2.4m is faster than 2.4l. Have you tried disabling variable reuse by - no-variable-reuse That's what Kyle recommends, till the bug is fixed.
Sunish On Nov 14, 1:54 pm, Sebastien Lelong <[email protected]> wrote: > Hi Rob, > > I can't explain why it runs slower on my server... (see last failure, again > due to a buildbot timeout). Maybe 2.4m is requiring more memory, and since > my PC swaps a lot, impacts are bigger ? > > I'll try to setup a higher timeout. > > Also, Sunish was (is ?) having problems with slower compilation with > JalEdit. Sunish, did you try last 2.4m beta compiler ? > > Cheers, > Seb > > 2009/11/14 Rob Hamerling <[email protected]> > > > > > > > Sebastien Lelong wrote: > > > > Just an idea. My script to compile all samples uses the SVN directory > > > structure (not the release structure with a single /lib include > > > directory). For the '-s' compiler parameter a string of > > > include-directories is built, but only those directories are included > > > which actually contain any .jal files. This string is with me about > > > 800 bytes and contains 25 directories. When I remember correctly you > > > do a similar thing, but are you also selective (skipping empty > > > directories)? > > > > No, I'm not selective, but it behaves the same whatever the compiler. > > > This might explain differences between your measures and mine, but > > > not within a same system. > > > True, I was thinking about the factor 10 difference. > > > More measurements: > > > Same run as previously, now on an Intel P4 2800 MHz, 512 MB RAM. > > First under eCS compiler 2.4l: 187 seconds, almost the same as on my AMD > > X2 3800! This was with the files on a JFS-formatted volume. The same but > > the files on a FAT32 volume (same harddisk): 291 seconds. The eCS FAT32 > > driver is known to be rather slow. > > > Then I did a run with under Windows. Regina Rex script interpretor, > > Compiler 2.4l (files on the FAT32 volume): 205 seconds, and the 2.4mbeta > > compiler: 196 seconds. > > > The measurements seem to be consistent: 2.4m is faster than 2.4l here, > > although under eCS a larger difference than under Windows. > > > It would take me some more time to repeat the runs under Linux and I > > doubt if it is very useful... > > > Just out of curiosity I tried parallel compilation (starting a > > new thread for every compilation) on my dual-core AMD. Now all samples > > compile in about 100 seconds, with both CPUs at 100%! > > > Regards, Rob. > > > -- > > Rob Hamerling, Vianen, NL (http://www.robh.nl/) > > -- > Sébastien Lelonghttp://www.sirloon.nethttp://sirbot.org --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "jallib" 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/jallib?hl=en -~----------~----~----~----~------~----~------~--~---
