hsgj@batcomputer.tn.cornell.edu (Dan Green) (04/26/87)
Here is a question for Perry of ASDG in particular, of interest to other hard disk developers, and for Commodore in general. The Amiga 2000 has this large hole in the front of the machine that is supposed to be for a hard disk. Will ASDG's new hard disk fit inside this hole? I sure hope that this product, and any other hard drives, are internal, as the Amiga 2000 seems big enough as is. Now here's the question for Commodore or other parties in the know. As I understand it (warning - my facts may not be correct) AmigaDOS has a pretty low throughput and can not keep up with fast hard disks. This is supposedly why current hard disks for the A1000 are not as fast as IBM (or is that CBM PC-10) PC hard disks. I also have heard that the Commodore-supplied RAM: driver "bypasses" a lot of AmigaDOS so that it can get extra speed. So the question is, (a) will Commodore be providing an upgraded AmigaDOS that has improved speed for hard disks, (b) will CBM write a HD Handler that will bypass AmigaDOS to get full speed, or (c) will developers such as ASDG write their own handlers to get full speed hard disk performance? Final note: In the business section of a certain newspaper there recently was an advertisement for the Commodore PC-10. Their slogan: "Its not just another Clone.... Its a Commodore!". Don't use this for the Amiga, please... -- Dan Green -- ARPA: hsgj%vax2.ccs.cornell.edu@cu-arpa.cs.cornell.edu UUCP: ihnp4!cornell!batcomputer!hsgj BITNET: hsgj@cornella
grr@cbmvax.cbm.UUCP (George Robbins) (04/27/87)
In article <814@batcomputer.tn.cornell.edu> hsgj@batcomputer.tn.cornell.edu (Dan Green) writes: > >Here is a question for Perry of ASDG in particular, of interest >to other hard disk developers, and for Commodore in general. >The Amiga 2000 has this large hole in the front of the machine >that is supposed to be for a hard disk. Will ASDG's new >hard disk fit inside this hole? I sure hope that this product, >and any other hard drives, are internal, as the Amiga 2000 seems >big enough as is. As I understand it, ASDG's hard disk controller is a board level product that will be offered in both the "zorro" and A2000 board form factors. I don't know if he will be offering a hard disk with the board. The A2000 has mounting points for 3 drives. Two are 3.5" size, one of which will normally be occupied by a floppy, and one is 5.25" size, suitable for either a half-height hard drive or a 5.25" floppy for the PC bridge card. There is also and external MAC+ SCSI compatible connector on the rear of the Commodore Hard Disk Controller card for hooking up externally mounted disks. I don't know if ASDG will have this feature. -- George Robbins - now working for, uucp: {ihnp4|seismo|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@seismo.css.GOV Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
andy@cbmvax.cbm.UUCP (Andy Finkel) (04/27/87)
In article <814@batcomputer.tn.cornell.edu> hsgj@batcomputer.tn.cornell.edu (Dan Green) writes: > >Now here's the question for Commodore or other parties in the know. >I also have heard that the Commodore-supplied RAM: driver "bypasses" a lot >of AmigaDOS so that it can get extra speed. It doesn't really bypass AmigaDOS...rather the RAM: handler has its own file handler which is quite a bit faster because it makes some assumptions about its physical media. >So the question is, >(a) will Commodore be providing an upgraded AmigaDOS that has improved >speed for hard disks, (b) will CBM write a HD Handler that will bypass >AmigaDOS to get full speed, or (c) will developers such as ASDG write >their own handlers to get full speed hard disk performance? We are currently doing bug-wringing on a filesystem handler for hard disks, that knows a bit about hard disks, and is able to get increased performance. (It works with the current AmigaDOS 1.2, through AmigaDOS's ability to use alternate filesystems) (That answers (a) and (b), I suppose. When ? Ah..in the fullness of time, at the appropriate juncture, in..., no, seriously, when we trust it.) > >-- Dan Green -- andy finkel {ihnp4|seismo|allegra}!cbmvax!andy Commodore/Amiga /or/ pyramid!amiga!andy } "Do not meddle with the affairs of wizards, for it makes them soggy and hard to light." Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.
daveh@cbmvax.cbm.UUCP (Dave Haynie) (04/27/87)
in article <814@batcomputer.tn.cornell.edu>, hsgj@batcomputer.tn.cornell.edu (Dan Green) says: > Keywords: Harddisks Amiga2000 AmigaDOS > Now here's the question for Commodore or other parties in the know. > As I understand it (warning - my facts may not be correct) AmigaDOS > has a pretty low throughput and can not keep up with fast hard > disks. This is supposedly why current hard disks for the A1000 > are not as fast as IBM (or is that CBM PC-10) PC hard disks. I also > have heard that the Commodore-supplied RAM: driver "bypasses" a lot > of AmigaDOS so that it can get extra speed. When most people refer to AmigaDOS being slow, they're referring instead to a few certain aspects of the current File System Handler. Based on the way that the current File System Handler stores directory entries, searching for a file based on a wild-card description or listing out a full directory are very much slower than the equivalent under MS-DOS. Finding a known file via a pathname, however, is very fast, based on the same structural reasons. And most of the other features fall somewhere in-between. The floppy drives, most hard disk drivers, and Perry's VD0: are all device drivers that use the floppy File System Handler as is. The RAM: device, however, is a totally different animal, comprised of a completely different file handler (apparently all one piece -- you can't access the RAM disk as an Exec device like you can the trackdisk.device for floppies or other similar drivers that use the floppy file system handler). The advantage to this is that this handler can know more about the device its dealing with than a general purpose handler, and it can even store files internally in a different structure if desired to make things go faster. The disadvantage to an alternate handler is that there are far more commands for a handler to implement than a device driver, so there alot more room for compatibility problems. The RAM: device, however, is not bypassing AmigaDOS at all, its just hooked into AmigaDOS at a handler, rather than device, level.n > Final note: In the business section of a certain newspaper there > recently was an advertisement for the Commodore PC-10. Their > slogan: "Its not just another Clone.... Its a Commodore!". > Don't use this for the Amiga, please... > > -- Dan Green A2000: Its not a PC Clone, its a Computer! -- Dave Haynie Commodore-Amiga Usenet: {ihnp4|caip|rutgers}!cbmvax!daveh "The A2000 Guy" BIX : hazy "These are the days of miracle and wonder" -P. Simon
dillon@CORY.BERKELEY.EDU (Matt Dillon) (04/28/87)
>When most people refer to AmigaDOS being slow, they're referring instead to >a few certain aspects of the current File System Handler. Based on the way >.. (1) If AmigaDOS were to sort the blocks when a user does an ExNext call, a directory listing would not be slow slow (2) If AmigaDOS were to cache whatever-the-hell it's tracking back for (when reading and writing LARGE files.. as in >30K)... tic tic tic traaack traack tic tic tic traack traack, then file load time would almost double, and file save time would also almost double. Fix the easy stuff first, I say. -Matt
ewhac@well.UUCP (Leo 'Bols Ewhac' Schwab) (04/30/87)
In article <8704281918.AA08964@cory.Berkeley.EDU> dillon@CORY.BERKELEY.EDU (Matt Dillon) writes: > (2) If AmigaDOS were to cache whatever-the-hell it's tracking back > for (when reading and writing LARGE files.. as in >30K)... > tic tic tic traaack traack tic tic tic traack traack, then file load > time would almost double, and file save time would also almost double. > ^^^^^^ ^^^^^^ Er, um... Don't you mean halve? Yeah, of course you do.... BTW, AmigaDOS doesn't need to cache file control blocks. Every data block points to the next data block, if any. Someone at MetaComCo was asleep at the editor, I'd say... _-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_ Leo L. Schwab ihnp4!ptsfa!well!ewhac The Guy in The Cape ..or.. well ---\ "Work FOR? I don't work FOR dual ----> !unicom!ewhac anybody. I'm just having fun." hplabs -/ ("AE-wack")
dillon@CORY.BERKELEY.EDU.UUCP (04/30/87)
:> time would almost double, and file save time would also almost double. :> ^^^^^^ ^^^^^^ : Er, um... Don't you mean halve? Yeah, of course you do.... Whoops.. Right. I meant 'halve'. -Matt