"8.0": (12, 0),
"8.1": (12, 1),
"8.2": (12, 2),
"8.3": (12, 3),
"8.4": (12, 4),
"8.5": (12, 5),
"8.6": (12, 7),
"8.7": (12, 8),
"8.8": (12, 9),
}


def get_minimal_required_driver_ver_from_ptx_ver(ptx_version: str) -> int:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Instead of driver_ver should we call this cuda_ver? Since someone could be using nvjitlink instead?

"""
Maps PTX ISA version to the minimal CUDA driver version.

Parameters
----------
ptx_version : str
PTX ISA version as a string, e.g. "8.8" for PTX ISA 8.8. This is the ``.version``
directive in the PTX header.

Returns
-------
int
Minimal CUDA driver version as 1000 * major + 10 * minor, e.g. 12090 for CUDA 12.9.

Raises
------
ValueError
If the PTX version is unknown.

Examples
--------
>>> get_minimal_required_driver_ver_from_ptx_ver("8.8")
12090
>>> get_minimal_required_driver_ver_from_ptx_ver("7.0")
11000
"""
try:
major, minor = _ptx_to_cuda[ptx_version]
return 1000 * major + 10 * minor
except KeyError:
raise ValueError(f"Unknown or unsupported PTX ISA version: {ptx_version}")



# Regex pattern to match .version directive and capture the version number
_ptx_ver_pattern = re.compile(r'\.version\s+([0-9]+\.[0-9]+)')
Comment on lines +89 to +90
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe a future note that this will slow down import time which some people care about and that we could potentially lazily initialize this instead if needed?



def get_ptx_ver(ptx: str) -> str:
"""
Extract the PTX ISA version string from PTX source code.

Parameters
----------
ptx : str
The PTX assembly source code as a string.

Returns
-------
str
The PTX ISA version string, e.g., "8.8".

Raises
------
ValueError
If the .version directive is not found in the PTX source.

Examples
--------
>>> ptx = r'''
... .version 8.8
... .target sm_86
... .address_size 64
...
... .visible .entry test_kernel()
... {
... ret;
... }
... '''
>>> get_ptx_ver(ptx)
'8.8'
"""
m = _ptx_ver_pattern.search(ptx)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Based on https://docs.nvidia.com/cuda/parallel-thread-execution/#source-format, the first line is always going to be .version, so we'd probably be better off avoiding the regular expression and always just look at the beginning of the string?

if m:
return m.group(1)
else:
raise ValueError("No .version directive found in PTX source. Is it a valid PTX?")