Additional timings.  After a first clone (from the PBUILD build) doing
"sage -br" to get back to the main branch took 8m29s.  Then doing a
new clone from the main branch took 7s.  Then another "sage -br" is
almost instantaneous, after which yet another clone is only 7s.

I'm not sure any of this helps, it certainly seems bizarre to me.

John

2008/5/8 John Cremona <[EMAIL PROTECTED]>:
> The following were all on the 64-bit Suse machine I normally build and
>  run Sage on.  It has 4 processors and I used  SAGE_BUILD_THREADS=2 or
>  3 (I forget).
>
>
>  Times to build 3.0.1:
>
>  Without PBUILD:  94m (user time)
>  With     PBUILD:  94m (user time)
>
>  Times to make clones:
>
>  Without PBUILD
>  First clone: 7.48s
>  Second clone (from main branch) 8.0s
>  Third clone: 7.4s
>
>  -- all as expected.
>
>  With PBUILD
>  First clone:     18m 7.7s
>  --looked like it was doing a build-all; I saw no evidence of more than
>  one process running at a time; it spent most of this time running gcc,
>  (re)"building the Sage library" I think
>  --There is definitely something wrong here if cloning is only supposed
>  to copy files!
>
>  Second clone (from clone branch): 7.2s  -- that's more like it!
>  Third cone (from main branch again): long again
>
>  I hope this info will help others find out what's up!
>
>  John
>
>  2008/5/8 mabshoff <[EMAIL PROTECTED]>:
>
>
> >
>  >  On May 8, 10:13 am, Dan Drake <[EMAIL PROTECTED]> wrote:
>  >
>  >  Hi Dan,
>  >
>  >
>  >  > As a final follow up to this thread, I tried 3.0.1 on another machine:
>  >  > this one has the same Linux distribution on it, but it's a Core 2 Duo
>  >  > (so no CPU mismatch!). I got the same behavior: pbuild took much longer
>  >  > (105 mins versus 75 mins) and 'make test' died on testdoc.py. A normal
>  >  > build and test went through just fine.
>  >
>  >  Can you actually compare the build times of pbuild and the old build?
>  >  It seems like we will see a 30 minute difference. Is that also on FC7
>  >  or is it a different distribution?
>  >
>  >
>  >  > I'll check out the fix that's coming up and see if those problems go
>  >  > away.
>  >
>  >  The fix will only get rid of the doctesting hang. We do not know so
>  >  far what causes the slowdown.
>  >
>  >  > Dan
>  >
>  >  Cheers,
>  >
>  >  Michael
>  >
>  >
>  > >  >
>  >
>

--~--~---------~--~----~------------~-------~--~----~
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/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to