sdg@uts.amdahl.com (Subrata Dasgupta) (07/26/90)
I have heard that running X on a Sun Sparcstation uses up close to half of the available mips. Is this true? Also, how much runtime memory is typically used by X ? How would these numbers vary is someone were to run the Motif window manager instead of xwm ? Thanks -sdg
mouse@LARRY.MCRCIM.MCGILL.EDU (07/27/90)
> I have heard that running X on a Sun Sparcstation uses up close to > half of the available mips. Is this true? Yes and no...Running X on a machine can use up anything from nearly zero to nearly all of the available cycles. It all depends on what you're doing, and is true regardless of the machine (unless the machine is incredibly slow, in which case the minimum percentage is higher). > Also, how much runtime memory is typically used by X ? I doubt there is a "typical" figure that really means anything. I'm on a Sun-3/50 with 4 Mb of core at the moment and it's swapping itself to death - but it's not clear how much of this can be ascribed to X itself, how much to my work habits, and how much to whatever programs happen to be running at the moment. > How would these numbers vary is someone were to run the Motif window > manager instead of xwm ? I don't know, never having used either of them - but I would guess that this is a minor variable which gets swallowed in the major ones of work habits, other programs, etc. der Mouse old: mcgill-vision!mouse new: mouse@larry.mcrcim.mcgill.edu (All: I'd've mailed this, but the poster didn't provide an address....)
marbru@auto-trol.UUCP (Martin Brunecky) (07/27/90)
In article <9007262124.AA18843@Larry.McRCIM.McGill.EDU> mouse@LARRY.MCRCIM.MCGILL.EDU writes: >> I have heard that running X on a Sun Sparcstation uses up close to >> half of the available mips. Is this true? > >Yes and no...Running X on a machine can use up anything from nearly >zero to nearly all of the available cycles. It all depends on what >you're doing, and is true regardless of the machine (unless the machine >is incredibly slow, in which case the minimum percentage is higher). > Half of available mips ??? Why not ALL ??? For whatever reason, people keep forgetting that to get 16x more performance than from VAX 11/780 they need not only 16x faster CPU chip, but also 16x more memory, 16x faster disk I/O channel .... The problem with SPARcstation is that at 8MB it can barely deal with resident portion of UNIX and Sun's X server (though the later is getting better). Now you want to add couple X applications, so that you start to exercise Unix paging/swapping algoritms - and the result is inevitable: you perform at or below that bashed, obsolete VAX 11/780, which at least (when it used Massbus) had decent I/O channels. Plus, it's paging algoritms (though far from being perfect) have been designed for multitasking, not a for a single thread. So, back to history. The key word for VAX 11/780 performance in the first half of 80's was: B U Y M O R E M E M O R Y !!!!!! For whatever reason, history still goes in circles. -- =*= Opinions presented here are solely of my own and not those of Auto-trol =*= Martin Brunecky marbru@auto-trol.COM (303) 252-2499 {...}ncar!ico!auto-trol!marbru Auto-trol Technology Corp. 12500 North Washington St., Denver, CO 80241-2404