d6a57d1dba32d6236b7e.newsdawg.com   
   From: "Dariusz Piatkowski"    
      
   On Fri, 8 Jul 2011 03:33:22 UTC, "Rich Walsh"  wrote:   
      
   > On Fri, 8 Jul 2011 01:57:37 UTC, "Dariusz Piatkowski"  wrote:   
   >    
   > > Well guys, I tried every single option I could think of, such as the   
   following:   
   > >    
   > > 1) no RWS   
   > > 2) no hrdwr speedup   
   > > 3) "PSD=OS2APIC.PSD /P=4 /APIC" (normal mode) replaced with    
   PSD=OS2APIC.PSD    
   > > /P=4 /PIC"   
   > > 4) Firefox started with 'safemode' (all add-ons and personalization   
   shut-off)   
   > > 5) this version (4.0.2b) appears to be better behaved, but once 3-4   
   windows are    
   > > started up (about 2-3 tabs each) the 4 CPU cores start getting maxed   
   out....   
   > >    
   > > I really am empty handed at this point in time, although I can tell that   
   the    
   > > Firefox configuration did not change (other then 4.0-pre to 4.0.1 & 4.0.2)   
   and    
   > > that the issue only appeared once multi-core system was activated.   
   >    
   > It seems fairly clear that the problem lies in your system's setup, so I'd   
   > strip it to the bare bones.   
   >    
   > First, back out the PMMERGE from FP6 - IIRC, it had problems.   
   >    
   > Next, disable all of your add-ons:  empty your startup folder, rem-out RUN=   
   > statements in config.sys, and deregister questionable WPS classes.   
   >    
   > FWIW... a few years ago, all my compiles started failing.  After I disabled   
   > everything then started adding stuff back, I discovered the cause was FT2LIB.   
   > It certainly wasn't what I expected because compilers, linkers, etc, are all   
   > text-mode apps while FT2LIB only loads into PM processes.  Still, it _was_   
   > the cause of the problem.  The moral here:  leave no stone unturned.   
      
   I agree with you...certainly somewhere in there I do have a problem...the big    
   question is how the heck do I get at it in a somewhat of an planned manner, as    
   opposed to just trying things on the fly so to speak.   
      
   Yes, I did hear enough reports that PMMERGE.DLL was a problem area in    
   FP6...strangely enough, doing a binary compare to the FP5 version only shows    
   differences where th BLDLEVEL info comes in:   
      
   1) FP5, 12-29-04   7:33p   1270499          49  PMMERGE.DLL   
      
   IBM:14.103   
      
   2) FP6,  7-05-06  10:28p   1270499          49  PMMERGE.DLL   
   IBM:14.105   
      
   Sure enough...size wise, they are the same.   
      
   I am currently running the matching FP6 SMP krnl & DOSCALL1.DLL files.   
      
   In my attempts to narrow this down I had started executing something similar   
   to    
   your idea regarding the CONFIG.SYS statments, classes, etc....so far I'm   
   really    
   going through Firefox related stuff though.   
      
   Regarding the particular Firefox/Mozialla hi-res timer value (SET    
   NSPR_OS2_NO_HIRES_TIMER=1), can you point out what specifically this causes    
   Firefox to do (or avoid doing)??? I'm curious if I can use that explanation to    
   further narrow down better what might be the root cause of the problem here?   
      
   Thanks!   
      
   --- Internet Rex 2.31   
    * Origin: NewsGuy - Unlimited Usenet $19.95 (1:261/20.999)