Serious "CPU usage" problem showing with HDtach

[computer]

Guys, I have a serious "CPU usage" problem showing with HDtach. Mobo is an Abit KG7 RAID w/900mhz Duron and WinSE. This is for a customer. I had a single IBM 60GXP on it that died yesterday, so I had to replace it with a WD800JB since that was the fastest thing I could get RIGHT NOW (locally from BB for $39.99 AR's). With the other 20gb IBM on it, HDtach showed the usual only 2-4% CPU utilization. With the WD800JB, it shows almost 24%!!!!! Setups are the SAME, single IDE drive on the RAID controller w/CDROM and CDRW on the PRI and SEC IDE controllers. WD drive is partitioned twice x 40gb using 16k cluster size. (I had to change that from 32k default because SE would not run scandisk using default cluster size).

Now the WD drive can not possibly require ~TEN TIMES the CPU utilization??? So what else could the problem be?
Thanks.

(BTW, checking the "System Monitor" shows anywhere from 12% kernel processor usage to 37%, both at rest. Just depends on how many app's were opened prior to System Monitor).

 

[myocardia]

Assuming you reinstalled Windows, I think you forgot to enable DMA for the hd. If you just copied the drive, I would check it anyway. If it were me, even if Windows says it's enabled (you know Windows), I would disable it, then reboot, and re-enable it.

edit: I haven't heard anything about WD drives using much more proc. power than other drives.

 

[computer]

Thanks for the reply. Yes, the WD was of course formatted and Windows installed clean on it (IBM drive was too dead anyway even had I wanted to copy it). DMA IS enabled in the device manager for the "SCSI controller" and "HPT370 ATA100 RAID controller" below it. Actually the options are "DMA If Possible" and "PIO" in the drop down. I changed them all to "PIO" and restarted then ran HDtach again, and got an INCREDIBLY lower transfer time from 51,xxxK to only 31xxK, and the CPU utilization was an ABSOLUTELY MIND BOGGLING 95%!! So, DMA was enabled.

 

[computer]

I switch it back of course to "DMA if possible" ran it again and got the same results as I originally did. I also changed the role of the computer to "Network server" and that didn't help the CPU usage.

 

[computer]

Does anyone know if I can move the HD from the RAID controller to the PRI IDE connector without screwing thing up? I did do this on my P4C800-E Deluxe (moved SATA Raptor from Promise controller to ICH5R and back) while testing it, and it booted up fine and ran ok, but seems like it caused some stability problems. So, I'm a bit leery of doing it to this other PC.

 

[adams828]

i'm curious about this too.. although my system seems to be running fine, i got the same result (~24%). but if it could be faster...

 

[computer]

Do you have the same mobo and HD?

 

[Jeff7181]

It's often wrong... pull up the task manager and watch CPU usage there while you run the test... when I did that it barely showed 5% CPU usage while the test was running, while HDTach reported something like 35%

 

[adams828]

nah, i've got a 2500+ @ 166fsb on an asus a7n8x-x. and a wd 1200jb

 

[Big Lar]

Have you run a virus check lately?

 

[adams828]

Originally posted by: Big Lar
Have you run a virus check lately?

yup, full scan with norton just a few hours or so before running hdtach

 

[Big Lar]

Just, to be sure, why not try www.trendmicro.com and do an online test.

 

[computer]

Jeff, this is WinSE.
Guys, I have a serious "CPU usage" problem showing with HDtach. Mobo is an Abit KG7 RAID w/900mhz Duron and WinSE.

 

[computer]

^

 

[adams828]

i dunno, mine is running winXP. supposedly ppl are saying that the hdtach reading isn't very accurate.. i dunno

 

[computer]

Yeah I saw something to that effect the other day when I was trying to find out about this. They opened the task manager (cont-alt-del in XP then "performance") and compared the CPU usage to that of HDtach while running it,....or running something. It was very different. Jeff above also mentioned that. Unfortunately you can't do that in SE since it doesn't have that function. I also wonder how much of an accurate comparison that would be since we don't know exactly HOW HDtach calculates CPU usage. It may turn out to be accurate....but then again, maybe not. HDtach may be using some kind of "algorithm" to determine CPU usage instead of CPU usage just in HDtach while it's running.

 

[Lynx516]

This may be a stupid surgestion but maybe installing IDE drivers might be usefull

 

[computer]

That's the first thing I always do; install mobo drivers. What was strange here is I updated the HPT drivers and not only the did the CPU usage problem get WORSE, but the benchmarks were also WORSE! I re-installed the OLD drivers, and the CPU usage went back to what it was previously (still really bad) as did the benchmarks go back to what they were, better. NONE of these chip makers for 3rd party mobo controllers ever seen to know what they are doing, be it Promise, Si, or HighPoint.

 

[Jeff7181]

Originally posted by: computer
Jeff, this is WinSE.
Guys, I have a serious "CPU usage" problem showing with HDtach. Mobo is an Abit KG7 RAID w/900mhz Duron and WinSE.

I assume you're talking about Windows ME since there's no such thing as Windows "SE" ... install motherboard monitor and configure it to show the CPU's duty cycle, or whatever MBM calls it.

 

[computer]

You never heard of Windows 98SE??? C'mon!

I could have sworn I mentioned that about the monitor. Guess I forgot. I always install the resource meter and system monitor as a habit. It was fluctuating wildly from ~4% kernel processor usage to 37%! Depended on if I opened up anything first. I never thought to keep it open when HDtach was running since I was in such a hurry to get it shipped out. I was not even sure if the monitor's "kernel processor usage" was the exact SAME THING as HDtach's "CPU utilization".

Big Lar I apologize for not acknowledging your posts. I must be getting this thread confused with another! The PC was never hooked up the net and nothing was installed except drivers off the CD's and HDtach. I got the updated HPT drivers from my PC (download them to it then put on a floppy). So, a virus was never an issue.