argh! scsi drive hdtach sucks

dowxp

Diamond Member
Dec 25, 2000
4,568
0
76
iknow hdtach sucks, but i have nothing else to compare it to. my scores come out 9.3/50 on a ibm 36zx 36gb 10k rpm! the 9.3 is fine with me, but why is it 50!?

i thoguht it was the write cache thing, i turned it on, nothing --. should it be turned on? im running on pdvia-ls with a LSI controller. sigh, 4-1 32's, latest lsi drivers.., fresh reinstall.

oh, im on fat32 partition in win2k..

my previous ide-raid scores:

link
 

dowxp

Diamond Member
Dec 25, 2000
4,568
0
76
hmm, tahts around what i have! so hold up. so u say my ide raid had a higher burst than a scsi? what would give me a higher burst then? atlas 10K II?
 

sharkeeper

Lifer
Jan 13, 2001
10,886
2
0


<< hdtach sucks? LOL. Name a better hard drive benchmark. >>



WB99 disk inspection tools. HD Tach's cpu utilisation result is useless because it doesn't allow the user to normalise the transfer rate to a constant. Thus faster disks will have higher cpu utilisation.

Atto bench is good for measuring STR at different block sizes, it also reports write speed without the need for a separate partition and is non destructive.

I/O Meter has its merits. HD Tach is a good fast and dirty way to see if you disks are malfunctioning or if you have something set up incorrectly. (i.e. DMA off, wrong SCSI mode, etc.)

Cheers!
 

Radboy

Golden Member
Oct 11, 1999
1,812
0
0
i agree w/ sharky.

burst is a fairly meaningless metric, as it's limited to data in the cache. you'll notice no real-world perf diffs from better/worse bursts.

even sustained transfers aren't very meaningful, cuz, in real-world apps, a drive rarely sustains x-fer for any significant period, unless you're working with video or audio files, or large gfx files.

it doesn't matter how high a x-rate a drive can sustain, if it spends all it's time *seeking* (which is what typically happens in most real-world apps)

the key metric you want to focus on, and the one that will have the greatest impact on real-world performance, is seek/access times.

yes, atto is a good disk bench.
 

sharkeeper

Lifer
Jan 13, 2001
10,886
2
0


<< << what would give me a higher burst then? >>


a x15 cheetah!
>>



My X15's burst in the mid 60's on an Adaptec 39160. Access time is in the low to mid 6's.

Burst speed is like engine RPM or even HP. Access time is like the quarter mile. :)

Cheers!
 

MichaelD

Lifer
Jan 16, 2001
31,528
3
76
Why is this Atto bench proggy better than HDTach? Not being a smartbutt; I don't know and would like to know if I should use it, instead of HDTach. Is it freeware?
 

Jgtdragon

Diamond Member
May 15, 2000
3,816
19
81
I have an Atlas 2 10k. I think my score is just a little bit better, not sure. Will let you know. I think the score should be the same since both is U160 drive at 10k.
 

dowxp

Diamond Member
Dec 25, 2000
4,568
0
76


<< Just checked your pic, your cpu usage seems abit high for a SCSI harddrive.

Here is my HD Tach score.

http://www.crosswinds.net/~johngt007/hdtach.jpg
>>



if you were talking about my link, its because that is an ideraid score. my scsi is like 2.7%, but oh well.


so the point is now, i shouldnt be worried right? i still havent really done anythign yet since im going to summer school and i just reformatted. perhaps its ok. i get 10/50 about..