Hi Juhani, On Thu 02 Jul 2009 17:49, Juhani Viheräkoski <moonsh...@kapsi.fi> writes:
>> > With recent changes in guile vm there are lots on improvements on the >> > Gambit benchmarks. >> >> Improvements compared to what? > > I should have been more concise.. You were concise, but not precise ;-) So here's a project, should you choose to be interested ;-) Sometimes we do things that we don't know exactly how they affect performance. We're working on getting decent measurement into Guile's build itself... I mean, we want to work on it :) But we're not measuring. I would really, really love a graph of performance over time on the various Gambit benchmarks. So for every git revision, a number (and standard deviation?) for every Gambit benchmark. Ideally displayed in tables and graphs. I just mention all of this to you because you said at one point that you liked systematically testing things ;-) >> This is not tail-recursive (because of the nested `ack' call in the >> `else' clause), which is why it can lead to a stack overflow. > > I understand that, but I reported this because interpreter (both 1.8.6 > and 1.9.0) run this test successfully but VM does not. Probably > real-world programs don't have so deep a recursion, although they might > when run with complex data.. This is a bug in the VM. The default stack is low, and we have no capacity to enlarge it on demand. We should keep it low and enlarge it on demand via overflow handlers. Peace, Andy -- http://wingolog.org/
