We work every day to make parallel programming faster and easier, so we are excited to introduce you to the new features of Parallelware Trainer 0.5 version. Take a look at the big changes!
- User can now select among three implementations of parallel scalar/sparse reductions:
- Built-in reductions (e.g. OpenMP/OpenACC reduction clause)
- Atomic access (e.g. OpenMP/OpenACC atomic construct)
- Explicit privatization and synchronization of reduction variables (e.g. combination of OpenMP private, barrier and critical/atomic)
- Improved GUI with new User Action LIst to support OpenMP/OpenACC-directive parallel templates
- New follow-on actions to fill in the template (e.g. CPU-GPU transfer ranges using OpenMP map to/map from and OpenACC copyin/copyout).
- New requirements to guarantee correctness (e.g. assert non-aliasing with restrict)
- New suggestions to improve performance (e.g. remove atomic if not needed)
- Improved GUI with a new dialog window for the creation of Parallelware pragmas
- Interactive on-the-fly syntax validation
- Clang-based reengineering of the Parallelware driver for C/C++
- GCC/Clang-compatible compilation flags (e.g. -D ,-I, -std)
- Clang-based reengineering of Parallelware front-end for C/C++
- Improved support for preprocessor macros
- Improved validation of Parallelware pragma parameters
- New documentation for the specification of Parallelware pragmas
- Bugfixes in GUI
You can try Parallelware Trainer for free 15 days.
Users will have full and free access to the tool for 15 days and will be able to learn parallel programming while improving their code.
✉ Get all our updates by subscribing to our newsletter.