raymund@sci.UUCP (Raymund Galvin) (10/09/87)
StepRate (or similar programs) can't help the overall disk performance as much as some people would like because the disk performance is dominated by by the time it takes to read a track after the head is positioned at the track. The drive step rate is not the dominate factor. Consider this: Total time to read a sector = Seek_time + Head_settle_time + Latency_time + Read_time For the amiga: Seek_time = 3ms/track (amiga 3.5 drive) Head_settle_time = ???? (usually negligible compared to the time it takes to do a multiple track seek and data read) Latency_time = 0 (neat amiga feature) Read_time = 200ms (one revolution at 300 rpm) For an 80 track drive: average random seek = 80/3 tracks (I think) average random seek time = 80 ms Thus, Total average time to read a sector = 80ms + 0ms + 0ms + 200ms = 280 ms If the drive step rate were reduced to zero, Total average time to read a sector = 80ms + 0ms + 0ms + 0ms = 200 ms What does all of this mean??? If you had a randomly fragmented disk, the BEST you could hope for is to reduce the effective average reading time by 28%. Reasons why you won't see a 28% effective speedup: 1. You can't make your drive step infinitely fast. 2. Even though amiga disks seem to get fragmented easily, it is unlikely that fragmentation will get as bad or worse than a randomized disk. 3. Disk buffers. 4. Multitasking. The cpu might be busy and not emitting step pulses to the drive at the correct rate (try resizing a window while disk activity is going on - the step rate goes way down). I hope this has helped, Ray Galvin