Le mardi 18 mai 2010 à 12:03 -0400, Christian Muise a écrit : > I'm starting to get mighty confused with the parallel attempts. > Fabian, Ronan, and Ondrej all have at least one branch doing the same > thing that is very closely related to my SoC goals -- removing the old > system and making the new one uber fast. Any suggestions on how to > unify all the threads? > I'm only attempting to remove the old system with as little changes to the rest of the code as possible, not to make the new one fast - actually, I seem to be doing the exact opposite. Fabian's attempt is old and got nowhere, so you can probably ignore it. I haven't looked too closely at Ondrej's recent work, but there seem to be a lot of changes unrelated to assumptions and I'm afraid it'll break many things.
> To your specific points, Ronan, I think the idea was to remove the > old assumptions from the backend and just have a stripped down simple > version that could easily be replaced by the new system when it's > ready (ie. fast). And to echo Ondrej's question, I'd love to know what > your metric was to tell the speed difference. I don't understand how it's possible to remove the old assumptions from the backend since most operations and algorithms rely on it. My approach is to make the new system sufficiently powerful to be able to replace the old one. Your role in this would be to make it fast, which is the other necessary precondition for the replacement. The metric is just the time taken to run the tests. Since it goes up by an order of magnitude or more for most tests, I don't need precise measurements to say that there's a severe slowdown. > Cheers > > On Tue, May 18, 2010 at 10:02 AM, Vinzent Steinberg > <[email protected]> wrote: > On 18 Mai, 07:31, Ronan Lamy <[email protected]> wrote: > > After my work on Predicates, I've made another attempt at > replacing the > > old assumption system with the new one and the results are > encouraging. > > After a few more relatively simple changes to the > assumptions system - > > that make it significantly more powerful - it appears that > it's possible > > to completely the old assumptions with the new ones with > only reasonably > > many failures, and very few seeming intractable. The biggest > problem is > > that the switch causes a very painful slowdown. It also > seems that a lot > > of tedious work is required to port all the capabilities of > the old > > system and thus fix the test failures. > > > > The (unpolished) code is in the branch "use-new-assump2" in > my github > > repo > [http://github.com/rlamy/sympy/commits/use-new-assump2]. I've > > gathered prerequisites to the switch passing all tests into > > > "new-assump" [http://github.com/rlamy/sympy/commits/new-assump]. > > > The old system did only trivial assumptions, I think these are > slower > with the new system. Probably we should use caching? > > See also issue 1884 [1]. > > Vinzent > > [1] http://code.google.com/p/sympy/issues/detail?id=1884 > > > -- > You received this message because you are subscribed to the > Google Groups "sympy" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to sympy > [email protected]. > For more options, visit this group at > http://groups.google.com/group/sympy?hl=en. > > > > > > -- > You received this message because you are subscribed to the Google > Groups "sympy" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to sympy > [email protected]. > For more options, visit this group at > http://groups.google.com/group/sympy?hl=en. -- You received this message because you are subscribed to the Google Groups "sympy" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/sympy?hl=en.
