pasquale@sgl (Pasquale Leone) (01/07/91)
I've seen many articles in the pc trade mags comparing the speed of various VGA cards and other articles comparing the speed of various 8514 cards and yet other articles comparing the speed of TMS34010 cards. As the owner of an ATI VGA1024 card who is thinking of buying a smart graphics card(8514, etc) I would like to know if it is worth the money. What I mean is how much faster is an 8514 going to be than my VGA. 10% ? Twice as fast ? 10 times as fast? Can anybody out there who has access to both a VGA equiped machine and one with a graphics coprocessor board please tell how much faster one is than the other. As a test consider having a full screen window (say notepad) in windows 3.0 that you then iconize. All of the backgorund windows and icons must be redrawn. How fast is it? Or consider just moving a window diagonally across the screen. If this has already been done in some article somewhere please let me know. pasquale@sgl.ists.ca
gpsteffl@sunee.waterloo.edu (Glenn Patrick Steffler) (01/08/91)
In article <16972@ists.ists.ca> pasquale@sgl (Pasquale Leone) writes: > As the owner of an ATI VGA1024 card who is thinking of buying > a smart graphics card(8514, etc) I would like to know if > it is worth the money. What I mean is how much faster is > an 8514 going to be than my VGA. 10% ? Twice as fast ? 10 times as fast? Pasquale, I have over two years experience in windows 3.0, and in this time I have used VGA boards from several different manufacturers, used an 8514/a PS/2, and tested a 34010 card. Given the phenominal speed of the DOS demos for both the 8514/a and the 34010 card, I was *extremely* dissapointed with the lack of performance realized in the windows environment. Most applications "massage" output before going to the video driver in order to make the update look smooth. Unfortunately, this renders graphics co-processors almost useless, as this pre-processing step is done entirely within the CPU memory by the main processor. The only way a fast video co-processor could help is if it could do the bit copy operation from main memory to video memory faster than say a simple VGA card which has no special hardware. Unfortunately, the main PC bus is so damn slow it inhibits any quick transfer of data from main memory to video memory. The number of wait states to write a single byte (be it one or two pixels) is almost 20!! As opposed to mabe 1.2 on average for main CPU memory. Therefore, from this information, and by the demos I have seen, I would recommend you hold off on a graphics board purchase, and possibly add more memory to your system, or increase the processor speed. Both these will allow windows programs to work with internal bitmaps, and keep more data around which would be discarded in a low memory situation. > Can anybody out there who has access to both a VGA equiped machine > and one with a graphics coprocessor board please tell how > much faster one is than the other. I would say that the only speedup you would realize would be with the` 256 colour VGA drivers. They are not optimized as yet, and in the next few months 30-40% speedups will be available by upgrades from the manufacturers, as microsoft releases better DDK kits. The internal bitmap drivers will be much faster in the next release for the 256 case. (rumour again...but founded on stuff I'va seen). >pasquale@sgl.ists.ca -- Co-Op Scum "Bo doesn't know software" - George Brett "The galaxial hearth steams the sea as the sky blood red embrasses darkness" -John Constantine (HellBlazer) Glenn Steffler
dave@exactus.UUCP (David Salas) (01/08/91)
> > Can anybody out there who has access to both a VGA equiped machine > > and one with a graphics coprocessor board please tell how > > much faster one is than the other. > > I would say that the only speedup you would realize would be with the` > 256 colour VGA drivers. They are not optimized as yet, and in the next > few months 30-40% speedups will be available by upgrades from the > manufacturers, as microsoft releases better DDK kits. > > The internal bitmap drivers will be much faster in the next release for > the 256 case. (rumour again...but founded on stuff I'va seen). > I have a NEC Multisync Graphic Engine (34010 and VGA on the same card). I also have both an ATI VGA wonder and a Genoa 5400 board. I have different Windows directories for each one of them (all of this working on a Lantastic Network).. I have my DGIS driver installed on the 34010 (MGE Card), at 1024x768 mode. Albeit slower than I care for sometimes, it is actually faster somewhat faster than the 4 bit (16 colors) drivers of both the ATI and the Genoa. I have installed the 256 color driver for the ATI, but found the performance to be totally unaceptable. Now, there is something that bothers me a bit, and perhaps you may have a good answer or explanation to it. I have been using Toolbook quite a bit lately, and when I open certain books, the 'opening screen' (the one showing the title of the book with the Asymetrix stuff), the screen speed of this operation is so slow that I have to wait for as much as 30 seconds for the whole book window to update. When I do this same operation on my VGA driver, this operation happens significantly faster. Is there any reason for this to be? Also, is it reasonable to expect that Windows 3.x will sometimes make better use of graphics co-processors like the 34010 or the 34020? +---------------------------------------+---------------------------------+ | David Salas | Exactus : David Salas | | President | Genie : EXAC-DAVE | | Exactus Information Service | UUCP : exactus!dave | | (707) 524-2548 @ 2400 (8N1) | Fax : (707) 524-2546 | | (707) 524-2553 @ 9600/1400 (V32/HST) | Voice : (707) 524-2547 | +---------------------------------------+---------------------------------+
rb9a@kelvin.seas.Virginia.EDU (Raul Baragiola) (01/13/91)
In article <5XFgV3w163w@exactus.UUCP> dave@exactus.UUCP (David Salas) writes: >[stuff deleted] > >Now, there is something that bothers me a bit, and perhaps you may have >a good answer or explanation to it. I have been using Toolbook quite a bit >lately, and when I open certain books, the 'opening screen' (the one >showing the title of the book with the Asymetrix stuff), the screen speed >of this operation is so slow that I have to wait for as much as 30 seconds >for the whole book window to update. When I do this same operation on my >VGA driver, this operation happens significantly faster. Is there any >reason for this to be? > Asymetrix (who thought of that name!!!) Toolbook was designed for the Intel 80986 processor. :-) Another way of looking at it: we got a free extra disk with Win 3 (that's the use for us with a 386/20). BTW, I agree with the comment about the "feature" of the hidden wallpaper. Raul -- Raul A. Baragiola \Internet: raul@virginia.edu Dept. Nuclear Engnr. and Engnr. Physics \Phone: (804)-982-2907 University of Virginia, Charlottesville, VA 22901 \ Fax: (804)-924-6270