wjc@ho5cad.ATT.COM (Bill Carpenter) (12/02/89)
Here is a simple question that probably has a simple but diappointing answer. Is there a way to measure access times on a UNIXpc? (By that, I mean without having to have any extra heavy-duty hardware.) In other words, if I have a drive that claims average seek times of, say 28ms, is there are way to measure if it's in the right ballpark? I'm not interested (much) in measuring stuff like rotation speed of the platters (sounds like an oscilloscope deal, huh?). Tx for any info. -- Bill Carpenter att!ho5cad!wjc or attmail!bill
jbm@uncle.UUCP (John B. Milton) (12/05/89)
In article <WJC.89Dec1203314@ho5cad.ho5cad.ATT.COM> wjc@ho5cad.ATT.COM (Bill Carpenter) writes: ... >In other words, if I have a drive that claims average seek times of, >say 28ms, is there are way to measure if it's in the right ballpark? >I'm not interested (much) in measuring stuff like rotation speed of >the platters (sounds like an oscilloscope deal, huh?). Sure, but it's crude 1. Shutdown UNIX. 2. Boot the floppy diag disk. 3. Press [Return] 4. At the "> " prompt enter "s4test" 5. At the "expert> " prompt enter "6,0" to re-cal the disk 6. At the "expert> " prompt enter "6,5" to do 1000 random seeks. 7. Time out step 6, repeat 5 and 6 a few times I've never done this, and I don't have good documentation on the drives I've got to compare measured values aginst claims. Ok, everybody **WHO KNOWS WHAT KIND OF DRIVE THEY HAVE** do the above random seek test 5 times and send me the time for all 5. Don't forget step 5 between each test, or you'll get crazy times. I'll run it on all the old 5M drives I've got and send out a summary. DO IT RIGHT NOW, or I'll never get it out! I'll cut it off in a week. The above test is NOT a blind seek, after each seek, a header is read to verify that the head got to the right track. John -- John Bly Milton IV, jbm@uncle.UUCP, n8emr!uncle!jbm@osu-cis.cis.ohio-state.edu (614) h:252-8544, w:469-1990; N8KSN, AMPR: 44.70.0.52; Don't FLAME, inform!
horn@rt3.cs.wisc.edu (Mark Horn) (12/09/89)
In article <618@uncle.UUCP> jbm@uncle.UUCP (John B. Milton) writes: >Sure, but it's crude > >1. Shutdown UNIX. >2. Boot the floppy diag disk. >3. Press [Return] >4. At the "> " prompt enter "s4test" >5. At the "expert> " prompt enter "6,0" to re-cal the disk >6. At the "expert> " prompt enter "6,5" to do 1000 random seeks. >7. Time out step 6, repeat 5 and 6 a few times > >I've never done this, and I don't have good documentation on the drives >I've got to compare measured values aginst claims. > >Ok, everybody **WHO KNOWS WHAT KIND OF DRIVE THEY HAVE** do the above I just did this. I have a seagate 251-1 drive that claims to have 28ms seeks. Well I did the test 4 times: test1 40ms test[2-4] 38ms Now, seagate offers a drive called the 251-0. This drive claims to have 40ms seeks. I paid for a 251-1. Did I get ripped off? Anyone who knows better than I, please let me know, because I paid the extra $$ to get the faster seeks, I damn well better get it!!!!! I bought this drive through an ad in Computer Shopper. Everything they sent me says that the drive is a seagate 251-1. The question is given that I'm getting 40ms seeks, did the company I bought this from rip me off, by telling me that they would send a 251-1 and actually sent me a 251-0? Or is there some other facter that is causing this slower seek time, and would it cause it in any other drive? The reason I ask, is if I decided I wanted to buy a drive that advertised Nms seeks, would I have to skew that slightly for my machine? Oh, I have a WD2010 controller chip. I used iv & the fast.o device driver (posted some time ago) to set the steprate to 14 . . . Just in case this matters. - sparkie -- p.s. Please reply to one of the following addresses and NOT the one listed above. Thanks! ___ ___ ___ ___ _ _ _ ___ / __\| . \/ . \| . \| |/ /|_|| _ | "Mothers Against Skunks Driving... \__ \| __/| || _ /| < | || _[ ...because stinking and driving don't mix" \___/|_| |_|_||_|\\|_|\_\|_||___| - heard on a madison radio station ARPA: harier!sparkie@cs.wisc.edu, sparkie@uhura.cs.wisc.edu UUCP: ...{harvard|rutgers|ucbvax}!uwvax!astroatc!nicmad!madnix!harier!sparkie