fmcphers@VTTCF.CC.VT.EDU (Frank McPherson) (12/04/90)
There have been postings to usenet lately claiming levels of speed with quantum 105 meg drives which I didn't think I'd been getting. I downloaded DiskSpeed from the fish archives at an FTP site. I've run the tests on my Quantum 105, which is in the rear bay of my A3000UX. I'm using whatever standard controller comes with the 3000. The is partitioned this way: 9 meg as an AmigaDos partition, named System2.0: The rest of that particular disk has been scarfed up by Unix. My question is basically this: What are the ways of improving peformance of my hard disk? I've defragmented the thing by backing it up and restoring it, so that shouldn't be the problem. When I used to dabble in the PC world, there was a nifty little parameter called Interleave. I have a pretty non-technical idea of what this is, I.E. I know the explanation of how the tracks are seperated on the disk, etc. Are there any existing utilities on the Amiga for determining the optimum interleave of a hard disk unit? If anyone is seriously interested in pursuing this thread farther, I can upload the results of my disk speed tests. (By that time I'll have the results for the 105 and the 170 (170 is entirely amigados)) I'd REALLY like to find out more about this. Thanks! -- Frank McPherson INTERNET: fmcphers@vttcf.cc.vt.edu --
milamber@caen.engin.umich.edu (Daryl Scott Cantrell) (12/04/90)
In article <9012040118.AA116694@vttcf.cc.vt.edu> fmcphers@VTTCF.CC.VT.EDU (Frank McPherson) writes: [...] >used to dabble in the PC world, there was a nifty little parameter called >Interleave. I have a pretty non-technical idea of what this is, I.E. I know >the explanation of how the tracks are seperated on the disk, etc. Are there >any existing utilities on the Amiga for determining the optimum interleave >of a hard disk unit? [...] Interleave isn't quite the same under SCSI. With an ST-506 (the usual PC protocol) controller/drive, the controller works "closely" with the drive, telling it things like which track to seek to and which surface to read. On a SCSI system, the hard drive has an something called an imbedded host adap- tor. This is part of the actual cicuit board "on" the drive, and the theory is that the SCSI controller in your A3000 only sees the drive as a series of blocks numbered 0 to n. The circuitry on the drive is built with the opti- mum interleave and such in mind and does the appropriate translations. One side affect of this is that if you have a drive with, say, 2 surfaces, 10 cylinders, and 10 blocks per track, it doesn't matter if you lie to the controller and tell it that the drive has 1 track with 200 sectors on it or 20 surfaces with 5 cylinders of 2 sectors, as long as you get 200 blocks total. >-- Frank McPherson INTERNET: fmcphers@vttcf.cc.vt.edu -- -- +---------------------------------------+----------------------------+ | // Daryl S. Cantrell | These opinions are | | |\\\ milamber@caen.engin.umich.edu | shared by all of // | | |// Welcome to the Machine. | Humanity. \X/ | +---------------------------------------+----------------------------+
blgardne@javelin.es.com (Blaine Gardner) (12/04/90)
fmcphers@VTTCF.CC.VT.EDU (Frank McPherson) writes: >upload the results of my disk speed tests. (By that time I'll have the >results for the 105 and the 170 (170 is entirely amigados)) I'd REALLY like >to find out more about this. Thanks! I replaced the stock Q105S with a Q170S on my A3000. With DiskSpeed 3.1 the peak speeds on the 105 were around 800K/sec, and with the 170 it was just over 1M/sec. The 170 mechanism is different from the 105 (it's taller and uses a lot more power), and is a fair bit faster too. -- Blaine Gardner @ Evans & Sutherland 580 Arapeen Drive, SLC, Utah 84108 blgardne@esunix.UUCP BIX: blaine_g {decwrl, utah-cs}!esunix!blgardne PLink: BlaineG DoD #0046 My other motorcycle is a Quadracer.