On Thursday, September 23, 2021 at 9:54:19 PM UTC-7 wst...@gmail.com wrote:
> Since Fortran is pretty central to this discussion, any chance you > could give a quick overview of the extent to which Sage depends on > Fortran in 2021? No big surprises: It comes in from openblas, numpy, scipy; and suitesparse (a dependency of cvxopt). openblas, of course, is not just used via these packages, but by many more of our C libraries. numpy&scipy of course make high quality wheels available for all platforms, but we probably shouldn't try to build our C libraries against their vendored openblas... I was reading about pyodide [1] recently and they had to work very > hard to build scipy for webassembly without using Fortran. Yes, this is very impressive work. Likely will be important for edge-computing type deployments. If we ever want to have Sage run in such an environment, modularization is the way to go - so that we can focus on subsystems with minimal compiled library dependencies. I imagine things might be better today, due to Flang, which seems to > be a fortran frontend (like Clang) for LLVM. That seems likely to > work well on MacOS. > We have a meta-ticket that tracks this: https://trac.sagemath.org/ticket/23926 But I have not followed the details. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/f1860056-f451-456f-a073-e5df8f175512n%40googlegroups.com.
