April 2 Present: Peter, Steve, Roland, Lucas, Deborah, Beyhan, Maxwell Rizzo, Cheng-Hsin, Zach, Nikolai Wallin Chair: Deborah Minutes: Steve
Release seems to be on track The TOV Gallery is updated Previous one is too big because too much was uploaded Need to link BH-NS collision to the gallery page TwopuncturesX is having trouble Cheng-Hsin says the Teukolsky-Lean test--fails on 2 procs, but not on 2 threads Is output written in consistent order? Cottonmouth - Split into 5 parts to obtain optimal, Zach asks whether these can run in parallel? Zach says that makes the point that higher order often needs different splittings Peter asked about the CPU performance CanudaX review, Lucas has started, needs doc folder (as does Cottonmouth). BHaHAHA review, Peter says it has a lot of good documentation New Gallery--all done, ready to go Boost--Roland has a script to reduce the tarball to 30M or so. Boost has its own build system, bjam. Hooked into make's parallel build system. Considered splitting it into multiple tarballs. Newest boost doesn't have boost::math::tools::eps_tolerance, so can't use it.
Workshop and School SOC meeting this afternoon
Atlassian retiring bitbucket issues thingy Zach will move that in May
Kyoto initiative comments? Who should we ask for updates? Zach, Philip Moesta, Beyhan, and Jay are involved. Zach is happy to be the Kyoto initiative liason.
Unanswered emails: Range error email, new issue created, using a different compiler made it go away
Tickets: ADIOS2 - Cheng-Hsin - looks at the build logs, int is size 4. On Delta. Has had the same issue on the local Ubuntu laptop. 2909 - Reminder to Roland.
"A question that sometimes drives me hazy: am I or are the others crazy?" --Einstein