Post your hard drive benchmarks!

weescott said:
On a P5K Deluxe and Vista 64 Ultimate:

weescott-HDTune.jpg
Click to expand...

That's really good for a 7200k drive. :cool:

Will be interesting to see how good the other makes do as well.
 
Why do RAID-0 and RAID-5 results here have slower seek times than most single drive setups?
Click to expand...

This is because every drive in the array needs to seek, and the operation is not complete until the slowest has finished. If you look at some of the HDTune benches, like the one above me, you will notice that the access times (yellow dots) deviate a fair bit from their average. If you imagion that you had to run a seek for a particular part of the disk multiple times, and could only accept the slowest, that's basically what's going on. The more disks you have, the more likely you are to get a slow seek.
 
Yea thats fair enough though I think you can get fancy syncronised drives or maybe thats what raid 2 used to do, not sure.

The transfer rate should always rise shouldnt it or is that only true when each drive has its own controller.
I think my drives must be conflicting or something :/ Each drive is 56meg/s so each Ide socket allows for 133meg or 100 at least, 50 from each drive at the same time so it shouldnt screw up raid but it is :confused:
 
Multiple drives on one cable is pretty bad for raid, because iirc only one device on the cable can be accessed at a time, and the performance advantages in raid come from being able to access multiple devices in parallel.

It will be strangling your array because whilst one disk is able to go at full speed, the second and third are having to take turns to send data on the cable, so your array performance is roughly half what you'd get if all three drives had their own cable.

Your best bet is to pick up a cheap 2 drive pci ide controller card and run software raid. OCUK only seem to be selling sata cards, but i'd expect it'd cost less than £5 on ebay.
 
Last edited:


Couldnt remember where i had posted my previous screenshot so uploaded again.

Sorry if i am being a bit dense but why is the transfer rate only around the 100MB/s when I have enabled SATAII on both drives so theoretically they should be running at 300MB/s.

Cheers very much.
 
SATAII only really effects the burst transfer rate from the hard disk cache, even top of the line 7200 drives can't read from the platters at more than about 80MB/s.

Your results do seem a little low for RAID0 though, i'm guessing it's using disks a few years old. Burst rate should be a lot higher also, blame your RAID controller for that. If it's plugged into the pci bus, the max you would ever see is 133MB/s, and usually a bit less than that becasue other stuff uses the pci bus.
 
Last edited:
Thanks for explaining Zarf, I actually have 4 seperate IDE sockets on my setup so maybe I can sort something out so long as windows can see the Raid card.

I guess the really expensive raid 5 cards have 4 sockets on them, 1 for each drive. My card cost a fiver off ebay :D

Schumi, those graphs hardly look bad. Whats the graph look like with 1 drive only. People spend big bucks to get 300mb/s you wont get that with just two drives
 
Lol, oh dear, they are actually brand new drives, using the onboard RAID controller with my MOBO. Oh well, they seem to be doin their job at the moment, cant complain.
Click to expand...

They wouldn't happen to be 7200.10's would they?
 
The ones Im using are 7200.7.

I tried a dynamic drive array over 3 seperate ports but I cant make it load BF2:PR as fast as a single 160gb samsung. 75 vs 81 seconds.
I do think cache makes a ton of difference to this raid though, for now seek time is everything.

HDtune wont recognise this raid so I cant post a graph, any other software apart from hdtach?
 
6bit8x0.jpg


53ewsj9.jpg



This is a test of logical devices rather the physical disks so the results are less accurate due to caching effects, etc.
Since caching is a real part of general windows use I still think it has some value, could someone give this benchmark a go so I can compare my dynamic drive array.

Download link - http://www.download.com/HardInfo-2005-Enterprise/3000-2651_4-10354010.html?tag=lst-0-1


You need to save the result then click visual compare to get the graph above
 
silversurfer said:
Schumi, those graphs hardly look bad. Whats the graph look like with 1 drive only. People spend big bucks to get 300mb/s you wont get that with just two drives
Click to expand...

Didnt get a chance to test them separately as RAIDed them as soon as i bought them. Have got a Seagate SATA I drive that i benchmarked and was considerably slower:

 
Quote:Originally Posted by Zarf
They wouldn't happen to be 7200.10's would they?



Yes they are, do they have a tendency to run a bit slower then in RAID?
Click to expand...

My hdtach benchmarks look very similar to yours, a fairly flat line capped at around 100MB/s when with 4 drives i should be looking at double that.

After tearing my hair out trying hundreds of different configs/drivers etc i discovered that the culprit is the latest Firmware on the 7200.10's , revision 3.AAK and possibly some of the previous ones.
I've emailed Seagate tech support about it, but it's been a week so far with no reply so i'm not too hopefull that they will send me a fix (or even achnowlege the problem)

What we really need is for a big hardware review site to kick up a stink - personally i get the impression that they released nice firmware for initial reviews and then nerfed it down to something less stressful so they would have less failed drives within the 5 year warranty. Extra Reliability is all well and good, but it's a different product than the one reviews led us to expect.

My mate has revision 3.AAE, and it works properly in RAID.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom