- Move sole logging tutorial to sys/tutorials
Kokkos: fix a bug in Kokkos::initialize. We need to set args.num_threads
Style: replace "!rank" with "rank == 0"In SF tutorial ex1 (src/vec/is/sf/tutorials/ex1.c), add parentheses around "rank == 0". This is relevant because ! binds more tightly than +, but == binds les
Style: replace "!rank" with "rank == 0"In SF tutorial ex1 (src/vec/is/sf/tutorials/ex1.c), add parentheses around "rank == 0". This is relevant because ! binds more tightly than +, but == binds lesstightly, so the result of the computation would otherwise change.
show more ...
Merge remote-tracking branch 'origin/jose/fix-manpages'Fix manpages to avoid errors reported by doctextSee merge request petsc/petsc!4309
Fix manpages: Input/Output Parameter --> Parameters
g++11: fix -Werror=misleading-indentation
Merge branch 'jacobf/2021-05-18/petscdevice-context' into 'main'Feature: PetscDeviceContextSee merge request petsc/petsc!3994
PetscLogGpuTime: skip event recording when not loggingThe present profiling approach waits eagerly on the event to record itstime, rather than deferring the cudaEventSynchronize until later (orus
PetscLogGpuTime: skip event recording when not loggingThe present profiling approach waits eagerly on the event to record itstime, rather than deferring the cudaEventSynchronize until later (orusing cudaEventQuery) when it isn't an expensive new synchronization. Aproper fix for this will be somewhat involved, but when we aren'tlogging (e.g., no -log_view) then there is no reason to hold upperformance (by making every elementary vector operation blocking).
Added PetscDevice to manage interaction and configuration of available devices.Added PetscDeviceContext class to manage stream interactions for CUDA and HIP devicebackends. The new class is design
Added PetscDevice to manage interaction and configuration of available devices.Added PetscDeviceContext class to manage stream interactions for CUDA and HIP devicebackends. The new class is designed to facilitate a fork-join parallelization model forasynchronous compute.
Merge branch 'barry/2021-07-09/add-dmdavecgetarraydofwrite' into 'main'Add DMDAVecGetArrayDOFWrite() and DMDAVecRestoreArrayDOFWrite()See merge request petsc/petsc!4161
Log movement of scalars between CPU and GPU seperate from vectors and large arraysCommit-type: profiling\spend 15m
Add CSV file format to petsc_tas_analysis.py
Update badsource rules- Zero or more spaces are not allowed for more returning types- zero spaces not allowed for for|if|while
Fix flamegraph output in parallelThe tree structure created by `PetscLogNestedTreeCreate` can be quiteconfusing due to how it tries to resolve event stacks that differ overranks. This meant that
Fix flamegraph output in parallelThe tree structure created by `PetscLogNestedTreeCreate` can be quiteconfusing due to how it tries to resolve event stacks that differ overranks. This meant that when run in parallel we would get interestingentries in the final flame graph with nonsense names like `542 -> 19`.To fix this we simply only add a line to the output file if we 'own' it(this is the case in the vast majority of cases).
Fix ascii_flamegraph outputPreviously the flame graph output would incorrectly output cumulativetime and also would not correctly reproduce the callstack.
Remove all double blank lines from sourceCommit-type: petsc-style/2h
Linter using libClang that checks PetscValidHeaderSpecifcType, PetscValidHeaderSpecific, PetscValidHeader, PetscValid[XXX]Pointer, PetscValidType, PetscCheckSameComm, PetscValidLogicalCollectiveXXX f
Linter using libClang that checks PetscValidHeaderSpecifcType, PetscValidHeaderSpecific, PetscValidHeader, PetscValid[XXX]Pointer, PetscValidType, PetscCheckSameComm, PetscValidLogicalCollectiveXXX for validity. Automatically generates diff files for problems with a known solution.
Use GPU event timer instead of CPU timer- Support CUDA and HIP- Initialize CUDA event timers along with CUDA- Update dev changes- Package-specific includes should not be placed in public include
Use GPU event timer instead of CPU timer- Support CUDA and HIP- Initialize CUDA event timers along with CUDA- Update dev changes- Package-specific includes should not be placed in public include files.
All MPIU_ functions except MPIU_File return MPI error codes for checkingUpdate checkbadSource to find use of CHKERRQ with MPIU_ functionsCommit-type: i.e. error-checking, optimization, bug-fix, p
All MPIU_ functions except MPIU_File return MPI error codes for checkingUpdate checkbadSource to find use of CHKERRQ with MPIU_ functionsCommit-type: i.e. error-checking, optimization, bug-fix, portability-fix, testing-fix, style-fix, feature, documentation, exampleFunded-by:Project:Time: hoursReported-by:Thanks-to:Development Tools: Vim, Emacs, Eclipse
CI: fix compile warnings when built with --with-debugging=0 COPTFLAGS=-O3
Add flamegraph generation option to logging outputThis commit also adds a Python script to convert the XML logging outputinto the same format.
F90: Encapsulate the usage of MPI in petsc.mod and import only needed definitions into other petsc modules.This requires an explicit use of petsc.mod [or mpi.mod or mpi_f18.mod] from user code
CHKERRQ() -> CHKERRMPI()
Convert MPI error type to PETSc error with string message for all MPI callsNow PETSc examples will ONLY return PETSc error codes and never MPI error codes directly so we can understand and post-pro
Convert MPI error type to PETSc error with string message for all MPI callsNow PETSc examples will ONLY return PETSc error codes and never MPI error codes directly so we can understand and post-process their errors better.The test harness will now automatically retry tests that fail with MPI, this may help with Intel MPI that produces seemingly random failures.Commit-type: error-checking/spend 30m
Do not log KSPSolve() and PCApply() from in PCSetUp()Do not log KSPSetUp() from within KSPSolve()Commit-type: logging/spend 90mReported-by: Stefano Zampini <stefano.zampini@gmail.com>
12345678910>>...20