Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!watmath!clyde!cbatt!ucbvax!CORY.BERKELEY.EDU!dillon From: dillon@CORY.BERKELEY.EDU.UUCP Newsgroups: comp.sys.atari.st Subject: Re: 70 folder limit? Message-ID: <8701290450.AA16048@cory.Berkeley.EDU> Date: Wed, 28-Jan-87 23:50:44 EST Article-I.D.: cory.8701290450.AA16048 Posted: Wed Jan 28 23:50:44 1987 Date-Received: Fri, 30-Jan-87 01:43:58 EST Sender: daemon@ucbvax.BERKELEY.EDU Organization: University of California at Berkeley Lines: 76 (Note that the article I'm responding to is included after my responce so people who are familar with it can abort their text display) Unfortunetly, relative hardware DMA speeds doesn't really matter when it comes to disk accesses, just as long as they are in the same ballpark. You must also take into account file buffers and the way the DOS optimizes disk space. A friend of mine has a 20Meg drive (forgot the brand) connected to his Mac's SCSI port which goes so fast, that it can read a file as fast as the disk rotates (e.g. one rotation to get a track). Theoretically, you can't get any faster than that. In terms of floppy drives, the original Mac had extremely slow floppies. The Mac+'s floppy drives are quite fast (quite a bit faster than the Atari's or the Amiga's). A 3-1/2 inch floppy spins at 300RPM, and thus you can theoretically get about 5 tracks a second throughput (Track= a given side, thus for a double sided drive it would take a rotation to read each side). The Amiga is about 50% efficient in terms of loading long data files from floppy (theoretical max @ 11 sectors/track = 28Kbytes/sec, Amiga gets around 15Kbytes/sec). The test is a simple one... simply start with a blank disk, write out a very large file, then see how long it takes to read it all in. (Anybody care to try on an Atari?). Since your reading large amounts of contiguous data, this doesn't take into account nominal OS buffering optimizations. The relative slowness of the Amiga's workbench icons popping up is due to the fact that the directory structure for an Amiga floppy is optimized for finding files rather than searching a directory. C-A made a big mistake by placing all the icons in separate files, requiring a directory search. However, this has nothing to do with nominal reading/writing speed for files. So, in retrospec, having fast hardware is nice and all, but its advantages become moot when limited by the physical device (floppy or harddrive) speed. The floppy drive is going to transfer at 250Kbits/sec no matter how fast the DMA interface is. -Matt >In fact, they make a fair point of stressing that the ST disk ports are about >the fastest around. If you look at the right specs, this seems to be true -- >for instance, Byte lists the top speed of data transmission of the DMA port >at about twice the top speed of the Mac Plus's SCSI port (I forget the >figures). Furthermore, one of the ST articles actually says that most hard >disks for the 512K Mac transmit at a speed barely above the ST's FLOPPY >ports, and that (this is the crucial item) the Mac's (serial) floppy ports >transmit at only a bare fraction of the blindin speed of the ST's (parallel) >floppy ports. > >Now, I should say that this was one of the original pieces of information >that attracted me to the ST. I find working with a floppy-based Mac about >as pleasant as I would imagine having boiling acid dripped into my eyes -- >after a while, the room spins along with the incessant whirring of the >floppy drives, which seem to take about 75% of your time. Furthermore, >my introduction to the Amiga was equally unpleasant -- if anything, judging >by the speed that a Window displayed its icons, Commodore seemed to have >achieved the impossible: floppy accesses even SLOWER than a Mac's. So >when Byte happily informed me that the ST's architecture, with its DMA >chips and its glue chip and its high-speed ports and its 8 MHz clock rate, >was a speed-machine, I thought "this is the 68K for me!". >Then there's the hard disk; all the dealers I spoke to insisted the >ST's DMA port was the fastest in the personal computer world -- certainly >Byte insisted it was twice as fast as a Mac+ SCSI; but the word was that >it was also even faster than an AT's drive. Why is it, then, that booting >the system with five desk accessories seems to take a million years? Why >does starting up a program seem to take so long, and MOST importantly, >how can it possibly be that copying files from one directory to another >on the HD seems to take such an amazingly long time?! If anything, >judging by my experience on XT's, I would be forced to say from copies >that the HD is no faster than an XT HD. Oops. >Well, these observations are (except for the Mac/Magic test) quite >subjective. What I am asking for now from the usenet community is any >objective hardware knowledge out there to shed some light on these >things -- is Byte magazine correct in saying the ST is the fastest of >the 68K machines, even in disk-accessing? What IS the truth about >this?