>> On Thursday, June 29, 2017 at 9:15:47 AM UTC+2, Kiran Kedlaya wrote: >> Over the years, various attempts have been made to implement good linear >> algebra over ZZ and QQ in Sage. (The initial motivation, which still remains >> quite relevant, was William's desire to compute with spaces of modular >> forms.) >> >> Can someone give me a rundown of what works well in Sage right now and what >> could use improvement (say, compared to what Magma provides)? >> >> Similar question over GF(p) for p small. p=2 seems to work well using m4ri, >> but there is a much steeper dropoff for p>2 vs p=2 than I'm seeing in Magma.
This is an incredibly hard question to answer, especially not knowing why you are asking it. E.g,: - because you’re writing a grant proposal? - because a student wants to a dev project on linear algebra that you’re advising? - because you want to say something in a book? in an article? - because you are designing/implementing an algorithm in Sage, and need to know what’s available to make better choices? Also, one critical thing to keep in mind is that surprisingly many Sage linear algebra algorithms ultimately depend on what numerical linear algebra library is linked in and **how it is built**. For example, ATLAS + special parallel build? OpenBlas? Built from source versus a binary, Etc.; this can very, very significantly influence performance for some benchmarks (easily a factor of 10x). Also, some Sage functions have a range of algorithm options, with now bad default choices (things may be more intelligent in Magma, since their source is closed so it’s much harder for users to even try to make choices manually... they have to automate everything... and they are also really good at it.). Clement Pernet may know a lot too about the general state of open source linear algebra. William >> >> Thanks, >> Kiran >> > > -- > You received this message because you are subscribed to the Google Groups > "sage-nt" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/sage-nt. > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "sage-nt" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send an email to [email protected]. Visit this group at https://groups.google.com/group/sage-nt. For more options, visit https://groups.google.com/d/optout.
