Add TACUDA package

overlay.nix

@@ -55,6 +55,7 @@ let
     stdenvClangOmpss2Nanos6 = final.stdenv.override { cc = final.clangOmpss2Nanos6; allowedRequisites = null; };
     stdenvClangOmpss2Nodes = final.stdenv.override { cc = final.clangOmpss2Nodes; allowedRequisites = null; };
     stdenvClangOmpss2NodesOmpv = final.stdenv.override { cc = final.clangOmpss2NodesOmpv; allowedRequisites = null; };
+    tacuda = callPackage ./pkgs/tacuda/default.nix { };
     tagaspi = callPackage ./pkgs/tagaspi/default.nix { };
     tampi = callPackage ./pkgs/tampi/default.nix { };
     upc-qaire-exporter = prev.callPackage ./pkgs/upc-qaire-exporter/default.nix { };

pkgs/tacuda/default.nix

@@ -0,0 +1,68 @@
+{ stdenv
+, fetchFromGitHub
+, automake
+, autoconf
+, libtool
+, gnumake
+, autoreconfHook
+, boost
+, cudaPackages
+
+, useGit ? false
+, gitUrl ? "git@gitlab-internal.bsc.es:task-awareness/tacuda/tacuda.git"
+, gitBranch ? "main"
+, gitCommit ? "35234f9445e6149a2bd38d119841e2485d6ee05e"
+}:
+
+let
+
+  release_ver = "2.1.0";
+  release = {
+    version = release_ver;
+    src = fetchFromGitHub {
+      owner = "bsc-pm";
+      repo = "tacuda";
+      rev = release_ver;
+      hash = "sha256-Cj3EiLVJSLvRv0ydeg7Vp4SpkniEqHkcWF+YOJQ8EcM=";
+    };
+  };
+
+  git = rec {
+    version = src.shortRev;
+    src = builtins.fetchGit {
+      url = gitUrl;
+      ref = gitBranch;
+      rev = gitCommit;
+    };
+  };
+
+  source = if (useGit) then git else release;
+
+in
+
+stdenv.mkDerivation {
+  pname = "tacuda";
+  inherit (source) src version;
+
+  enableParallelBuilding = true;
+  separateDebugInfo = true;
+
+  nativeBuildInputs = [
+    autoreconfHook
+    automake
+    autoconf
+    libtool
+    gnumake
+  ];
+
+  patches = [ ./fix_config.patch ];
+
+  configureFlags = [ "--with-cuda-include=${cudaPackages.cudatoolkit}/include" ];
+
+  buildInputs = [
+    boost
+    cudaPackages.libcublas
+    cudaPackages.cuda_cudart
+  ];
+
+}

pkgs/tacuda/fix_config.patch

@@ -0,0 +1,13 @@
+diff --git a/m4/cuda.m4 b/m4/cuda.m4
+index 23f5c94..8f9b534 100644
+--- a/m4/cuda.m4
++++ b/m4/cuda.m4
+@@ -40,7 +40,7 @@ search_libs="cuda cublas cudart"
+ required_libs=""
+ 
+ m4_foreach([function],
+-           [cuInit,
++           [
+ 	    cublasSgemm,
+             cudaStreamCreate,
+             cudaLaunchKernel,