merge with kernels from MH's master thesis

[LbmBenchmarkKernelsPublic.git] / src / BenchKernelD3Q19ListAaPv.c

diff --git a/src/BenchKernelD3Q19ListAaPv.c b/src/BenchKernelD3Q19ListAaPv.c
index 8ae0c2cd1cf735c6b36990c74050dc0795a20293..4590823190f08e291534f3f76501502cf592dedb 100644 (file)
--- a/src/BenchKernelD3Q19ListAaPv.c
+++ b/src/BenchKernelD3Q19ListAaPv.c
@@ -76,6 +76,8 @@ void FNAME(D3Q19ListAaPvKernel)(LatticeDesc * ld, KernelData * kernelData, CaseD
                KernelStatistics(kd, ld, cd, 0);
        #endif
 
+       X_KERNEL_START(kernelData);
+
        // TODO: outer openmp parallel
 
        for(int iter = 0; iter < maxIterations; iter += 2) {
@@ -139,6 +141,8 @@ void FNAME(D3Q19ListAaPvKernel)(LatticeDesc * ld, KernelData * kernelData, CaseD
 
        } // for (int iter = 0; ...
 
+       X_KERNEL_END(kernelData);
+
 #ifdef VTK_OUTPUT
        if (cd->VtkOutput) {
                kd->PdfsActive = kd->Pdfs[0];
@@ -235,7 +239,7 @@ static void KernelEven(LatticeDesc * ld, KernelData * kernelData, CaseData * cd)
 
        #define I(index, dir)   P_INDEX_3((nCells), (index), (dir))
 
-       for (int index = indexStartVec; index < indexStopVec; index += VSIZE) {
+       for (int index = indexStartVec; index < indexStopVec; index += VSIZE) { // LOOP list-aa-pv-even-vec
 
 
                #define X(name, idx, idxinv, _x, _y, _z)        JOIN(vpdf_,name) = VLDU(&src[I(index, idx)]);
@@ -373,7 +377,7 @@ static void KernelEven(LatticeDesc * ld, KernelData * kernelData, CaseData * cd)
 
        } // loop over fluid nodes
 
-       for (int index = indexStopVec; index < indexStop; ++index) {
+       for (int index = indexStopVec; index < indexStop; ++index) { // LOOP list-aa-pv-even-scalar
 
                #define X(name, idx, idxinv, _x, _y, _z)        JOIN(pdf_,name) = src[I(index, idx)];
                D3Q19_LIST