[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Poor Sun timings, as competition...

    Date: Mon, 30 Jul 1990 9:32:02 PDT
    From: price%iris.usc.edu@usc.edu (Keith Price)

    The orignial discussion of the virtual memory bugs was regarding
    the SparcStations (I, I+), 

The article I saw posted to Sun-Spots described a problem in the entire
Sun-4 line, not just the Sparcstation I.  The first sentence of the
third paragraph was "This problem exists on all Sun-4 systems."

			       which, at least locally, are configured to never
    contain more than 16MBy actual memory to eliminate such problems.  

Having only 16MB of physical memory avoids the bug, but it will page
more if you run jobs with large working sets.  The bug causes memory
beyond about 16MB to be unused by the virtual memory system (but it will
still be usable for I/O buffers).

								       Of course,
    Sun tries to sell you all kinds of memory when you buy a SparcStation.  At least
    these current problems are recognized as bugs, unlike some memory size 
    problems on the Sun3s.

If you're running large applications, paging excessively may result in
worse performance than the bug produces.

By the way, we've sped up our Sun-4's considerably by not running the
"swapper" daemon.