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 2.2.02Apr31Mar2726252419171615131110987628Feb262524111026Jan2015109865231Dec3029232017161410989876223Nov2120171596429Oct28161514130Sep29Implement benchmarks for naive kernelImplement naive 3D kernelImplement naive 2D kernelImplement naive convolution for 1D kernelAdded assertions to getView/getConstView methods in Array/ArrayView/Vector/VectorView.Fixed data corruption error in the Array move-constructorMerge branch 'JK/MPI' into 'develop'Added function MPI::optimizeRanks and an example how to use itAllowed passing custom communicators to resolveAndLoadDistributedMesh and loadDistributedMeshCleaned up header files in Merson, StaticMerson, Euler and StaticEulerBreak up cyclic dependency between header files in TNL/MPIReplaced MPI_Comm with MPI::Comm in distributed data structuresAdded RAII wrapper for MPI communicatorsSimplified MPI::getRankOnNode with MPI_INFO_NULLRefactor handling of negative tags in DistributedNDArraySynchronizer to allow skipping communication over the periodic boundaryRefactor operators for SyncDirection in DistributedNDArraySynchronizerMerge branch 'TO/petsc' into 'develop'Disabling cmake test for PETSc in benchmarks.Fixed __launch_bounds__ parameters for Ampére 8.6 architectureFixed initialization of threadsPerSegment in CSRHybridKernel to avoid compiler warningsMerge branch 'JK/doc' into 'develop'Removed old copyright header from CMakeLists.txtRemoved CPack configuration from CMakeLists.txtDocumentation: added section on code linting to CONTRIBUTING.mdDocumentation: added section on code formatting to CONTRIBUTING.mdDocumentation: added interactive rebase to CONTRIBUTING.mdNDArray: fixed subarrays for arrays with static sizesGEM: fixed usage of MatrixReader::readMtxGEMGEMGEM: fixed usage of parseCommandLineGEM can now calculate with multiple grid (matrix size 8000 rows +), better matrix reading and memory operations implemented. Text almost done.MPI device to device transfare for pivoting GEM done.MPI device to device for non-pivoting GEM done!MPI with device to device for non-pivoting gem done.Mapping of threads in gems (MPI, noMPI) changed.Went back in commits for old GPU code. Now MPI turns on MPI code and without MPI turns on old GPU code.Repaired double problem. MPI works on GPU but it is slow.Repaired without MPI witout pivoting.Readable code, MPI pivoting working, makefile.inc switch WITH_MPI yes/no madePivoting with MPI done, unreadable code!Error repaired, cuda MPI without pivoting working