--On Tuesday, May 25, 1999, 11:25 AM -0700 "Laurence Rozier" lrozier@thepattern.com wrote:
When I start PWS, performance throughout the rest of my Squeak environment takes a nose dive. Simple things like selecting a window have a noticeable delay. I tried tweaking the secondsBetweenBackups := 1
- (60 * 60) and the Delay forMilliseconds: 100 behaviors in the
loopOnPort:loggingTo: method but to no avail. Any suggestions?
My guess is that your Squeak version is > 2.1. Something happened somewhere around 2.2/2.3 that causes this problem. I've seen it on Macs and Windows. 2.1, OTOH runs fine.
Worse, not everyone seems to see it. I've spent some hours digging around and I've tenatively concluded that it's a VM problem. Next step, I suppose, is to compare the c code for the network primatives between 2.1 and 2.4.
I wouldn't mind this as much as it's not *just* the UI or other processes slowing down--the server slows waaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaay down. (No, the vowels were *not* superfluous! :))
I have yet to test Comache (the next generation PWS) though I intend to this week. If it exhibits the problem, well, it'll be another data point!
My current advice is that if you see this problem, drop back to 2.1. If anyone *is* running 2.2 or greater wthout this problem, please let me know!
Cheers, Bijan.
bparsia@email.unc.edu wrote:
--On Tuesday, May 25, 1999, 11:25 AM -0700 "Laurence Rozier" lrozier@thepattern.com wrote:
When I start PWS, performance throughout the rest of my Squeak environment takes a nose dive. Simple things like selecting a window have a noticeable delay. I tried tweaking the secondsBetweenBackups := 1
- (60 * 60) and the Delay forMilliseconds: 100 behaviors in the
loopOnPort:loggingTo: method but to no avail. Any suggestions?
My guess is that your Squeak version is > 2.1. Something happened somewhere around 2.2/2.3 that causes this problem.
I am running under 2.4.
I've seen it on Macs and Windows. 2.1, OTOH runs fine.
Worse, not everyone seems to see it. I've spent some hours digging around and I've tenatively concluded that it's a VM problem. Next step, I suppose, is to compare the c code for the network primatives between 2.1 and 2.4.
I wouldn't mind this as much as it's not *just* the UI or other processes slowing down--the server slows waaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaay down. (No, the vowels were *not* superfluous! :))
I just tried PWS under 2.1 and can really see the difference!
I have yet to test Comache (the next generation PWS) though I intend to this week.
Is Comache available for download? I'd be more than willing to help with testing. Thanks,
Laurence
If it exhibits the problem, well, it'll be another data point!
My current advice is that if you see this problem, drop back to 2.1. If anyone *is* running 2.2 or greater wthout this problem, please let me know!
Cheers, Bijan.
bparsia@email.unc.edu writes:
My current advice is that if you see this problem, drop back to 2.1. If anyone *is* running 2.2 or greater wthout this problem, please let me know!
Hi, I run 2.4c and have run 2.3 on both a 68k powerbook as well as a PPC powerbook, both of them older and slower, and notice no real slowdown with PWS. I do pretty heavy development work on a PWS application on 100mhz 603e (think 120mhz Pentium) and the UI is just as fast whether or not I run PWS. I also run a Swiki on my 040 powerbook (think slow 486 or so) with Scamper and it runs fine. It doesn't slow down any when I start up PWS, and trust me here, I'd notice almost any slowdown, it's slow enough as it is :-)
Looking at PWS Class loopOnPort:loggingTo: I wonder if the problem with other systems isn't in the Delay forMilliseconds:100 send at the bottom? If you didn't delay, or, if the delay was calculated wrong it would spin very fast. Maybe...
cheers
bruce
squeak-dev@lists.squeakfoundation.org
-
bparsia＠email.unc.edu -
Laurence Rozier -
Mark A. Schwenk -
O'NEEL Bruce