merge with kernels from MH's master thesis

[LbmBenchmarkKernelsPublic.git] / src / BenchKernelD3Q19AaVec.c

diff --git a/src/BenchKernelD3Q19AaVec.c b/src/BenchKernelD3Q19AaVec.c
index f79e6cd5495495808ddb63723677e266acc68937..f67bccbfa57065c8b70eaf06eb2928771a85234b 100644 (file)
--- a/src/BenchKernelD3Q19AaVec.c
+++ b/src/BenchKernelD3Q19AaVec.c
@@ -123,6 +123,8 @@ void FNAME(D3Q19AaVecKernel)(LatticeDesc * ld, KernelData * kd, CaseData * cd)
 
        Assert((maxIterations % 2) == 0);
 
+       X_KERNEL_START(kd);
+
        for (int iter = 0; iter < maxIterations; iter += 2) {
 
                // --------------------------------------------------------------------
@@ -215,6 +217,8 @@ void FNAME(D3Q19AaVecKernel)(LatticeDesc * ld, KernelData * kd, CaseData * cd)
 
        } // for (int iter = 0; ...
 
+       X_KERNEL_END(kd);
+
        #ifdef VTK_OUTPUT
 
        if (cd->VtkOutput) {
@@ -331,7 +335,7 @@ static void KernelEven(LatticeDesc * ld, KernelData * kd, CaseData * cd) // {{{
 
                for (int x = bX; x < eX; x += 1) {
                for (int y = bY; y < eY; y += 1) {
-               for (int z = bZ; z < eZ; z += VSIZE) {
+               for (int z = bZ; z < eZ; z += VSIZE) { // LOOP aa-vec-even
 
                        #define I(x, y, z, dir) P_INDEX_5(gDims, (x), (y), (z), (dir))
 
@@ -523,7 +527,7 @@ static void KernelOdd(LatticeDesc * ld, KernelData * kd, CaseData * cd)  // {{{
 
                for (int x = bX; x < eX; ++x) {
                for (int y = bY; y < eY; ++y) {
-               for (int z = bZ; z < eZ; z += VSIZE) {
+               for (int z = bZ; z < eZ; z += VSIZE) { // LOOP aa-vec-odd
 
                        #define I(x, y, z, dir) P_INDEX_5(gDims, (x), (y), (z), (dir))