Commit 43fdac63 authored by Tomáš Oberhuber's avatar Tomáš Oberhuber
Browse files

Fixed bug in very slow 3D boundary conditions in CUDA.

parent 5b742b0a

src/TNL/Meshes/GridDetails/GridTraverser_impl.h

+16 −16
Original line number Diff line number Diff line
@@ -840,6 +840,7 @@ processEntities( 
   const int& stream,

   const GridEntityParameters&... gridEntityParameters )

{

   // TODO: Fix the grid sizes inside the for loops -- even in 2D

#ifdef HAVE_CUDA   

   if( processOnlyBoundaryEntities && 

      ( GridEntity::getDimensions() == 3 || GridEntity::getDimensions() == 0 ) )
@@ -848,17 +849,15 @@ processEntities( 
      const IndexType entitiesAlongX = end.x() - begin.x() + 1;

      const IndexType entitiesAlongY = end.y() - begin.y() + 1;

      const IndexType entitiesAlongZ = end.z() - begin.z() + 1;

      const IndexType entitiesAlongXY = entitiesAlongX * entitiesAlongY;

      const IndexType entitiesAlongXZ = entitiesAlongX * ( entitiesAlongZ - 2 );

      const IndexType entitiesAlongYZ = ( entitiesAlongY - 2 ) * ( entitiesAlongZ - 2 );

      

      dim3 cudaBlocksAlongXY, cudaBlocksAlongXZ, cudaBlocksAlongYZ;

      cudaBlocksAlongXY.x = Devices::Cuda::getNumberOfBlocks( entitiesAlongXY, cudaBlockSize.x );

      cudaBlocksAlongXY.y = Devices::Cuda::getNumberOfBlocks( entitiesAlongXY, cudaBlockSize.y );

      cudaBlocksAlongXZ.x = Devices::Cuda::getNumberOfBlocks( entitiesAlongXZ, cudaBlockSize.x );

      cudaBlocksAlongXZ.y = Devices::Cuda::getNumberOfBlocks( entitiesAlongXZ, cudaBlockSize.y );

      cudaBlocksAlongYZ.x = Devices::Cuda::getNumberOfBlocks( entitiesAlongYZ, cudaBlockSize.x );

      cudaBlocksAlongYZ.y = Devices::Cuda::getNumberOfBlocks( entitiesAlongYZ, cudaBlockSize.y );

      cudaBlocksAlongXY.x = Devices::Cuda::getNumberOfBlocks( entitiesAlongX, cudaBlockSize.x );

      cudaBlocksAlongXY.y = Devices::Cuda::getNumberOfBlocks( entitiesAlongY, cudaBlockSize.y );

      cudaBlocksAlongXZ.x = Devices::Cuda::getNumberOfBlocks( entitiesAlongX, cudaBlockSize.x );

      cudaBlocksAlongXZ.y = Devices::Cuda::getNumberOfBlocks( entitiesAlongZ - 2, cudaBlockSize.y );

      cudaBlocksAlongYZ.x = Devices::Cuda::getNumberOfBlocks( entitiesAlongY - 2, cudaBlockSize.x );

      cudaBlocksAlongYZ.y = Devices::Cuda::getNumberOfBlocks( entitiesAlongZ - 2, cudaBlockSize.y );

      

      

      const IndexType cudaGridsAlongXY_x = Devices::Cuda::getNumberOfGrids( cudaBlocksAlongXY.x );

      const IndexType cudaGridsAlongXY_y = Devices::Cuda::getNumberOfGrids( cudaBlocksAlongXY.y );
@@ -872,6 +871,11 @@ processEntities( 
      

      const cudaStream_t& s1 = pool.getStream( stream );

      const cudaStream_t& s2 = pool.getStream( stream + 1 );

      const cudaStream_t& s3 = pool.getStream( stream + 2 );

      const cudaStream_t& s4 = pool.getStream( stream + 3 );

      const cudaStream_t& s5 = pool.getStream( stream + 4 );

      const cudaStream_t& s6 = pool.getStream( stream + 5 );

      

      for( IndexType gridIdx_y = 0; gridIdx_y < cudaGridsAlongXY_y; gridIdx_y++ )

         for( IndexType gridIdx_x = 0; gridIdx_x < cudaGridsAlongXY_x; gridIdx_x++ )

         {
@@ -900,8 +904,6 @@ processEntities( 
                    gridIdx_y,

                    gridEntityParameters... );

         }

      const cudaStream_t& s3 = pool.getStream( stream + 2 );

      const cudaStream_t& s4 = pool.getStream( stream + 3 );

      for( IndexType gridIdx_y = 0; gridIdx_y < cudaGridsAlongXZ_y; gridIdx_y++ )

         for( IndexType gridIdx_x = 0; gridIdx_x < cudaGridsAlongXZ_x; gridIdx_x++ )

         {
@@ -930,8 +932,6 @@ processEntities( 
                    gridIdx_y,

                    gridEntityParameters... );

         }

      const cudaStream_t& s5 = pool.getStream( stream + 4 );

      const cudaStream_t& s6 = pool.getStream( stream + 5 );

      for( IndexType gridIdx_y = 0; gridIdx_y < cudaGridsAlongYZ_y; gridIdx_y++ )

         for( IndexType gridIdx_x = 0; gridIdx_x < cudaGridsAlongYZ_x; gridIdx_x++ )

         {