hard drives & raid - benchmark and compare!

[supershanks]

How do you guys get numbers like these? Mine are so low compared to yours

Don't get to concerned it's all about averages.

 

If i have a single raid across 4x80gb I get this:-

The average of 196.5 is across all 329gb. Note Max is 240 at start of drive & 120 at end the speed decreasing as access time increases.

 

Now my fast stripe uses the 1st 68gb of the array (1st 17gb of each drive) so is using the fastest part of the drive. The comparison is almost an exact match

The average of the fast stripe is of course much higher.

 

Similarly the average for the 2nd stripe =186.1 as you expect lower than the average for the whole array 196.5.

 

The Burst values in these shots seem screwed & are down to the impact of having Write back Cache enabled.

 

So in fairness if you run with a single raid & create a partition the same size as my "fast volume" you will all things being equal:) hit the same performance, it's just that the HD tools only work with physical drives & raid volumes & so the speed of your fast partition is just aggregated.

 

The main advantage of Matrix is the ability to have a 2nd raid volume operating on raid 1,5 etc., so you can maximise Performance Volume0 & security Volume1

 

luck:D

[Polizei]

So if I had 2 RAID stripes, like 2 RAID 0 arrays on the same 2 drives, I would start seeing numbers like yours?

[supershanks]

Like i proved in pic 2, the outer array will show a much faster average. You won't be toatally there as i'm using x4 drives compared to your x2, but yep you'll see faster numbers the 2nd stripe if raid 0 will be lower but not by much, like i said it's around averaging.

[Polizei]

I get it now... Ok. If Im going to dual boot with XP and Vista, Im going to stick XP Pro on the earlier (outer) partition so its faster because Im not sure how Im going to like Vista so it can run slower if need be. I can always format the Vista partition and merge it with the rest of the diskspace.

 

I dont know... Ive got time to figure that out, not getting Vista till April or May when I can get it though my college for $20.

 

The bigger numbers become moot at a certain point I think... After a certain speed, you dont notice the difference anymore. If I could crack 1GB/s with my drives, I dont think I would notice anything faster than that like most people cant tell a difference between 80FPS and 200FPS.

[supershanks]

yes that makes good sense like i told grafton earlier, i'm running

XPSP2 - Fast Raid

 

Slow Raid:-

XP64 60.5Gb

VISTA32 RC2 60.5gb

VISTA64 RC2 60.5gb

Short term Data 60.5gb

 

When i move to Vista?? , or get fed up of playing with the other os's, or start filling up my Data drives. i may consider switching to raid1 or 5 for the 2nd stripe.

 

luck:D

[Tonezorz]

So my main rig went down yesterday, (just the PSU so far, being RMA'd) and it's sparked some interest in me to re-do things a little.

 

I haven't reformatted the sucker since my original build, where I have the two Hitatchi 80GB's in Raid-0 16k/4k

 

I have a third 80GB sata drive that i'd like to throw into the mix once I get my replacement PSU back, but my concern is it's not SATA II compatible, while the other two Hitachi's already are sata II enabled.

 

Should I disable the Sata II functionality of the Hitachi's before i throw the Segate into the mix? (For a 3 drive Raid-0)

 

Also, should I go with something like a 32k stripe 4k cluster? I know the "double the drives double the stripe" rule, but I'm adding just 1, not doubling..

 

Thanks!

 

EDIT: Also, which/where are the newest nf4 raid drivers?

[Polizei]

I cant speak from experience, so if you have time, why not try both? Post test results for both if its not trouble.

 

I would disable SATA II on the HItachis just so there isnt any conflict... Im paranoid of troubleshooting errors because Im not the greatest at it. I couldnt turn my computer on when I first built it and it took me half an hour to figure out I had the front panel connectors hooked up wrong - HDD LED and power switch reversed.

[supershanks]

Funny you should say that

but my concern is it's not SATA II compatible

I got my drives wednesday & was a bit busy , wasn't until saturday that i noticed in Matrix storage console that i'd forgotten to enable sata2 in my new drives:rolleyes: Think my 1st matrix post here is with the mixed drives.

 

But other than that my drives are identical (1 has a different firmware version) so only you can really say how thoise drives will mix 'n match.

I'd certainly have a mind to try it. Though the cost of an extra hitachi sata2 a'int that great that's why i went 4 it.

 

If your just going for a single stripe i'd look larger 64k but if using it for a fast stripe use smaller. Though in fairness as i showed earlier i started with 128 then tried 64 which didn't really show any benefit at all.

 

luck:D

[Polizei]

Like I said before, when I pick up Vista, Im going to dual boot with it and XP Pro and put both on 32k stripe sizes and see if that speeds things up. I might try 16, but if you can (do) show that 16k offers no performance gain over 32k, I will stick with 32k.

 

My 64k stripe RAID 0 array is noticably faster than the single Seagate I used for my boot drive before... never going back, haha.

 

I love cold boot to Windows desktop in about 20 seconds.

[supershanks]

If u check post My post Post#1484shows comparisaon of going from original 128 stripe to 64. Performance increase in negligable whilst cpu usage shows almost x2, can't see much point in going to 32 on those results.

 

 

I love cold boot to Windows desktop in about 20 seconds.

Funny u should mention that:-

PCMark05

 

luck:D

[Polizei]

What am I looking for?

[supershanks]

sorry:- post# 1484

Edit: -Results

not worth the change

HDTune

64

Pics =

1) HDTUNE 64 stripe

2)HDTune 128 Stripe

 

Bottom pic = 128v64 stripe comparison

HDTACH Compararison Blue=-128 Red = 64