PDA

View Full Version : Why is my hard drive performance crap?


Feek
06-01-2007, 20:20
A while ago, I had a nasty on my PC and ended up reformatting the drives and starting from scratch.

Ever since then, I've suffered with poor hard drive performance.

The system is a Shuttle SN41G, all the correct and latest manufacturers drivers are loaded. It has an XP-3200 CPU installed and the drive configuration is thus:

Disc 0 - The boot drive is an Hitachi Deskstar 7K250 [HDS722525VLAT80] 250Gb
Disc 1 - Second drive is a Seagate ST3120023A 120Gb

The CD/DVD is a Samsung SH-S182M which I've just fitted.

I can't remember the exact layout but I think I've got the two hard drives on one IDE channel and the CD on the second

The page file is on the second drive, each drive has only one partition.

Everything is bang up to date. The only thing different from before is that I think I had the drives the other way round with the smaller drive as the boot drive and the larger one as the secondary drive.

But it's just sluggish, it feels slow and sometimes it is actually painfully slow and I have to reboot just to get any kind of performance back in.

Neither drive is fragmented, I have diskeeper running on them both.

The only thing I've just noticed is that both drives are set as basic rather than dynamic. Could that cause performance issues? I can't think of any real reason that it should.

OS is XP Professional, the system has 1Gb RAM.

Any suggestions?

Zirax
06-01-2007, 20:38
Only thing I can think of right now is have you optomized windows by turning off all the un-needed crap? ie messenger etc?

Did you move the hdds around on the ide cables? ie set two to master or having master/slave but plugging them in the wrong order on the cable? Done that before now, while it still works it can get very slow

Mark
06-01-2007, 20:45
Yup - cables is what I'm thinking. Check the master/slave jumpers match the sockets on the cable, and if all else fails, try a different cable (assuming you can - getting at the IDE cable on my SFF system is nigh on impossible without gutting the system).

Basic vs Dynamic should make sod all difference (indeed I'd probably expect Dynamic to be the slower if anything). All my disks are set to Basic and always have been - don't see the point of Dynamic unless you're fiddling with partitions lots.

Feek
06-01-2007, 20:49
No, I've not tweaked windows but it's something I didn't do before but I suppose it could be a cable thing. Call me a n00b but I didn't think it actually made any difference on IDE cables as to which connector the drives went on.

Which should be where?

Mark
06-01-2007, 20:52
Googled...

If using the 80-wire cable, attach the blue end connector to the system board or host controller, the gray middle connector to the Slave, and the black end connector to the Master.

It's supposed to only make a difference if using Cable Select (which I avoid like the plague), but I'd still say it's a good idea to get it right, particularly if you're having problems.

Zirax
06-01-2007, 20:55
I kept forgetting as well. It should go mobo->slave->master.

Handy to remember. Probably most apt for here is "the slave gets spit-roasted by the mobo and master"

Feek
06-01-2007, 21:21
Is it also worth trying to split the drives onto different IDE channels?

Mark
06-01-2007, 21:24
If you've only got two channels and have a DVD/CD-ROM and two disks, then you have the correct setup - disks on primary, CD on secondary. Any other combination (save removing the CD entirely) will yield worse performance on at least one of the disks.

Feek
06-01-2007, 22:24
Right then, system pulled apart, cables checked.

It was:

Primary IDE
Boot drive as master, on only connector on cable.

Secondary IDE
CD as master, on end connector.
Second hard drive as slave, on middle connector.

Now I've changed it all around

Primary IDE
Boot drive as master, on end connector on cable.
Second hard drive as slave, on middle connector.

Secondary IDE
CD as master, on end connector.

Will run it for a few days and see if there's any noticeable difference.

Mark
06-01-2007, 22:27
Sounds good. If the original master cable only had one connector, it might have even been a 40-wire one, which would have been disastrously bad.

Feek
06-01-2007, 22:29
No, it was 80 core. I was using the original shuttle cables and that was the very short one.

I need some HD benchmarking software so at least I can see what they're doing.

Garp
06-01-2007, 22:42
HDTach for the win!
http://www.simplisoftware.com/Public/index.php?request=HdTach

Mark
06-01-2007, 22:42
http://www.simplisoftware.com/Public/index.php?request=HdTach

Think that's what most people use - haven't tried it myself.

Feek
06-01-2007, 22:51
Should have done before and after I suppose, but I'll give it a try.

Thanks both :D

Garp
07-01-2007, 00:36
Should have done before and after I suppose, but I'll give it a try.

Thanks both :D

If you post what you're getting now it should give us some hint at whether its good or not.

Feek
07-01-2007, 01:33
http://img443.imageshack.us/img443/5161/drivesvd5.png

Drive in blue is my boot drive and the red one is my second drive, the one I use mainly for storage of iso files waiting to be burned. I thought that they were the other way around, that the boot drive was a bit slower but I got that wrong.

Page file is on the red drive. Hmmmm.

mejinks
07-01-2007, 01:50
Its not a million miles away in read speed from a SATA 300 drive:

http://www.markgerry.plus.com/bd/hd-tach.jpg

Garp
07-01-2007, 15:15
Its not a million miles away in read speed from a SATA 300 drive:

How'd you figure that then? His burst speed is attrocious! Its barely half what you achieve!


edit: Doh...

Neither of those drives are SATA are they?

My IDE hdd which I use for backups manages 72.7mb/s burst, and 52.2 MB/s average speed which is still better than any of Feek's speeds, so may be some indication of concern.

Feek
07-01-2007, 15:36
My Hitachi drive has a maximum data transfer rate of 100 MB/s and it's nearly hitting that on burst. The sustained data rate is 29.7 to 61.4 so the average of 50.1 MB/s is what I'd expect.

The Seagate again has a maximum rate of 100MB/s sustained is quoted at 27 to 44 so 35.1MB/s is about right.

Shows the limitations of IDE compared to newer hardware.

Feek
20-03-2007, 19:01
Ahhh, Necroposting for the win.

Even knowing the figures were accurate on the old box, it never felt quick, it actually felt sluggish.

http://www.ukrm.org/feek/better.png

This is better! Much faster and it feels blindingly quick.

I'm happy :D