From 84786ebecd08ca6bf61ebda039ffc608a7f88083 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Dani=C3=ABl=20de=20Kok?= Date: Sun, 14 Feb 2021 11:15:02 +0100 Subject: [PATCH] cudnn_cudatoolkit_11: remove libcudnn_cnn_infer.so exception With the update to CuDNN 8.1, we do not get the "maximum file size exceeded" error anymore when patchelf'ing libcudnn_cnn_infer.so. This change removes the exception for that library. --- pkgs/development/libraries/science/math/cudnn/generic.nix | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) diff --git a/pkgs/development/libraries/science/math/cudnn/generic.nix b/pkgs/development/libraries/science/math/cudnn/generic.nix index 59c0b7f4498..e5f09f142dd 100644 --- a/pkgs/development/libraries/science/math/cudnn/generic.nix +++ b/pkgs/development/libraries/science/math/cudnn/generic.nix @@ -38,13 +38,7 @@ stdenv.mkDerivation { # See the explanation in addOpenGLRunpath. postFixup = '' for lib in $out/lib/lib*.so; do - # patchelf fails on libcudnn_cnn_infer due to it being too big. - # Most programs will still get the RPATH since they link to - # other things. - # (https://github.com/NixOS/patchelf/issues/222) - if [ "$(basename $lib)" != libcudnn_cnn_infer.so ]; then - addOpenGLRunpath $lib - fi + addOpenGLRunpath $lib done '';