From 5163822e2633e95ee6c716c8c6c22e53bd5e4691 Mon Sep 17 00:00:00 2001 From: Brant Robertson Date: Wed, 17 Jul 2024 16:21:51 -0700 Subject: [PATCH] Removing the gravitational terms does not have an enormous impact. Still spurious heating of gravity is removed from half-step update. Need to make sure this isn't an indexing issue with the half step update. But also removing the DE synchronization after the half-step. --- src/integrators/VL_3D_cuda.cu | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/src/integrators/VL_3D_cuda.cu b/src/integrators/VL_3D_cuda.cu index d67a90643..2f37dc69e 100644 --- a/src/integrators/VL_3D_cuda.cu +++ b/src/integrators/VL_3D_cuda.cu @@ -254,8 +254,9 @@ void VL_Algorithm_3D_CUDA(Real *d_conserved, Real *d_grav_potential, int nx, int GPU_Error_Check(); #endif // MHD +/* // Select and synchronize the half step - // ADDED IN TESTING + // ADDED IN TESTING, what happens without it? #ifdef DE cuda_utilities::AutomaticLaunchParams static const de_select_launch_params(Select_Internal_Energy_3D, n_cells); hipLaunchKernelGGL(Select_Internal_Energy_3D, de_select_launch_params.get_numBlocks(), @@ -266,7 +267,7 @@ void VL_Algorithm_3D_CUDA(Real *d_conserved, Real *d_grav_potential, int nx, int n_fields); GPU_Error_Check(); #endif // DE - +*/ // Step 4: Construct left and right interface values using updated conserved variables #if defined(PLMP) or defined(PLMC) @@ -385,6 +386,8 @@ void VL_Algorithm_3D_CUDA(Real *d_conserved, Real *d_grav_potential, int nx, int // Compute the divergence of Vel before updating the conserved array, this // solves synchronization issues when adding this term on // Update_Conserved_Variables_3D + + // Does this need to use the updated pressure from the half-updated conserved quantities? cuda_utilities::AutomaticLaunchParams static const de_advect_launch_params(Partial_Update_Advected_Internal_Energy_3D, n_cells); hipLaunchKernelGGL(Partial_Update_Advected_Internal_Energy_3D, de_advect_launch_params.get_numBlocks(),