Package Details: tensorflow-cuda-git 1.12.1+54837+g03796cf1dda-1

Git Clone URL: https://aur.archlinux.org/tensorflow-git.git (read-only, click to copy)
Package Base: tensorflow-git
Description: Library for computation using data flow graphs for scalable machine learning (with CUDA)
Upstream URL: https://tensorflow.org/
Keywords: artificial_intelligence google machine_learning neural_network
Licenses: Apache
Conflicts: tensorflow-cuda
Provides: tensorflow-cuda
Submitter: swiftscythe
Maintainer: None
Last Packager: swiftscythe
Votes: 2
Popularity: 0.000000
First Submitted: 2017-05-27 18:50 (UTC)
Last Updated: 2021-04-14 00:37 (UTC)

Latest Comments

1 2 Next › Last »

MarsSeed commented on 2022-06-18 23:59 (UTC) (edited on 2022-07-05 21:11 (UTC) by MarsSeed)

Bazel's build setup (beginning phase) fails with a lot of ceremony. But the gist of it is this (first line is lol):

Error in fail: 
System Library Configuration Error: Invalid system lib set: pcre

After removing pcre from _set_build_env()'s export TF_SYSTEM_LIBS variable in PKGBUILD, the main Bazel build phase is able to commence.

(Note: tried to set pcre2 but that lead to the same error as with pcre.)

EDIT:

Unfortunately I couldn't finish the build, because I have an old laptop with 11.6 gigs of RAM and this build maxed it out and the processes were killed by the OOM.

swiftscythe commented on 2022-05-21 09:12 (UTC)

@ramatullah I haven't built this package in a long time... But I know that GCC 12.1 has problems with CUDA (I have to use clang for it to work)

Are you interested in maintaining this package?

ramatullah commented on 2022-05-20 12:07 (UTC)

What is the maximum GCC version you've successfully compiled with?

Kamilcuk commented on 2021-04-07 10:53 (UTC)

Umm, the prepare() function is kind-of invalid. You can't set environment variables from prepare(), I mean, you can set it, but you can't expect them to be visible in build() stage. The build fails when prepare() and build() are run in separate stages - like first makepkg -o and then followed by separate makepkg -e - makepkg -e will cause configure.py ask for some settings, because prepare() was not run. The build most notably fails to build properly when using yay AUR helper.

Consider applying like the following trivial patch to just move variable settings to be inside build():

--- i/PKGBUILD
+++ w/PKGBUILD
@@ -42,7 +42,9 @@ prepare() {

   [ -d ${srcdir}/tensorflow-cuda ] && rm -rf ${srcdir}/tensorflow-cuda
     cp -r ${srcdir}/tensorflow ${srcdir}/tensorflow-cuda
+}

+_set_build_env() {
   # make sure the proxy variables are in all caps, otherwise bazel ignores them
   export HTTP_PROXY=`echo $http_proxy | sed -e 's/\/$//'`
   export HTTPS_PROXY=`echo $https_proxy | sed -e 's/\/$//'`
@@ -96,6 +98,7 @@ prepare() {
 }

 build() {
+  _set_build_env
   cd ${srcdir}/tensorflow
   export CC_OPT_FLAGS="-march=native -O3"
   export TF_NEED_CUDA=0

Thanks!

ruestique commented on 2020-05-21 21:54 (UTC) (edited on 2020-05-21 21:54 (UTC) by ruestique)

have to make symlink in /usr/bin

sudo ln -s bazel bazel-3.0.0-linux-x86_64

ruestique commented on 2020-04-03 11:41 (UTC)

How to compile it with modern CPU's instructions support ?

entriphy commented on 2020-01-21 08:37 (UTC)

Had to downgrade from Bazel 2.0.0 to 1.2.1. Otherwise, it fails with a Python "subprocess" error. This is not caused by the PKGBUILD, however.