Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!usc!elroy.jpl.nasa.gov!turnkey!orchard.la.locus.com!fafnir.la.locus.com!fafnir.la.locus.com!dlt From: dlt@locus.com (Dan Taylor) Newsgroups: comp.sys.amiga Subject: Re: UNIX sys V4.0 Message-ID: Date: 31 Oct 90 21:19:11 GMT References: <9010025@hpfcso.HP.COM> <15404@cbmvax.commodore.com> <16001@hydra.gatech.EDU> <15464@cbmvax.commodore.com> <22912@grebyn.com> <16102@hydra.gatech.EDU> Organization: Locus Computing Corporation, Los Angeles, California Lines: 53 ken@dali.gatech.edu (Ken Seefried iii) writes: >In article <22912@grebyn.com> ckp@grebyn.UUCP (Checkpoint Technologies) writes: >> >>But the 030 plus the blitter is able to do bitblits plus run >>applications faster than the 030 alone. This is the reason to >>have a graphics co-processor, so that your applications get more CPU >>because the CPU is spending fewer cycles doing graphics. >This is, of course, absolutely wrong (though it is a common >fallicy). The actual advantage gained from a graphics coprocessor >depends on many things. For example, setup time for the graphics >chip can have a big influence on how useful it is. Another big >area (especially in low-end machines) is memory architecture. >Keith Packard (keith@expo.lcs.mit.edu) is a useful source of hard >data on the relative usefulness of graphics co-processors, >especially under X. In general, a graphics chip *can* be an >advantage, but it isn't always. I don't know anything about the >Amiga blitter implimentation, so I can't comment on what may or may >not be it's problems. If you can't comment, why did you? >An additional disadvantage of the graphics chips is that, being >somewhat specialty items, they do not track advances in technology >as fast a general purpose CPUs. >>Now if you tell me that the 030 alone *can* run applications and do >>graphics faster than the 030 plus the blitter, well then I won't believe >>you... I am inclined to agree here. What I want to see is the tests performed and the results. The 68030 is bandwidth limited, as are the co-processors. Where was the performance gain? What was the blitter used for? Area fills, i.e. tiling? Font copies? Line drawing? I would believe that the '030 is faster than the blitter for many operations, I've liked the '020 & '030 for a long time because of their graphics capabilities. But I do find it difficult to believe that the blitter was useless for X, actually slowing operation of the total system. >Never good to make statements like that when totally lacking hard >evedence... Please take your own advice. >-- > ken seefried iii "A snear, a snarl, a whip that > ken@dali.gatech.edu stings...these are a few of > my favorite things..." -- * Dan Taylor * The opinions expressed are my own, and in no way * * dlt@locus.com * reflect those of Locus Computing Corporation. *