• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: Parallel builds on Intel Mac failing..... start loosing trust in MacOS X....
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Parallel builds on Intel Mac failing..... start loosing trust in MacOS X....

  • Subject: Re: Parallel builds on Intel Mac failing..... start loosing trust in MacOS X....
  • From: Fons Rademakers <email@hidden>
  • Date: Wed, 22 Nov 2006 01:30:48 +0100
  • Keywords: CERN SpamKiller Note: -51 Charset: west-latin
  • Organization: CERN
Hi,

I see this problem on both my MacPos's when all 4 CPU's are at ~100% load compiling. Using memtest I've basically excluded bad RAM issues. I've now disabled mdimport on my project's source and build directories (although the system should never fail with dyld problems under whatever load). Lets see if that helps.

Cheers, Fons.


David Fang wrote: 
  does anybody on this list ever do a parallel build using e.g. "make -j 4"
on a quad-core MacPro. For me this is a total hit and run proposition. It
basically always results in random MacOS X crashes related to the dynamic
loader. See for example:


Hi,
	As a data point, I do make -j{2,4} builds on powerpc G4, G5, and
Intel Core Duo iMac every day with no problems.  I do get occasional
fork/vfork errors failing to launch too many processes, but that's not
XCode's/make's problem.  [Yes, there are people on this list who've never
opened XCode.  :P Call me archaic.] One final tip regarding mdimport: I
always disable spotlight on my build directories (Preferences:Spotlight),
which are always separate from by source directories (VPATH building).
This way spotlight can happily index the source tree and not bother with
built objects.

Fang

Host Name:      proof
Date/Time:      2006-11-21 10:46:42.804 +0100
OS Version:     10.4.8 (Build 8L2127)
Report Version: 4

Command: as
Path:    /usr/libexec/gcc/darwin/i386/as
Parent:  as [21696]

Version: ??? (???)

PID:    21697
Thread: Unknown

Link (dyld) error:

Symbol not found: _vm_allocate
   Referenced from: /usr/libexec/gcc/darwin/i386/as
   Expected in: /usr/lib/libSystem.B.dylib

---------------------------------------------------------------------

Host Name:      proof
Date/Time:      2006-11-21 10:58:44.082 +0100
OS Version:     10.4.8 (Build 8L2127)
Report Version: 4

Command: rootcint_tmp
Path:    utils/src/rootcint_tmp
Parent:  make [21741]

Version: ??? (???)

PID:    27236
Thread: Unknown

Link (dyld) error:

Symbol not found: __ZTTSt14basic_ofstreamIwSt11char_traitsIwEE
   Referenced from: /usr/lib/libstdc++.6.dylib
   Expected in: /usr/lib/libstdc++.6.dylib


In the console.log I see in this case:

Nov 21 10:46:43 proof crashdump[21707]: as crashed
Nov 21 10:46:43 proof crashdump[21707]: crash report written to:
/Users/rdm/Library/Logs/CrashReporter/as.crash.log
Nov 21 10:56:17 proof crashdump[22254]: mdimportserver crashed
Nov 21 10:56:18 proof crashdump[22254]: crash report written to:
/Users/rdm/Library/Logs/CrashReporter/mdimportserver.crash.log
Nov 21 10:56:44 proof crashdump[23327]: mdimportserver crashed
Nov 21 10:56:44 proof crashdump[23327]: crash report written to:
/Users/rdm/Library/Logs/CrashReporter/mdimportserver.crash.log
Nov 21 10:57:58 proof crashdump[25756]: mdimportserver crashed
Nov 21 10:57:58 proof crashdump[25756]: crash report written to:
/Users/rdm/Library/Logs/CrashReporter/mdimportserver.crash.log
Nov 21 10:58:10 proof crashdump[26247]: mdimportserver crashed
Nov 21 10:58:10 proof crashdump[26247]: crash report written to:
/Users/rdm/Library/Logs/CrashReporter/mdimportserver.crash.log
Nov 21 10:58:44 proof crashdump[27238]: rootcint_tmp crashed
Nov 21 10:58:44 proof crashdump[27238]: crash report written to:
/Users/rdm/Library/Logs/CrashReporter/rootcint_tmp.crash.log

Also, can the mdimportserver not handle the many files created during a
large compilation?

These problems are on a MacPro 3GHz with 4GB RAM, but I see the same
problems on MacBook Pro's.

After a fresh reboot the compile works fine (1.5 millions lines of C++,
takes when everything goes well about 8 min), but after a while (a few
hours) I start having these errors


Cheers, Fons.



David Fang
Computer Systems Laboratory
Electrical & Computer Engineering
Cornell University
http://www.csl.cornell.edu/~fang/
	-- (2400 baud? Netscape 3.0?? lynx??? No problem!)


--
Org:    CERN, European Laboratory for Particle Physics.
Mail:   1211 Geneve 23, Switzerland
E-Mail: email@hidden              Phone: +41 22 7679248
WWW:    http://fons.rademakers.org           Fax:   +41 22 7669640
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden






References: 


 >Re: Parallel builds on Intel Mac failing..... start loosing trust in MacOS X.... (From: David Fang <email@hidden>)






    

  • Prev by Date:
Re: Can anyone explain what the function argument declaration: ( int  _Tp::* )  means ?

    • 

  • Next by Date:
Re: why don't custom views have a "tag" in IB?

    • 

  • Previous by thread:
Re: Parallel builds on Intel Mac failing..... start loosing trust in MacOS X....

    • 

  • Next by thread:
Re: XCode misbehaving on Intel

    • 

  • Index(es):

      

    • Date
      • 

    • Thread
      • 

    

    •