drz@csri.toronto.edu (Jerry Zarycky) (03/02/89)
[ what line eat I am the owner of an Amiga 2000, a Miniscribe 8051S (40 meg SCSI drive) and a 2090A hard drive controller card. I am having trouble loading 640x400 4 bitplane screens from my harddrive. I recall reading an Amazing Computing issue late last year which compared various controller cards and they mentioned that the Commodore card exhibited some peculiar behavior when within Deluxe Paint and a 640x400 4 bitplane screen was being loaded. I assumed the problem was some interaction between Deluxe Paint and the 2090A, but now I find that it is an inherent limitation of the 2090A, since I run into this problem using a simple file display utility. I realize that the card's software uses DMA to do the I/O and that it is largely being soaked up by this sort of screen, but why does it slow down to such an extreme crawl? I remember that the Supra card ran about as fast as the 2090A (using DMA as well) but they had no such problem with these screens. I take it that this is a shortcoming with the software in the EPROMS. So I ask Commodore "Will this ever be fixed?" I know that a workaround is to flip to a low-res screen and the transfer will be quickly completed, but I find this to be unacceptable for a supposedly professional machine. Since Commodore is shipping these cards with the 2000HD and probably the 2500, I'm sure there will be lots of novice users out there who will not understand what is going on and the Amiga will get another bad rap as a "slow" hard drive machine. So what's the solution? Is there a fix for the software in the works? Or should I sell my 2090A card and buy a Supra card (or a Comspec card, or a Microbotics card, or a ....) I'm as loyal as the next guy, but this is getting to me.... Jerry Zarycky Usenet: {uunet,watmath}!csri.toronto.edu!drz CSNET: drz@csri.toronto.edu EAN: drz@csri.toronto.cdn BITNET: drz@csri.utoronto
jesup@cbmvax.UUCP (Randell Jesup) (03/04/89)
In article <8903022014.AA25100@queen.csri.toronto.edu> drz@csri.toronto.edu (Jerry Zarycky) writes: >I am having trouble loading 640x400 4 bitplane screens from my harddrive. ... >I realize that the card's software uses DMA to do the I/O and that it is >largely being soaked up by this sort of screen, but why does it slow down >to such an extreme crawl? Because it has a 64-byte FIFO for buffering, and with 640x{200|400} 4 bitplanes there are no free cycles for DMA while the beam is on screen, and it overflows often. >I take it that this is a shortcoming with the software in the EPROMS. >So I ask Commodore "Will this ever be fixed?" Well, I'd say the problem is with the hardware :-), but there are software workarounds we plan to implement when next we upgrade the driver. -- Randell Jesup, Commodore Engineering {uunet|rutgers|allegra}!cbmvax!jesup
poirier@giants.dg.com (Charles Poirier) (03/07/89)
In article <6138@cbmvax.UUCP> jesup@cbmvax.UUCP (Randell Jesup) writes: <In article <8903022014.AA25100@queen.csri.toronto.edu> drz@csri.toronto.edu (Jerry Zarycky) writes: <>I am having trouble loading 640x400 4 bitplane screens from my harddrive. <... <>I realize that the card's software uses DMA to do the I/O and that it is <>largely being soaked up by this sort of screen, but why does it slow down <>to such an extreme crawl? < < Because it has a 64-byte FIFO for buffering, and with 640x{200|400} 4 <bitplanes there are no free cycles for DMA while the beam is on screen, and <it overflows often. For me, "often" is 9 times out of 10, i.e. tenfold slowdown. Oops. <>I take it that this is a shortcoming with the software in the EPROMS. <>So I ask Commodore "Will this ever be fixed?" < < Well, I'd say the problem is with the hardware :-), but there are <software workarounds we plan to implement when next we upgrade the driver. Fine, but *when* will that be? I've been waiting for a driver that fixes this problem a good 8 months. My 2090A is soon going sailing off a nearby cliff. Charles Poirier