-
Improve implementation of NeuBndCond, DirBndCond, RobBndCond, and AddTraceIntegral and tidy LoopExecution 5 of 5 checklist items completed
-
Tidy execution model and uniformise template parameter 5 of 5 checklist items completed
-
Tidy CUDA implementation and update const auto consistency 5 of 5 checklist items completed
-
Add SYCL BwdTrans Operator to redesign 5 of 5 checklist items completed
-
Allow wrapper array around a existing raw pointer 6 of 7 checklist items completed!1848updated
-
Add CUDA gitlab-ci 5 of 5 checklist items completed
-
Fix multilevel IterativeStaticCond with absolution tolerance 6 of 6 checklist items completed
-
Add missing SYCL backend operators and kernels to allow automatic compilation 5 of 5 checklist items completed
-
-
Replaced MPI_Init with MPI_Init_thread to avoid deadlocks from scotch 7 of 7 checklist items completed
-
Add SYCL backend for AssmbScatr and add generalized atomic function to LoopExecution 5 of 5 checklist items completed
-
Feature/addtraceintegral avx 3 of 7 checklist items completedupdated
-
-
Add Kokkos backend for BwdTrans, IProductWRTBase, IProductWRTDerivBase, and PhysDeriv operators 5 of 5 checklist items completed
-
Fix padding and vector width for device backend and add initial multi-GPU implementation for ConjGrad 5 of 5 checklist items completed
-
Add synthetic turbulence generation for the compressible solver 6 of 7 checklist items completedupdated
-
Fix shared memory for the SYCL backend 4 of 4 checklist items completed
-
Update Cmake for SYCL compilation to use icpx compiler 5 of 5 checklist items completed
-
Draft: Add PhysDerivSumFacKernels and Unittest codes 0 of 7 checklist items completed
- Closed
-
- 9
updated -
Add PhysDeriv SYCL Kernels 5 of 5 checklist items completed
