Package Details: cuda-10.1 10.1.243-1

Git Clone URL: https://aur.archlinux.org/cuda-10.1.git (read-only, click to copy)
Package Base: cuda-10.1
Description: NVIDIA's GPU programming toolkit
Upstream URL: https://developer.nvidia.com/cuda-zone
Licenses: custom:NVIDIA
Provides: cuda-sdk, cuda-toolkit
Replaces: cuda-sdk, cuda-toolkit
Submitter: tinytangent
Maintainer: tinytangent
Last Packager: tinytangent
Votes: 2
Popularity: 0.000000
First Submitted: 2020-01-01 08:13 (UTC)
Last Updated: 2020-01-01 08:13 (UTC)

Dependencies (6)

Required by (7)

Sources (4)

Latest Comments

ashirviskas commented on 2020-09-07 08:07 (UTC)

Could you remove replaces line like it is in cuda-10.0? Otherwise you can't have multiple CUDA versions on your machine.

boernjamin commented on 2020-03-23 08:18 (UTC) (edited on 2020-03-23 15:21 (UTC) by boernjamin)

Hello,

I have the same problem. Currently nvidia-440-xx-utils are installed but tried it also with nvidia-418-xx-utils. The output I get is the following:

Preparing...
Cloning cuda-10.1 build files...
Checking cuda-10.1 dependencies...
Resolving dependencies...
Checking inter-conflicts...

Building cuda-10.1...
==> Making package: cuda-10.1 10.1.243-1 (Mo 23 Mär 2020 08:44:42 CET)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Found cuda_10.1.243_418.87.00_linux.run
  -> Found cuda.sh
  -> Found cuda.conf
  -> Found cuda-findgllib_mk.diff
==> Validating source files with sha512sums...
    cuda_10.1.243_418.87.00_linux.run ... Passed
    cuda.sh ... Passed
    cuda.conf ... Passed
    cuda-findgllib_mk.diff ... Passed
==> Removing existing $srcdir/ directory...
==> Extracting sources...
==> Starting prepare()...
cuda_10.1.243_418.87.00_linux.run: line 441: exec: -t: invalid option
exec: usage: exec [-cl] [-a name] [command [arguments ...]] [redirection ...]
==> ERROR: A failure occurred in prepare().
    Aborting...

thanks for helping

-------------------------

tl;dr: install the package over the cli

Solution to this specific problem:

I am using manjaro and because i am relatively new to Linux i was using the GUI version of pamac.

As layed out in this forum-post (https://forum.manjaro.org/t/exec-t-invalid-option-while-building-cuda-8-0-and-cuda-9-2-with-pamac/63876) the error is caused by the gui version because it wants to open xterm. After installing Xterm i was running in other problems.

The final solution was to install the package over the cli with the command

pamac build cuda-10.1

tinytangent commented on 2020-01-18 01:12 (UTC)

@creat89 Indeed these options do not appear in the output of sh cuda_10.1.243_418.87.00_linux.run --help

But they are from the PKGBUILD in previous CUDA-10.1 community repo https://git.archlinux.org/svntogit/community.git/plain/trunk/PKGBUILD?h=packages/cuda&id=fcbd4c73582f2c4ddd0e271388b864050811e232

I guess they may be deprecated but still supported.

Seems the PKGBUILD works on my side. Could you please paste some error messages you encounter here?

tinytangent commented on 2020-01-15 02:46 (UTC)

@creat89

Thank you for your feedback. Actually this PKGBUILD is modified from the official community repo without careful checks so there could be some mistakes. I will look into this problem.

creat89 commented on 2020-01-14 17:03 (UTC)

Hello,

The code of the PKGBUILD doesn't work without some modifications. I haven't arrived to install completely the package. But as far as I can tell line 29 should be modified, as for the current version of CUDA, there is no option --target or --no-exec.