The twenty-fourth month of work focused on work in the following areas:
- Adding support for 4D operations
- Added a distributed matrix
- Updates to keep Phylanx current with HPX master
- Work on Blaze CUDA
Organizational Activities and Results
- Weekly group-meeting, minutes:
- August 1st: https://drive.google.com/open?id=1wRM31d-o2lz8eCv-tFytFUoXAdFm8F_w
- August 8th: https://drive.google.com/open?id=1-NcnPO9_yX2Ky86wDn5DuIiwz7DnWT_9
- August 14th: https://drive.google.com/open?id=1St2ySIyksRoQGEhZFh3i8mFn1hDTAI1d
- August 21st: https://drive.google.com/open?id=1aEnRGezyC0W9Z18OwsXDmM3yXjTUipWT
- Phylanx Seminars:
Development Activities and Results
- Tiling
- Theory:
- Derived results related to the online minute and the maxcut problems
- Currently looking at some online submodular optimization problems
- Implementation
- Begun working on distributed objects and primitives to test tiling theory
- Theory:
- Primitives & Algorithms
- Blazetensor is now mandatory (PR #1010)
- Distributed matrix has been added (PR #1013)
- Spatial_2d_padding operation has been added (PR #1018)
- Expand_dims now has 4D support (PR #1021)
- 4D support was added to random operation (PR #1024)
- 4D support was added to len operation (PR #1025)
- 4D support was added to extract_shape operation (PR #1029)
Repository Activity August 3rd – September 3rd:
Code statistics
------------------------------------------------------------------------------- Language files blank comment code ------------------------------------------------------------------------------- C++ 475 18551 6067 93478 C/C++ Header 311 6903 4318 33671 Python 137 2577 1795 6796 CMake 109 954 791 4018 YAML 2 24 77 670 Markdown 2 17 0 46 Dockerfile 2 6 28 44 CSS 1 14 10 38 HTML 2 5 6 22 DOS Batch 2 0 0 12 INI 1 0 0 8 ------------------------------------------------------------------------------- SUM: 1044 29051 13092 138803 -------------------------------------------------------------------------------
Impact on Other Projects
- Blaze Tickets:
- Pull Request: Blaze CUDA compatibility: adding “IsCUDAAssignable” typetrait and “BLAZE_HOST_DEVICE” macro (PR #35)
- Blaze Tensor
- Updating 4D arrays (PR #33)
- Fixing things to conform with latest changes in blaze (PR #34)
- Updating quatslice (PR #35)
- Adapting for latest changes in Blaze (PR #37)
- Making sure slice of a tensor transposition gives correct result (PR #38)
- Adding trans operation for 4D arrays (PR #39)
- Adding missing QuatSliceTrait for CustomArray (PR #40)
- Fix compilation problem by removing specialization of member functions. (PR #42)