[racket-dev] mouse cursor changes cause frequent collections?

2011-07-25 Thread John Clements
I have the log showing right now, and I see that moving the mouse in a way that causes mouse cursor changes (for instance, back and forth between the definitions & interactions windows) causes collections of about 32K on the order of every eight to ten mouse cursor changes, something like 2K of

Re: [racket-dev] Pre-Release Checklist for v5.1.2

2011-07-25 Thread Michael Sperber
Stephen Chang writes: > I've made a push that should fix both these problems. The push should > be included in the release. > > The stepper was intercepting rendering of lists so that infinite lists > could be handled, but I've restored deferring to render-to-sexp for > non-lazy lists. Works fo

Re: [racket-dev] can we write these four lines of C in performant racket?

2011-07-25 Thread Sam Tobin-Hochstadt
On Mon, Jul 25, 2011 at 7:51 AM, Matthew Flatt wrote: > > Here are some timings for 1000 iterations on 2^20-element inputs > (32-bit mode, Mac Book Pro 2.53 GHz): > >  C as above, gcc -02          :  1409 >  C with indirections, gcc -O2 :  4041 >  C as above, gcc -O0          :  6425 >  C with ind

Re: [racket-dev] can we write these four lines of C in performant racket?

2011-07-25 Thread Matthew Flatt
At Sat, 23 Jul 2011 14:42:15 -0400, John Clements wrote: > This C code adds the content of one buffer to another one, with no checking. > The corresponding racket code runs about 10x slower. Do you folks think that > it > should be possible to do better? (One salient fact: these are > shorts--