What's Changed

Pull devcontainer from Docker Hub by @nicolasbock in #709
Set bash shell for devcontainer user by @nicolasbock in #710
Fix CMake warning by @nicolasbock in #711
Add some default extensions to devcontainer by @nicolasbock in #712
Update Conda badges by @nicolasbock in #713
Add missing sync calls after magma functions by @jeanlucf22 in #714
Rocsparse add targeting OLCF Frontier by @mewall in #716
Add script to install rocsparse headers on OLCF machines by @mewall in #717
Fix bug in indent.sh script by @mewall in #722
Fix OMP offload build memory leaks by @mewall in #720
Fix progress rocsparse tests by @mewall in #723
Remove support for ellsort format by @jeanlucf22 in #724
Add doxygen package for API builds by @nicolasbock in #725
Update API documentation by @nicolasbock in #726
Updated tranpose operations to support non-symmetric matrix structure. by @oseikuffuor1 in #727
Implement noinit allocation for dense and distributed2d by @jeanlucf22 in #728
Bug fixes for norm calculation. by @oseikuffuor1 in #729
Remove unused variables, duplicate code by @jeanlucf22 in #730
Reorganize "domain" code by @jeanlucf22 in #732
Separate real offload diagonalize routines. by @jmohdyusof in #733

Full Changelog: v2.3.0...v2.4.0

Contributors

jmohdyusof, nicolasbock, and 3 other contributors

Assets 2

12 Apr 17:59

nicolasbock

v2.3.0

0404165

Release v2.3.0

What's Changed

Also lint the shell scripts by @nicolasbock in #528
Fix element_multiply with magma by @jeanlucf22 in #581
Fix bcast function for dense format with magma by @jeanlucf22 in #586
Update bml_allocate.c by @tokshgithub in #376
Clean up CI tests by @nicolasbock in #587
Add example for assembly generation by @nicolasbock in #588
Simplify BLAS/LAPACK find in cmake by @jeanlucf22 in #583
Add scripts for MAGMA builds on spock and crusher by @mewall in #585
Add BLA_VENDOR to build.sh by @nicolasbock in #589
Fix which line is highlighted by @nicolasbock in #590
Add key by @nicolasbock in #591
Fix MAGMA case for write/read distributed2d by @jeanlucf22 in #593
Remove unnecessary build time checks by @jeanlucf22 in #594
Fix build to use ScaLapack by @jeanlucf22 in #595
Implement rocSOLVER dsyevd for MAGMA build by @mewall in #597
Add build script for summit at OLCF by @jeanlucf22 in #596
Clean up build script with magma on Summit by @jeanlucf22 in #600
Build scripts for crusher offload by @mewall in #601
Clean up c-tests by @jeanlucf22 in #606
Enable use of IBM XL compilers with openMP Offload support by @oseikuffuor1 in #605
Add simple test to verify OpenMP offload by @jeanlucf22 in #607
Added cusparse functionalities for ELLPACK. by @oseikuffuor1 in #609
Add file to run CI on GPU at OLCF by @jeanlucf22 in #611
Only check push events on master branch by @nicolasbock in #614
Intel offload by @jmohdyusof in #610
Add pre-commit hook script by @nicolasbock in #613
Upgrade CI to llvm-12 by @nicolasbock in #615
Upgrade stale action to latest release by @nicolasbock in #616
Enable CI testing on OS X by @nicolasbock in #172
Add BML_VERSION by @nicolasbock in #622
Switch to OpenBLAS on crusher to avoid fortran errors by @mewall in #624
Add sanity check for build.sh options by @nicolasbock in #620
Minor revisions to cuSPARSE code and offload flags by @mewall in #626
Add indented files to gitignore by @nicolasbock in #629
Add build script for GPU offload on Summit by @jeanlucf22 in #630
Fix offload bug in bml_set_diagonal_ellpack() by @mewall in #631
Add less to CI container by @nicolasbock in #636
Strengthen test get_set_diagonal by @jeanlucf22 in #625
Update bml_add_dense_typed.c by @cnegre in #638
Offload mods for crusher development by @mewall in #633
Turn complex code off when BML_COMPLEX off by @jeanlucf22 in #639
Fix no-complex build with MPI by @jeanlucf22 in #641
Fail at build time if magma asked for but not found by @jeanlucf22 in #643
Lessen cmake requirements for non-GPU build by @jeanlucf22 in #640
Add rocSPARSE method for ellpack multiply by @mewall in #645
Store result of failed lint job by @nicolasbock in #648
Fix no-complex build for bml_elemental code by @jeanlucf22 in #649
Support latest magma on crusher by @mewall in #650
Updated deprecated Dockerfile commands by @nicolasbock in #655
Replace _Complex with complex by @nicolasbock in #651
Test with OpenBLAS as an alternative by @nicolasbock in #656
Add missing include statement by @nicolasbock in #658
Optimize openmp offload for rocsparse by @mewall in #657
Look for Lapack functions in OpenBLAS by @jeanlucf22 in #654
Specify magma version to use for CI on Ascent@OLCF by @jeanlucf22 in #659
Remove check for fabs by cmake by @jeanlucf22 in #661
Change macro DO_MPI to BML_USE_MPI by @jeanlucf22 in #666
Couple with ELPA by @jeanlucf22 in #665
Fix some compiler warnings related to calloc by @jeanlucf22 in #670
Remove OpenCL from CMakeLists.txt by @jeanlucf22 in #673
Fix functions/values used for multiply x2 test by @jeanlucf22 in #669
Fix accumulate_offdiag for distributed2d GPU by @jeanlucf22 in #675
Simplify offload code using BML_OFFLOAD_CHUNKS macro by @mewall in #663
Added cusparse capability for bml_transpose_ellpack. by @oseikuffuor1 in #674
Fix tolerance in matrix norm test by @jeanlucf22 in #676
Fix conflict with ELPA macro by @jeanlucf22 in #677
Add MPI to GPU testing by @jeanlucf22 in #678
Clean up OpenMP flags setup in CMakeLists.txt by @jeanlucf22 in #682
Modify macro to please clang compiler by @jeanlucf22 in #680
Fix link line in bml.pc by @jeanlucf22 in #679
Fix cmake ROCm dependencies on OLCF crusher by @mewall in #683
Move CMake package definitions by @nicolasbock in #685
Update container workflow by @nicolasbock in #686
Renamed default image to Bionic by @nicolasbock in #687
Add Docker Hub badge by @nicolasbock in #689
Adjust tolerance float matrix multiply tests by @jeanlucf22 in #690
Add new function to print version number by @jeanlucf22 in #691
Expose bml_print_version to Fortran by @nicolasbock in #694
Print diff if linter failure by @nicolasbock in #696
Fix rocsparse multiply using rocm 5.3 by @mewall in #695
Debug bml_add_ellpack and Fortran build on crusher by @mewall in #700
Dense offload by @jmohdyusof in #701
Add build time option of using syevd by @jeanlucf22 in #702
Split function bml_diagonalize_dense_single/double_real by @jeanlucf22 in #704
Upgrade base image to Focal by @nicolasbock in #688
Minor fix to enable debug build w magma by @jeanlucf22 in #705
Fix memory management bug in rocsparse multiply by @mewall in #706
Add devcontainer definition by @nicolasbock in #707
Add local user to devcontainer by @nicolasbock in #708
Version bump to v.2.3.0 by @nicolasbock in #699

Full Changelog: v2.2.0...v2.3.0

Contributors

jmohdyusof, nicolasbock, and 5 other contributors

Assets 2

28 Jan 18:37

nicolasbock

v2.2.0

80d66a2

Release v2.2.0

Feature updates and bug fixes

What's Changed

Fix dir local variables by @nicolasbock in #529
Modifications to support additional offload builds by @mewall in #527
Enable override of variables for tests by @nicolasbock in #532
added sparse matrix example by @finkeljos in #534
Fix copy and setters by @oseikuffuor1 in #536
Fix paths in tag update script by @nicolasbock in #537
Modification to enable Cray fortran build by @mewall in #538
Fix Summit build script by @jeanlucf22 in #540
Change the way MPI datatype is built for dense case by @jeanlucf22 in #543
Add CUDA libraries when linking if CUDA is found by cmake by @mewall in #544
set_row distributed2d by @jeanlucf22 in #548
Bug fix in transpose for distributed2d by @jeanlucf22 in #549
Add CUDA libraries when linking if CUDA is found by cmake by @mewall in #545
Split container jobs into separate jobs by @nicolasbock in #550
Add Conda badge by @nicolasbock in #552
Update webpage with new badge by @nicolasbock in #553
Add bml_get_deep_type function by @jeanlucf22 in #551
Bug fix and add test for fixed function by @jeanlucf22 in #554
Add functionalities for distributed2d gershgorin by @jeanlucf22 in #555
Switch to non-blocking MPI recv in multiplication by @jeanlucf22 in #556
Use irecv in distributed transpose by @jeanlucf22 in #558
Add get_sparsity function to distributed2d by @jeanlucf22 in #559
Fix issues with ScaLapack call by @jeanlucf22 in #560
Enable testing Fortran interface with MPI by @jeanlucf22 in #561
Minor fixes in outputs for MPI runs by @jeanlucf22 in #563
Fix matrix size for fortran tests by @jeanlucf22 in #564
Add workflow to check for stale issues and PRs by @nicolasbock in #565
Fix permission issue with stale workflow by @nicolasbock in #567
Bug fix in datatype for normalize_distributed2d by @jeanlucf22 in #569
Element-wise matrix multiplication and corresponding tests are added by @ares201005 in #562
Automate API doc building by @nicolasbock in #435
Ignore local worktrees by @nicolasbock in #573
Fixup the style of the rst documentation by @nicolasbock in #574
Add RTD badge by @nicolasbock in #575
Fix for sphinx<2 (as used on RTD) by @nicolasbock in #576
Fix readthedocs build by @nicolasbock in #577
Make list style consistent by @nicolasbock in #578
Upgrade pinned package versions by @nicolasbock in #579
Shorten paths in __FILE__ expansion by @nicolasbock in #370