home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Shareware Grab Bag
/
Shareware Grab Bag.iso
/
090
/
rlltests.arc
/
RLLTESTS
Wrap
Text File
|
1986-06-30
|
4KB
|
126 lines
These are the results of some testing I did late last winter on
three different hard drive controllers. I compared a Western
Digital "Cobra" card, an Adaptec 2010A (predecessor to the 2070A)
and an Adaptec 2070A (the famed RLL controller). I wanted to
confirm for myself the claims being made about the 2070 and compare
it against other products available to me.
In order to measure the speed of the controllers with as little
system overhead as possible, I set up a small RAM disk to contain
the programs I needed for the testing. The files I used to auto-
mate the procedure are listed below.
For the actual speed test, Norton's DT (Disk Test) was used. I
chose this program because it appeared to read the whole disk in
a sequential manner. I divided the total size of the disk as
reported by CHKDSK by the number of seconds DT took to execute
to get the kbits/sec values in the charts.
An AT&T PC 6300 with PC-DOS 3.10 was used for all tests. I have
found that different brands of machines may give different results
and may even vary between machines of the same line.
The best interleave for your computer is something you have to
determine based on what your applications are. Adaptec states in
their manual:
"2 to 1 is the optimum interleave factor for the IBM PC/XT. This
typically will give the highest performance on the IBM PC/XT. The
optimum interleave factor is a function of the number of devices
making requests on the bus, and the number of blocks transferred
at one time by the applications software. Experimentation with
different interleave factors is the best way of determining the
optimum interleaving factor for your application."
Jerry Wallace
Mpls, MN
---------------------------------------
You can find Jerry and others at Terrapin Station BBS!
A BBS for programmers and technically oriented users!
612/623-0152 24 hours/day 1200/2400
---------------------------------------
DISKTEST.BAT:
time 00:00:00
d:dt c: < d:kb
time
KB:
d
---------------------------------------
Seagate ST-225
Adaptec 2070A Adaptec 2010A
Total
bytes: 31,344,640 21,309,440
Inter Kbits Kbits
leave Time /sec Time /sec
----- ---- ----- ---- -----
3 7:15 576 5:55 480
4 7:57 526 6:33 434
5 8:39 483 7:14 393
---------------------------------------
Rodime 204e
Adaptec 2070A Adaptec 2010A
Total
bytes: 32,571,392(x2) 22,142,976(x2)
Inter Kbits Kbits
leave Time /sec Time /sec
----- ---- ----- ---- -----
3 7:29 580 6:03 488
NOTE: The WD Cobra does not appear here because I did not
have the necessary table on the controller BIOS. I
did not run any more than an interleave of 3 because
previous results seemed to show that this was the best
interleave for my machine. (Plus it took 50 minutes
per test; I didn't have the patience!)
---------------------------------------
MMI 3.5" 10MB
Adaptec 2070A Adaptec 2010A WD Cobra
Total
bytes: 15,572,992 10,584,064 10,584,064
Inter Kbits Kbits Kbits
leave Time /sec Time /sec Time /sec
----- ---- ----- ---- ----- ---- -----
1 -- --- 3:38 388 6:44 210
2 -- --- 3:56 359 6:50 207
3 2:34 809 2:12 641 7:05 199
4 2:54 716 2:33 553 7:26 190
5 3:15 639 2:54 487 6:33 215
6 3:23 614 3:14 436 3:15 434
7 3:43 559 3:22 419 3:35 394
8 4:04 511 3:42 381 3:57 357
9 4:24 472 4:03 348 4:04 347
NOTE: These really are the values I got last winter! They
seemed fast to me, so I checked them again and got
the same results. These were done with an un-modified
DOS, so the cluster size is 4k. On the same disk with
a 2k cluster size, the 2010A at an interleave of 3 runs
at about 480 Kbits/sec.
Also, the 2070 does not offer interleaves of 1 or 2 in
it's configuration procedure, hence they are not presented
here.