Anders Johansson anjohan

## lmpotf_lib.patch
diff --git a/src/library.cpp b/src/library.cpp
index cdc0a81..2d75d49 100644
--- a/src/library.cpp
+++ b/src/library.cpp
@@ -482,10 +482,10 @@ char *lammps_command(void *handle, const char *cmd)

   BEGIN_CAPTURE
   {
-    if (lmp->update->whichflag != 0)
-      lmp->error->all(FLERR,"Library error: issuing LAMMPS commands "

## clangoffload.md

      
        
          
            
              
              1 file
            
          
          
            
              
              2 forks
            
          
          
            
              
              0 comments
            
          
          
            
              
              6 stars
            
          
        
        
          
              
          
          
            
                anjohan
                / clangoffload.md
            
            
              Created
              December 3, 2020 02:37
            
              
                Build Clang with OpenMP Target Offloading for NVIDIA GPUs
              
          
        
      
        
  
      
    Note to self
This works with Clang 12 on Manjaro, and maybe other Linux versions. CUDA 11.1 is installed with pacman in /opt/cuda.
Download

git clone https://github.com/llvm/llvm-project.git
First build
	diff --git a/src/library.cpp b/src/library.cpp
	index cdc0a81..2d75d49 100644
	--- a/src/library.cpp
	+++ b/src/library.cpp
	@@ -482,10 +482,10 @@ char lammps_command(void handle, const char *cmd)

	BEGIN_CAPTURE
	{
	- if (lmp->update->whichflag != 0)
	- lmp->error->all(FLERR,"Library error: issuing LAMMPS commands "