Mark_Weiser.PARC@xerox.com (11/01/90)
We have the patch. It made a very significant difference in our pmeg thrashing on our 16MB and 28MB sparcstations. Some of our real everyday applications went form 80% system to 20% system. That is like going from 3MIPS to 12MIPS. Sun, great work on this patch. (Why Sun is first making users run a program to see if you "really need it", when they never did this on any other of their bug fixes or patches, is beyond me. Maybe they don't consider shipping an O.S. that doesn't let you use much more than 16MB of a 64MB-capable machine a bug, just a misfeature? Or maybe its considered a hardware problem, and this is just a workaround (my opinion too, actually).) -mark
trevor@media-lab.media.mit.edu (Trevor Darrell) (11/21/90)
In article <1990Oct31.235802.22736@rice.edu> Mark_Weiser.PARC@xerox.com writes: >We have the patch. It made a very significant difference in our pmeg >thrashing on our 16MB and 28MB sparcstations. Some of our real everyday >applications went form 80% system to 20% system. That is like going from >3MIPS to 12MIPS. Sun, great work on this patch. Is there a PMEG patch needed for non-sparcstation sun-4's? I have 4/260 running SunOS 4.0, and run lisp jobs that exceed 64M of virtual.