blom@cc.helsinki.fi (04/05/91)
Does graphics co-processor speed up Windows applications? I am trying to find out ways to improve the screen drawing speed of CorelDraw. Tests with 386 and 486 systems have shown that a more efficient processor or increased amount of RAM have only a marginal effect. Now I am considering video boards with a graphics co-processor like the NEC Graphics Engine. Tests (e.g. PC Magazine 4/91) indicate that they would have a significant effect on drawing speed with MS-Windows. This doesn't necessarily apply to application programs like CorelDraw. Therefore I would be grateful to hear experiences about high-end graphics boards and Windows application programs. Tom Blom Univ. of Helsinki, Finland
gpsteffl@sunee.waterloo.edu (Glenn Steffler) (04/06/91)
In article <1991Apr4.170207.5822@cc.helsinki.fi> blom@cc.helsinki.fi writes: >Does graphics co-processor speed up Windows applications? Quick answer: NO Long drawn out solution: Wait until a fast EISA or MCA adapter implements many of the Windows calls on board. The current crop of 34010 co-proc'd boards suck largely. They don't implement lines, stretchblits, brushes or any of the other Windows primatives on board. They only implement stuff like simple blits. The drivers for these boards are slower than the V7vram board we have at work...(and the ATI Wonder is almost as fast). I have yet to see the XGA adapter (as we have no MCA machines), but from what I have heard it is plenty fast on bitblts. I doubt they buffer line calls or implement internal stretchblits. Still don't have an idea how much faster in 256 colour mode it is than a standard 640x480 VGA adapter. >I am trying to find out ways to improve the screen drawing speed >of CorelDraw. Tests with 386 and 486 systems have shown that a >more efficient processor or increased amount of RAM have only a >marginal effect. Try a large cache (256->512k) with fast < 20ns access time. Since most of the operations would be in cached memory you get big time speed improvements. I have two 386/25's : one has 256k of nasty 20ns cache, the other has 64k of 35ns cache. The former machine rates the same approximately on Norton SI 4.5, but is more than twice as fast on graphics calls and the like. (same video/etc) >Now I am considering video boards with a graphics co-processor >like the NEC Graphics Engine. Tests (e.g. PC Magazine 4/91) >indicate that they would have a significant effect on drawing >speed with MS-Windows. This doesn't necessarily apply to application >programs like CorelDraw. Therefore I would be grateful to hear >experiences about high-end graphics boards and Windows application >programs. I have seen this board, and was seriously unimpressed by the Windows video driver. The DOS demo was ungodly fast (almost as good as a silicon graphics work station demo I saw last year). >Tom Blom >Univ. of Helsinki, Finland -- Windows Sumo Wrestler "Bo doesn't know software" - George Brett --(Windows 3.0, a combination of modern moodring technology and voodoo)-- "I guess she had a way, of making every night seem bright as day" `I Don't Believe In Love` -Queensryche (Oper. Mindcrime) Glenn Steffler
flusekw@ucs.indiana.edu (WILLIAM FRANKLIN FLUSEK) (04/10/91)
>In article <1991Apr4.170207.5822@cc.helsinki.fi> blom@cc.helsinki.fi writes: >>Does graphics co-processor speed up Windows applications? >Quick answer: NO >Long drawn out solution: (Stuff deleted ...) >I have yet to see the XGA adapter (as we have no MCA machines), but >from what I have heard it is plenty fast on bitblts. I doubt they >buffer line calls or implement internal stretchblits. Still don't >have an idea how much faster in 256 colour mode it is than a standard >640x480 VGA adapter. I have seen it run and it looks to me to be faster, but I haven't had the time or inclination to do some actual testing. I do like the 640x480 64K color mode quite a bit. (1024x768 256 color is not much better, but is faster than my 8514/A board). (Stuff Deleted) >>Now I am considering video boards with a graphics co-processor >>like the NEC Graphics Engine. Tests (e.g. PC Magazine 4/91) >>indicate that they would have a significant effect on drawing >>speed with MS-Windows. This doesn't necessarily apply to application >>programs like CorelDraw. Therefore I would be grateful to hear >>experiences about high-end graphics boards and Windows application >>programs. I think there is some interesting information about the ATI 8514/Ultra board in the 4/16/91 PC Magazine (page 43). They show it to be much faster than the NEC MultiSync Graphics Engine on some Windows operations and slightly faster on others without the Crystal Fonts. With these fonts, it is still faster in some things. I am hoping to get one of these in the next month or so (when I can afford one and a 1024x768 non-interlaced monitor.). But between the speed, the non-interlaced image, Bitstream font support and MCA availability, it is definitely on the horizon. (This board has both connectors on it, MCA and ISA 16-bit.) >>Tom Blom >>Univ. of Helsinki, Finland >-- >Windows Sumo Wrestler "Bo doesn't know software" - George Brett > --(Windows 3.0, a combination of modern moodring technology and voodoo)-- >"I guess she had a way, of making every night seem bright as day" >`I Don't Believe In Love` -Queensryche (Oper. Mindcrime) Glenn Steffler Bill Flusek, Indiana University Internet: flusekw@ucs.indiana.edu Bitnet: flusekw@iubacs .sig - Hey, I barely have time to do all the playing on my computer I want to do, let alone think of a cute .sig file. So sue me.
rob@pcad.UUCP (Ralph Brown) (04/17/91)
I saw an Ultra 8514 at a show running the BYTE benchmarks next to an IBM 8514A. The ATI was visibly much faster at graphics, for the price it seems very interesting.