nixpkgs/pkgs/development/r-modules/patches/WideLM.patch
Bjørn Forsman 22321f2e58 cudatoolkit: don't move $out/include to $out/usr_include
This effectively reverts 86c283824f
("If cuda headers are presented to nix [...]") and all the following
workarounds that was added due to that commit.

As far as I can tell[1] this hack isn't needed anymore. And moving
includes to $out/usr_include causes pain for cudatoolkit users, so
better get rid of it.

In patches that did more than the $out/usr_include workaround, I only
changed the line back to $out/include instead of re-generating the
patches and fully removing the changed line.

[1]: I build tested blender and caffe, and temporarily added
recurseIntoAttrs to rPackages and haskellPackages so that nox-review
could get proper coverage. However, many of the packages do not build
even before this patch. I also built CUDA samples with cudatoolkit7
that ran fine.
2015-09-10 10:35:40 +02:00

20 lines
686 B
Diff

diff -ru -x '*~' WideLM_orig/src/Makefile.in WideLM/src/Makefile.in
--- WideLM_orig/src/Makefile.in 2012-02-17 04:28:05.000000000 +0900
+++ WideLM/src/Makefile.in 2014-10-25 18:54:49.110011921 +0900
@@ -12,12 +12,12 @@
#compiler/preprocessor options
R_INCS := @R_INCL@ @RCPP_INCL@
CC_ARGS := @CU_PARAMS@ -Xcompiler @R_CPIC@
-CU_INCS := -I@CUDA_HOME@/include
+CU_INCS := -I@CUDA_HOME@/include
CU_ARCH := -gencode arch=compute_20,code=sm_20
#linker options
-LD_PARAMS := -Xlinker "@RCPP_LDFLAGS@"
-RCU_LIBS := @R_LIB@ -L@CU_LIBDIR@ -lcublas
+LD_PARAMS := -Xlinker "--export-dynamic -fopenmp -L${R_HOME}/lib -lR"
+RCU_LIBS := -Xlinker "-L@CU_LIBDIR@ -lcublas"
all : WideLM.so