🚀 go-pugleaf

In <4g3q25$2h5g@news-s01.ny.us.ibm.net>, iceware@ibm.net said:

> The long story.

> Using an adaptec 2940w or 2940UW, using our product IWSS I can bench a
>Seagate barracuda ST32550W out at between 6.9->7.2 MB when reading from
>cylynder zero for approx 80MB of data. Now if your drive is delivering
>the 5MB of data per second then the speedup is marginal, it would be
>better to have more memory, in order to reduce paging altogether. The
>only place where this falls apart is, when I state that our product
>benches the drive at that transfer rate, it is from reading the data off
>of the platters, yes, prefetch is still in use, but as the entire track
>is being read in one I/O, and the same track is never read twice, the
>drive is being forced to go back to the platters. Some disk benchmarks,
>are continuously hammering on the drives cache, which artificially
>inflates the numbers.  

I recently benched a 2940 with a Seagate Hawk (2GB, 5400RPM, SCSI2-F) and it
showed transfer rates of 6.7MB/sec for both reads and writes for file sizes
up to 4MB.  (The benchmark creates the file and times the write speed and
the read speed, the tests were done with no software disk caching).  Sounds
to me like there isn't much of a benefit to going to SCSI2-UW :)

John Corrigan
CorrComp

--
jcorigan@teleport.com (John Corrigan)



.. I use OS/2 2.0 and I don't care who knows!
-- MR/2 2.26 #314