Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!dsinc!bagate!cbmvax!jesup From: jesup@cbmvax.commodore.com (Randell Jesup) Newsgroups: comp.sys.amiga.hardware Subject: Re: Hardrives and the A3000 Message-ID: <20249@cbmvax.commodore.com> Date: 2 Apr 91 05:20:54 GMT References: <1991Mar19.222827.16244@en.ecn.purdue.edu> <17727@crdgw1.crd.ge.com> <1991Mar20.220547.4677@en.ecn.purdue.edu> <20023@cbmvax.commodore.com> <4216.tnews@tower.actrix.gen.nz> Reply-To: jesup@cbmvax.commodore.com (Randell Jesup) Organization: Commodore, West Chester, PA Lines: 73 In article <4216.tnews@tower.actrix.gen.nz> johnv@tower.actrix.gen.nz (John Veldthuis) writes: >Quoted from <20023@cbmvax.commodore.com> by daveh@cbmvax.commodore.com (Dave Haynie): >> >> - Got your MASK, MaxTransfer, and Memory Type values set up for the >> fastest transfers. > >Can someone please explain how to use the MASK value. The documentation on >it is very scimpy. >I have an A2620 in my Amiga and I have lockup troubles when DMA'ing to 32 >bit memory with my HardFrame Harddisk and want to set the MASK so that the >HardFrame only uses the 4 meg of Fast memory starting at $00600000 as >buffers There was some motherboard/2630 issue like this, perhaps it affects 2620's as well (dave?) A pullup or some such if my vague recollection is right. As for Mask, leave it to it's default. For a A2000 system, it should default to something like 0x00ffffff, or 0x00fffffe (depending on controller/ setup software). For A3000's, either 0x7ffffffc or 0x7ffffffe are ok for the built-in scsi (you can replace 7 with f as well, really doesn't matter). The real tricks to scsi speed are: Enough buffers (I usually use 30-250, depending on what I use the partition for). Correct mask (you'll find some difference between 0x7ffffffe and 0x7ffffffc, though which is faster depends on what you're doing with the partition.) "good" setup of the drive params (HDToolbox makes a good guess, but often a human with the drive manual can produce a better one. Note that with Zone-recorded drives like quantums, the optimum setup varies across the disk, and the guess HDToolbox makes will be best at the start of the disk. Disk speed is usually fastest at the start of the drive on ZR drives, slowest on the end. My Swift ST1480N goes from 2MB/s on the first 100 meg to 1.5MB/s by the last 100 meg (400Meg, 4400rpm, 3.5" drive)). fast ram: transfers to chip ram are slow, make sure the test program didn't run out and start allocating chip (standard 2M A3000's will often run out with diskspeed). Background stuff: Turn off your commodities, screen blankers, mouse toys, etc, etc. Often cna make a big difference. SCRAM vs non-SCRAM: Noticable difference (haven't checked it with disks, but it makes a ~15% difference to CPU benchmarks). May not be a major factor. Fragmentation: obvious. Less obvious is how full the drive is. If the test pushes it past the 50% mark (even if it's totally non-fragmented), access will slow down, since it won' be contiguous any more. May interact with fragmentation. The only way to get totally safe numbers is with a freshly formatted partition (also means files/data will be near the root block and bitmaps, speeding writes). Caches: obvious. On 2620/30's, you want to use SetCPU to put stuff in ram instead of ROM. Not an issue on the 3000. Program: I recomment DiskSpeed, though it needs some improvements (%CPU used, differing offsets of transfers, noticing failed writes, check small-file read/write as well as large file with small buffers etc). DiskPerf, diskperfa: just say no. I'm sure I'll think of more... -- Randell Jesup, Keeper of AmigaDos, Commodore Engineering. {uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.commodore.com BIX: rjesup Disclaimer: Nothing I say is in anything other than my personal opinion. Thus spake the Master Ninjei: "To program a million-line operating system is easy, to change a man's temperament is more difficult." (From "The Zen of Programming") ;-)