Jakub Klinkovský pushed to branch main at Arch Linux / Packaging / Packages /
cuda
Commits:
41f7a845 by Jakub Klinkovský at 2026-06-30T22:48:50+02:00
Revert replacing CCCL headers
Reverts 79aec4a48cd58cff99210c055ba5cedcc5cf7c3d and
f6face96eda229de828e1e1575bd11f9d2c2742d
- - - - -
7b4ef69c by Jakub Klinkovský at 2026-06-30T22:53:58+02:00
upgpkg: 13.3.1-1
- - - - -
2 changed files:
- .SRCINFO
- PKGBUILD
Changes:
=====================================
.SRCINFO
=====================================
@@ -1,6 +1,6 @@
pkgbase = cuda
pkgdesc = NVIDIA's GPU programming toolkit
- pkgver = 13.3.0
+ pkgver = 13.3.1
pkgrel = 1
url = https://developer.nvidia.com/cuda-zone
install = cuda.install
@@ -72,10 +72,10 @@ pkgbase = cuda
sha512sums =
5df94583c7d082045fffb5c055b163b5242e7975d4de19e334c2d4b2e0e8caf8a930a5ff82b800f54d45ac53d9be0044afd893fb649b461594981693a5256836
sha512sums =
200b23fa74c486a3b1d003d3326163a7d2318dd3a02a0fe488fef3c4ae713cd0dac251b751edc723696fb3d04d4e5f2c11d3b68a0b5f4621388ca28a387ba421
sha512sums =
0a49ce7972ac24ac1f1bd229a4fad3a37747d3e8123c4c1c4521c19fbc308b85f3ee6bf8fbee420aac2335c57aa6b524c20277c59abd7202590144f7ed79c24a
- source_x86_64 =
https://developer.download.nvidia.com/compute/cuda/13.3.0/local_installers/cuda_13.3.0_610.43.02_linux.run
- sha512sums_x86_64 =
9e882eba72b4a986f9ab25afb9566d54f8d73c88aa0634a6e3cd986f05c4c90fd6c4a8cc55f87fb4cedc0bda3a560079f9856a075c5c61b3b773787cc89792fe
- source_aarch64 =
https://developer.download.nvidia.com/compute/cuda/13.3.0/local_installers/cuda_13.3.0_610.43.02_linux_sbsa.run
- sha512sums_aarch64 =
14dfc609f987177ea4bab2b66468d17cf75c8f340392151352c03843ed17d62d98e416e5b642df21ae6d31a6ef48002f1fda515024d5f70905f44307896fd02b
+ source_x86_64 =
https://developer.download.nvidia.com/compute/cuda/13.3.1/local_installers/cuda_13.3.1_610.43.02_linux.run
+ sha512sums_x86_64 =
76a8297a127d643284e03cabcba075f02b570a6406bbaece303d2c10b09be351ed6794e07c1170e587444576f4809f01ce9d8677972f3477d5255a1f7275b04c
+ source_aarch64 =
https://developer.download.nvidia.com/compute/cuda/13.3.1/local_installers/cuda_13.3.1_610.43.02_linux_sbsa.run
+ sha512sums_aarch64 =
07f8904b9cf94e62260f363f43a3ea12011e5b8715c967be39f3655fef54ed3675b2673c18e4529ac6cfa64c8ae461ad8125979ef265004aceca0d0a435397b6
pkgname = cuda
optdepends = gdb: for cuda-gdb
=====================================
PKGBUILD
=====================================
@@ -4,7 +4,7 @@
pkgname=(cuda)
pkgbase=cuda
-pkgver=13.3.0
+pkgver=13.3.1
# Before upgrading, make sure that we ship at least this version of
# nvidia-utils as otherwise we'll get stuff such as #7.
_driverver=610.43.02
@@ -83,8 +83,8 @@
sha512sums=('ff6ee01a6838e51804911c50f817c8e1545d71aa1b34cd937a546d42642097f81d5
'5df94583c7d082045fffb5c055b163b5242e7975d4de19e334c2d4b2e0e8caf8a930a5ff82b800f54d45ac53d9be0044afd893fb649b461594981693a5256836'
'200b23fa74c486a3b1d003d3326163a7d2318dd3a02a0fe488fef3c4ae713cd0dac251b751edc723696fb3d04d4e5f2c11d3b68a0b5f4621388ca28a387ba421'
'0a49ce7972ac24ac1f1bd229a4fad3a37747d3e8123c4c1c4521c19fbc308b85f3ee6bf8fbee420aac2335c57aa6b524c20277c59abd7202590144f7ed79c24a')
-sha512sums_x86_64=('9e882eba72b4a986f9ab25afb9566d54f8d73c88aa0634a6e3cd986f05c4c90fd6c4a8cc55f87fb4cedc0bda3a560079f9856a075c5c61b3b773787cc89792fe')
-sha512sums_aarch64=('14dfc609f987177ea4bab2b66468d17cf75c8f340392151352c03843ed17d62d98e416e5b642df21ae6d31a6ef48002f1fda515024d5f70905f44307896fd02b')
+sha512sums_x86_64=('76a8297a127d643284e03cabcba075f02b570a6406bbaece303d2c10b09be351ed6794e07c1170e587444576f4809f01ce9d8677972f3477d5255a1f7275b04c')
+sha512sums_aarch64=('07f8904b9cf94e62260f363f43a3ea12011e5b8715c967be39f3655fef54ed3675b2673c18e4529ac6cfa64c8ae461ad8125979ef265004aceca0d0a435397b6')
prepare() {
sh cuda_${pkgver}_${_driverver}_linux*.run --target "${srcdir}" --noexec
@@ -125,6 +125,13 @@ build() {
rm -r "$lib"
done
+ # Replace CCCL headers with symlinks to the cccl package
+ rm -r "${_prepdir}"/opt/cuda/targets/x86_64-linux/include/cccl/
+ ln -s /usr/include/cccl
"${_prepdir}"/opt/cuda/targets/x86_64-linux/include/cccl
+ rm -r
"${_prepdir}"/opt/cuda/targets/x86_64-linux/lib/cmake/{cccl,cub,libcudacxx,thrust}
+ rmdir "${_prepdir}"/opt/cuda/targets/x86_64-linux/lib/cmake/
+ ln -s /usr/lib/cmake "${_prepdir}"/opt/cuda/targets/x86_64-linux/lib/cmake
+
# Fix location of CUPTI headers and libs, otherwise some software fails to
build
#
https://gitlab.archlinux.org/archlinux/packaging/packages/tensorflow/-/work_items/19#note_273055
mv "${_prepdir}"/opt/cuda/extras/CUPTI/include/*
"${_prepdir}"/opt/cuda/include/
View it on GitLab:
https://gitlab.archlinux.org/archlinux/packaging/packages/cuda/-/compare/79aec4a48cd58cff99210c055ba5cedcc5cf7c3d...7b4ef69c5f5f0adf5c197ce6db21e242d4d0e188
--
View it on GitLab:
https://gitlab.archlinux.org/archlinux/packaging/packages/cuda/-/compare/79aec4a48cd58cff99210c055ba5cedcc5cf7c3d...7b4ef69c5f5f0adf5c197ce6db21e242d4d0e188
You're receiving this email because of your account on gitlab.archlinux.org.
Manage all notifications: https://gitlab.archlinux.org/-/profile/notifications
| Help: https://gitlab.archlinux.org/help
