wsdwgk@eutrc3.urc.tue.nl (Guido van Rooij) (11/29/89)
Hello. I recently bought a PC with a 30 meg harddisk, a Miniscribe 3438. The problem is that the harddisk-controller combination (a controller of unknown origin) is very slow, coretest reports 25 K/s. First I believed it was the interleave, but spinrite said it still needed 26-32 revolutions to read one track at any interleave. There must be something wrong... What is the problem, and what can be done about it ?? Greetings guido
bradley.grigor@canremote.uucp (BRADLEY GRIGOR) (11/30/89)
TO: wsdwgk@eutrc3.urc.tue.nl In regards to: %I recently bought a PC with a 30 meg harddisk, a Miniscribe 3438. %The problem is that the harddisk-controller combination (a controller %of unknown origin) is very slow, coretest reports 25 K/s. First I %believed it was the interleave, but spinrite said it still needed %26-32 revolutions to read one track at any interleave. There must be %something wrong... What is the problem, and what can be done about it Guido, I think your interleave must still be wrong. I believe you must do another low-level format, deliberately specify a new interleave factor, then run SPINRITE again to obtain new results. Remember that if your best interleave is 6:1, then anything less (like 5:1, 4:1, etc.) will give aweful performance, while anything higher (like 7:1) will not appear so bad. So, if going from x:1 to x-1:1 causes performance to degrade significantly, then your best interleave factor is x:1. --bag --- ~ EZ-Reader 1.21 ~ Play TelePoker via Bell Alex in Montreal!
ngeow@bu-cs.BU.EDU (Yee Ngeow) (12/05/89)
In article <89120107511745@masnet.uucp> bradley.grigor@canremote.uucp (BRADLEY GRIGOR) writes: >TO: wsdwgk@eutrc3.urc.tue.nl >In regards to: > >%I recently bought a PC with a 30 meg harddisk, a Miniscribe 3438. >%The problem is that the harddisk-controller combination (a controller >%of unknown origin) is very slow, coretest reports 25 K/s. First I >%believed it was the interleave, but spinrite said it still needed >%26-32 revolutions to read one track at any interleave. There must be >%something wrong... What is the problem, and what can be done about it > >Guido, I think your interleave must still be wrong. I believe >you must do another low-level format, deliberately specify a new >interleave factor, then run SPINRITE again to obtain new results. >Remember that if your best interleave is 6:1, then anything less >(like 5:1, 4:1, etc.) will give aweful performance, while anything >higher (like 7:1) will not appear so bad. So, if going from x:1 >to x-1:1 causes performance to degrade significantly, then your >best interleave factor is x:1. Also, are you installing the controller onto an AT/286 type machine? If you are, then it can only make things worse. It is possible that you need something like 1:10 interleave to make it work. Some AT bus are too fast for earlier controllers to catch up, thus you need to slow down the already slow transfer rate. Try first to format 1:6, do only a couple of tracks, check it again, then if it doesn't work, try 1:7 ... all the way down until you get it right. Be patient, it might take some time. But it is worth it. 26K/Sec? That's like a floppy drive!!! Kwong > >--bag >--- > ~ EZ-Reader 1.21 ~ Play TelePoker via Bell Alex in Montreal!