kowalski@svax.cs.cornell.edu (Jeffrey Kowalski) (12/14/89)
Hi, FIRST: I've got a Firefox (vs3520) running Ultrix. It's supposed to be a fast graphics machine, but it seems slow running decwindows. When I run Xgb (the X server for this type of machine), I get an error saying that it can't find microcode /usr/lib/ucode/vs3520. It runs without it, but I have a feeling that it's running "dumb," i.e. it's not using the double buffer, super-slicko hardware. I've hunted everywhere and cant find the ucode. So, what's the scoop? NEXT: the 3520 is a 2 processor machine, but I can't tell. I mean, it sez that "processor b is started," but a ps gives no indication of parallelization, even at the processor level. On a Gould, a ps will tell you which cpu your job's on. So what gives? Does the machine run both cpus or not really. Please help. Running slowly, jeff please respond via email, since I don't read news often.
vixie@decwrl.dec.com (Paul A Vixie) (12/29/89)
Sorry, can't help with your first question, I don't know. >> NEXT: >> the 3520 is a 2 processor machine, but I can't tell. I mean, >> it sez that "processor b is started," but a ps gives no indication of >> parallelization, even at the processor level. On a Gould, a ps will >> tell you which cpu your job's on. So what gives? Does the machine >> run both cpus or not really. On Ultrix versions up through 3.1, the parallel processing isn't visible through normal means. You can see it with iostat and you can certainly feel the difference if you run a GNU Make in parallel mode :-). The multiprocessor support gets a lot better in 4.0, but until then all it's really doing is allocating a processor at schedule-time -- EACH schedule time. Your job doesn't get "allocated" to a processor. Even in 4.0 it won't be "allocated", it's a many-processors-to-many-jobs mapping that takes place continuously. Paul Vixie DEC WRL -- Paul Vixie Work: vixie@decwrl.dec.com decwrl!vixie +1 415 853 6600 Play: paul@vixie.sf.ca.us vixie!paul +1 415 864 7013