news@udenva.cair.du.edu (netnews) (07/09/90)
i have an Imprimis St-1100 (i think); it is an IDE 128MB before i got it , i saw it perform 15.5 millisecond seeks. (average) on my machine, i doesn't do better than 16.4; what factors cause the variation? it was in an SX at the store, my machine is a 386-25. thanks tim
cjp@beartrk.beartrack.com (CJ Pilzer) (07/09/90)
In article <13423@udenva.cair.du.edu>, news@udenva.cair.du.edu (netnews) writes: > i have an Imprimis St-1100 (i think); it is an IDE 128MB > > before i got it , i saw it perform 15.5 millisecond seeks. (average) > on my machine, i doesn't do better than 16.4; > > what factors cause the variation? it was in an SX at the store, my machine > is a 386-25. > Since the disk is over 32mb, there is a good chance that it was partitioned into more than one partitions. The size of the partition will affect the average seek rate as measured by most if not all tests. After all if the partition is smaller there are not as many cylinders for the head to move from the first cylinder to the last cylinder of a logical partition. The average seek rate is about one-half the time, it takes for the head to move from from the first to the last cylinder. Therefore, if the partitions were of a different size, the results of the tests would be different. Of course, if the partitions as measured in both tests were the same size I guess you need some other explanation. -- cj
jca@pnet01.cts.com (John C. Archambeau) (07/10/90)
news@udenva.cair.du.edu (netnews) writes: >i have an Imprimis St-1100 (i think); it is an IDE 128MB > >before i got it , i saw it perform 15.5 millisecond seeks. (average) >on my machine, i doesn't do better than 16.4; > >what factors cause the variation? it was in an SX at the store, my machine >is a 386-25. The seek shouldn't vary. But there are disk caching programs that can make the drive appear that it has a faster seek than it really does. It wouldn't surprise me if they cached it in the store on the sales floor demo unit. 16.4 ms isn't really that bad of a seek time, in fact, it's quite good. // JCA /* **--------------------------------------------------------------------------* ** Flames : /dev/null | Small memory model only for ** ARPANET : crash!pnet01!jca@nosc.mil | Unix? Get the (*bleep*) out ** INTERNET: jca@pnet01.cts.com | of here! ** UUCP : {nosc ucsd hplabs!hd-sdd}!crash!pnet01!jca **--------------------------------------------------------------------------* */