QUDA 0.4.0 release
Closed Apr 3, 2012
100% complete
Here's what I believe we need in place for a 0.40 release:
- Reference clover implementation -> spilled to 0.4.1
- Fix slow performance introduced with multi-dimensional parallelization.
- Make the blas tuning more robust and faster through caching and default parameters.
- Understand instabilities in GCR solver.
- Fix the reference gauge fixed dslash.
- Fix clover …
Here's what I believe we need in place for a 0.40 release:
- Reference clover implementation -> spilled to 0.4.1
- Fix slow performance introduced with multi-dimensional parallelization.
- Make the blas tuning more robust and faster through caching and default parameters.
- Understand instabilities in GCR solver.
- Fix the reference gauge fixed dslash.
- Fix clover code generator for standalone kernel generation.
Any other suggestions?
This milestone is closed.
No open issues remain. View closed issues or see open milestones in this repository.