On Sun, Jan 4, 2015 at 7:26 AM, Travis Scrimshaw <[email protected]> wrote: > Hey William, > >> >> Does your change impact performance? > > > I don't completely know. I believe it will not as I think most of the > time people would create an instance of ManinSymbol from ManinSymbolList > (plus it uses caching that was there prior to these changes), so there won't > be a performance hit from that. There's a good probability that > ManinSymbolList should be a UniqueRepresentation, but the manin_symbols() > method is currently cached, so we won't see a performance hit that way. So > if you're creating instances of (subclasses of) ManinSymbolList directly > frequently, then you might see a hit, but I suspect that's bad usage. Thus I > don't see any uses where it could result it a slowdown, but perhaps > ManinSymbolList is being used in ways I don't know of. I will run some > timings with some possible ways I see this being used.
The ManinSymbolList code is I think performance critical. It's at the core of some modular symbols/modular forms algorithms. Changing performance there by a constant factor could potentially slow down everything. Because of caching you have to be extremely careful with benchmarking, since you can't run a test more than once during a session (since clearing caches can't be trusted). For example, you write "but the manin_symbols() method is currently cached, so we won't see a performance hit that way."... which might be true, but also might just mean that the way you benchmark doesn't see anything. All that said -- looking over your patch -- I don't think there will be any performance hit; maybe your code will even be faster. > > > Hey John. > >> Can you copy this post to sage-nt? >> >> John > > > I applied to sage-nt, so once I join, I will cross-post. > > Best, > Travis > > -- > 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 [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. -- William Stein Professor of Mathematics University of Washington http://wstein.org -- 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 http://groups.google.com/group/sage-nt. For more options, visit https://groups.google.com/d/optout.
