lS6eozFPqZQxfTz6kYVeiccMyOERtvrOX6wOLsugB3zVLGieA==   
   HzzNK1nH5VogDW2HPZ5KCknb0G4ZKg2j5wqghk+Im7691FQh4btah72dNkdYnyNj   
   ZxNciEGZHwxEyCh3N2CdpxodpjOkkNTkw6kuAbwwUe4raVi5"; mail-complain   
   s-to="abuse@albasani.net"   
   From: "Ruediger Ihle"    
      
   On Thu, 14 Jul 2011 06:45:52 UTC, Dave Yeo  wrote:   
      
   > Perhaps we should be recommending everyone to set NSPR_OS2_NO_HIRES_TIMER ?   
      
   AFAIK, eCS ships with that option set to 1 by default for quiet some time.   
      
      
   > Is the alternative of using DosQuerySysInfo(QSV_MS_COUNT, QSV_MS_COUNT,    
   > &msCount, sizeof(msCount)); that much more inaccurate?    
      
   The granularity is about 32ms. Maybe down to 8ms, when CLOCKSCALE=4    
   is given in CONFIG.SYS, but not sure about that...   
      
      
   > Windows just uses timeGetTime(), I have no idea of how fine of    
   > resolution it has.   
      
   Hmm. Last time I looked at the code (a long time ago), they were   
   using QueryPerformanceCounter() and QueryPerformanceFrequency().   
      
      
   Microsoft writes:   
     The default precision of the timeGetTime function can be five    
     milliseconds or more, depending on the machine. You can use the   
     timeBeginPeriod and timeEndPeriod functions to increase the    
     precision of timeGetTime. If you do so, the minimum difference   
     between successive values returned by timeGetTime can be as    
     large as the minimum period value set using timeBeginPeriod    
     and timeEndPeriod. Use the QueryPerformanceCounter and    
     QueryPerformanceFrequency functions to measure short time    
     intervals at a high resolution   
      
      
      
   --    
   Ruediger "Rudi" Ihle [S&T Systemtechnik GmbH, Germany]   
   http://www.s-t.de   
      
      
   --- Internet Rex 2.31   
    * Origin: S&T (1:261/20.999)