Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
No matching results
An error occurred while fetching branches. Retry the search.
An error occurred while fetching tags. Retry the search.
Created with Raphaël 2.2.014Feb1093230Jan292524231916826Dec2120151413121110765432130Nov2928262221201918171615141312111098765432131Oct3029272524232221191716MpiCommunicator cheks MPI_CUDA awarness only running really in more MPI processes.Added std::flush to TNL_MPI_PRINT to avoid mixed output.Improved macro TNL_MPI_RINT to work even wihout MPI.Avoided duplicity in MPI data type wrapper implementation.Fixed type of verbose in tnl-diff.Fixed getObjectType failure.Adding MPIDimsCreateError exception.Refactoring of DistributedGridIO.Set-up of the distributed grid subdomains overlaps is taking the actualy applied periodic boundary conditions into account.Throwing exception when MPI_Dims_create is going to be called with wrong dimensions set-up.Added #ifdef HAVE_CUDA to Cuda::synchronizeDevice.Fixing setup of grid subdomain overlaps for periodic boundary conditions.Subdomain overlaps for the periodic boundary conditions is set only when periodic neighbor is not the same MPI process.Fixing mesh function evaluator to work with more general RealType.Making function adapter to work with other types than float and double.Fixing DitributedGridIO to work with general mesh functions.Refactoring of DistributedGridIO_MeshFunction.Added int and general type to MPITypeResolver.Fixing mesh function evaluator to work with more general RealType.Making function adapter to work with other types than float and double.Fixing data types in distributed grid IO.Additional fix of tnl-view for support of int and bool type in mesh functions.tnl-view can convert mesh functions holding even int and bool.Fixing DitributedGridIO to work with general mesh functions.tnl-diff exact-match tells coordinates of mismatches.Merge branch 'traversers-optimizations' into 'develop'Merge branch 'traversers-optimizations' of mmg-gitlab.fjfi.cvut.cz:mmg/tnl-dev into traversers-optimizationsRemoved TARGET_LINK_LIBRARIES and SET_TARGET_PROPERTIES from CMakeLists.txt in the benchmark of traversersFixing benchmarks for traversers with BCRemoved useless HAVE_CUDA from tnl-benchmark-traversersDisabled unused parameters in tnl-benchmark-traversersFixing benchmarksTurned off the build of traverser benchmark until its splitted into several files.Added build parameter --with-profiling.Added check of the benchmark results.Benchmarks: set minTime = 0.0 by default due to backwards compatibilityAdded simple cell test.Traversers benchmark is using asynchronous parallel for.Added asynchronous mode to ParallelFor.MeshFunction refactoring.