From: Steven H. R. <st...@sh...> - 2009-03-20 03:28:05
|
I'm interested in testing Psyco V2 generator support. What exactly should I check out to do this? Regards, Steve |
From: Christian T. <ti...@st...> - 2009-03-30 22:49:25
|
On 3/20/09 4:26 AM, Steven H. Rogers wrote: > I'm interested in testing Psyco V2 generator support. What exactly > should I check out to do this? http://codespeak.net:/svn/psyco/v2/dist Sorry about the late answer. Frequency is so deadly low that I almost don't expect anything. Maybe this will change, soon :-) cheers - chris -- Christian Tismer :^) <mailto:ti...@st...> tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ |
From: Steven H. R. <st...@sh...> - 2009-03-30 23:55:13
|
Christian Tismer wrote: > http://codespeak.net:/svn/psyco/v2/dist > > Sorry about the late answer. Frequency is so deadly low that > I almost don't expect anything. > > Maybe this will change, soon :-) > > cheers - chris Thanks Christian: Hopefully, your current efforts will attract some attention. Is generator support enabled? I saw a note about instructions in setup.py for enabling this, but didn't see anything promising when I looked. Regards, Steve |
From: Christian T. <ti...@st...> - 2009-03-31 23:43:17
|
On 3/31/09 1:53 AM, Steven H. Rogers wrote: > Christian Tismer wrote: >> http://codespeak.net:/svn/psyco/v2/dist >> >> Sorry about the late answer. Frequency is so deadly low that >> I almost don't expect anything. >> >> Maybe this will change, soon :-) >> >> cheers - chris > Thanks Christian: > > Hopefully, your current efforts will attract some attention. Is > generator support enabled? I saw a note about instructions in setup.py > for enabling this, but didn't see anything promising when I looked. Rev. 62678 from March 7, 2009, had the following comment: """ This is Psyco V2 beta 2 ======================= Generators are completely re-written. We still use a compact object, but state is loaded/saved at every entry or exit to a generator. This is faster, much more difficult to implement, and has the advantage that it does not overly stress promotion. This is still a beta, to be replaced. This version uses an explicit self variable to hold generator state. It will go away, again, but it was helpful for the transition phase. The calling of the generators has a big overhead, which will be optimized in the next beta. Anyway, this is a version that can be used! """ Besides that, how about *trying* things, yourself? How hard is it to figure out, if generators get accelerated or not? Anyways, generators are enabled, and that can be deduced from the fact that the disabling code is gone. :-) ciao - chris -- Christian Tismer :^) <mailto:ti...@st...> tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ |
From: Steven H. R. <st...@sh...> - 2009-04-01 03:13:11
|
Christian Tismer wrote: > Besides that, how about *trying* things, yourself? > How hard is it to figure out, if generators get accelerated or > not? > > Anyways, generators are enabled, and that can be deduced from > the fact that the disabling code is gone. :-) > > ciao - chris Actually, I did try it and saw no significant change relative to v1.6 for my test case, one of the SimPy examples. I didn't see any disabling code, but I thought I might be missing something. I'll put together a more comprehensive test suite and do some more testing. Anything else I can do to help? Later, Steve |
From: Christian T. <ti...@st...> - 2009-04-01 12:04:40
|
On 4/1/09 4:02 AM, Steven H. Rogers wrote: > Christian Tismer wrote: >> Besides that, how about *trying* things, yourself? >> How hard is it to figure out, if generators get accelerated or >> not? >> >> Anyways, generators are enabled, and that can be deduced from >> the fact that the disabling code is gone. :-) >> >> ciao - chris > Actually, I did try it and saw no significant change relative to v1.6 > for my test case, one of the SimPy examples. I didn't see any disabling > code, but I thought I might be missing something. I'll put together a > more comprehensive test suite and do some more testing. Anything else I > can do to help? You can run psycobench.py and see generator acceleration. If it does not work in another example, it can be something else not supported, like free/cell variables. cheers - chris -- Christian Tismer :^) <mailto:ti...@st...> tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ |