https://pfspipe.ipmu.jp/jira This file is an XML representation of an issue en-us 8.3.4 803005 13-09-2019 https://pfspipe.ipmu.jp/jira/browse/INFRA-10 Software Development Infrastructure <p>This may or may not be a good idea, but I'd like to think about it.</p> <ul class="alternate" type="square"> <li>Is Blitz++ still maintained?</li> <li>Does it require lapack or blas?</li> <li>Is it fast?</li> <li>Does it play nicely with numpy?</li> </ul> INFRA-10

Convert Blitz++ code to Eigen

Story Major Done Done aritter rhl Wed, 16 Jul 2014 17:44:31 +0000 Thu, 23 Jun 2016 18:19:02 +0000 Thu, 23 Jun 2016 18:19:02 +0000 0 2 <p>Given that STELLA uses Blitz++ as its primary image class we may need to move this up. One possibility would be to keep Blitz++ internally for now, and construct Blitz++ arrays from afw::Images. Note that afwImage supports a getArray() method to recover a strided array, and that Blitz++ can use such arrays in its ctors (<a href="http://ehc.ac/p/blitz/mailman/message/2406262/" class="external-link" rel="nofollow">http://ehc.ac/p/blitz/mailman/message/2406262/</a>)</p> <p>The blitz-intensive functions FiberTrace::SlitFunc and FiberTrace::MkSlitFunc as well as all blitz functions from Math.h and Utils.h needed by them have been moved to MathBlitz.h and UtilsBlitz.h. Many Math.h functions as well as MpFit have been changed to use ndarrays instead of blitz arrays. The remaining source code is now free of blitz. Tests and Tasks have been changed to the new environment.</p> <p>The Issue is ready for review...</p> <p>Everything related to the Blitz++ library has been converted to Eigen or ndarrays.</p> Blocks INFRA-25 Relates INFRA-8 Development Epic Link INFRA-13 Rank 0|0nmhuq: Story Points 8.0