Raptors in RAID 0? Any benefit? [Archive]

masher

01-16-2006, 09:34 AM

I've heard that these drives were "made to be RAIDed"...
The older raptors don't support TLER, which means putting them in a Raid array is a risky proposition. Besides doubling your chance of a hardware failure, you also can easily lose data from a temporary drive dropout.

The new 150GB raptors have TLER, but still, you're doubling your failure rate. And there aren't a whole lot of desktop scenarios that see real benefit from RAID. You can usually get much better performance just by running them as two independent disks (e.g. OS+swap file+utils on 1 drive, apps on other...)

TheDoucheMan

01-16-2006, 09:44 AM

There is some gains... but this isn't the fourm or site to ask that question for the admins are against raid at all costs.

Sorry if people get fed up about answering the raid 0 question 10 times a day. Most people have realized that the extra risk to the data is not worth the minimal increase in performance. I ran a raid 0 array, and then ran the drives separately I didn't really notice that much difference, and I was doing small file transfers as well as large file transfers. Many people swear by raid 0 because of their seat of the pants benchmarking, I personally think that raid 0 is waaaaay overhyped.

To the OP
If you are worried about losing data, or if they are going to be the only drives in the computer, then don't do raid 0. You need a storage drive so that you don't have to keep ANYTHING important on the raid array, only replacable files.

DougLite

01-16-2006, 10:19 AM

There is some gains... but this isn't the fourm or site to ask that question for the admins are against raid at all costs.The admins aren't against RAID at all, it's one mod (me). Also, I've never used mod power to steer the RAID debate.

Furthermore, I'm not always against RAID-0. As I've pointed out, it's very useful for content creation scratch disks and disk to disk backup. My position all along is that it is neither cost effective nor worth the pitfalls for minimal gains in some desktop access patterns and even performance drops in others.

Ockie

01-16-2006, 11:00 AM

Also, I've never used mod power to steer the RAID debate.

It explains your sticky that you made?

:p

Celeryman

01-16-2006, 11:00 AM

I saw almost no benefit running my 74GB raptors in Raid 0. I only benched them a couple of times. The problem was, the feel of going from one to two did not change. I am talking about load times in games mostly. The one thing I did notice is that MS Office started a lot faster!

DougLite

01-16-2006, 11:02 AM

It explains your sticky that you made?

:pIt wouldn't have mattered...as fast as it was getting posts, it would have stayed on the front page of the DSS index for a week, stuck or not :p

The title I used did more than the sticky :eek:

rodsfree

01-16-2006, 02:35 PM

It wouldn't have mattered...as fast as it was getting posts, it would have stayed on the front page of the DSS index for a week, stuck or not :p

The title I used did more than the sticky :eek:

And here I was thinking that the OP of this thread knew how to get an argument started. :eek: :D

http://www.hardfolding.com/ftag1.php/mem/2172.png (http://www.hardfolding.com?go=38&id=2172)

djnes

01-16-2006, 08:28 PM

There is some gains... but this isn't the fourm or site to ask that question for the admins are against raid at all costs.
Actually, they are pro-factual information. Furthermore, it's RAID0 that's a joke...not RAID in general.

unhappy_mage

01-16-2006, 09:49 PM

<at risk of starting another flamewar, I'm posting in this thread>
It explains your sticky that you made?

:p
DL had 23 posts (unless I counted wrong), I had 18. I made a lot more radical statements about raid.

DL's not the only one who feels that way about Raid 0 - there are some marginal gains in certain cases, but for most situations it makes more sense to use multiple single disks. If you buy another disk, put the games on one disk, and the OS, pagefile, and apps on the other disk.

The ultimate solution for more performance is usually more ram, if you're not maxed out yet. If you don't have 2gb (at least) that's a better upgrade. If you're playing a game that doesn't use a lot of ram (any ut99 fans in the house? :D), you can get a software ramdisk (I bought the one from ramdisk.tk for $6) and load the games to that for really fast load times.

http://www.hardfolding.com/ftag1.php/mem/150072 (http://www.hardfolding.com?go=38&id=150072&tm=33)http://www.hardfolding.com/utag.php/mem/1392.png (http://www.hardfolding.com?go=36&id=1392)

natermeister

01-16-2006, 11:25 PM

I was leaning against RAID 0. At some point I'd love to run 3 150GB Raptors in a RAID 5 array. To bad the controllers are pricey. RAID 5 appears to be much more useful than RAID 0. The only problem is if one drive fails, you're in trouble. I suppose you could always have a hot spare ready, but that adds even more cost.

djnes

01-16-2006, 11:41 PM

In RAID5, if one drive fails, you are still okay. You would shut down the PC, and have WD replace the bad drive. Then you'd stick it in, and let the array rebuild, and you'd be fine. I just did this with a RAID5 array of 4 IDE drives.

Ultra Wide

01-16-2006, 11:49 PM

a Raptor RAID review:

Raptor in RAID configuration (http://www.gamepc.com/labs/view_content.asp?id=raptor150raid&page=1)

mrbigshot

01-17-2006, 12:11 AM

i cant comment on raptors in raid i can comment on raid 0 performance, test was done with a epox 8kda3j mobo, athlon64 3000+ @ 2.4 and 1 gig patriot tccd. sorry its older but its what i was using a year and a half ago.

this was a single 40 gig maxtor ata 133 7200 rpm drive using a ide/ sata adaptor
http://server3.uploadit.org/files/mrbigshot-hdtach.jpg

this was the same drive and another identical added to it, again another adaptor added
http://server3.uploadit.org/files/mrbigshot-nvraid.JPG

pretty god damn clear difference to me. windows boots much faster, games and heavy transfer task suck as moving large files like movies get done much quicker.

it it no more risky to run raid over a single drive, if a drive is gunna die its gunna go weather there is 1 drive or 50. i would like to see everyone here follow samrt storage though. windows, games, no important info on raid. important files on a single storage drive

DougLite

01-17-2006, 01:07 AM

I wouldn't be too confident in those claims actually standing up to objective measurement. Looking at your HDTach results, average access time measurements come in a full 1.1ms higher on the striped array, and that's without the additional impact of rotational latency from accessing both disks simultaneously without spindle sync.

Also, your results indicate that you are using at least one DiamondMax Plus 8, which is limited to a 2MB buffer. An 8MB or 16MB buffer drive is going to blow away the application level performance of a DM+8, regardless of how many of them are striped together.

Yoshiyuki Blade

01-17-2006, 03:17 AM

I had a RAID 0 setup before the new Raptor came out. Once it did I sold them as FAST as possible (put em up for auction about 2 days after the announcement). I wanted to get a decent sale before the new Raptors started showing up on eBay, and I'm glad I made that move... sold them for $270, so I'm basically getting the approx. same capacity + better performance - 2x chance of data loss for about $40 more from ZZF (nice math formula there ;)).

Anyway, I haven't had the chance to experience the difference between the normal configuration vs RAID 0, but I didn't expect much from them. I basically wanted more than just 74 GB to put my stuff in, while putting the rest in a storage drive. I've heard about how the 16 MB cache really started helping the 7200 drives catch up to the Raptors in performance, and now that the new Raptors have that, I'm hoping for a tangible increase. It should be here in a couple days.

djnes

01-17-2006, 09:19 AM

pretty god damn clear difference to me. windows boots much faster, games and heavy transfer task suck as moving large files like movies get done much quicker.

it it no more risky to run raid over a single drive, if a drive is gunna die its gunna go weather there is 1 drive or 50. i would like to see everyone here follow samrt storage though. windows, games, no important info on raid. important files on a single storage drive
It is pretty clear of the difference....read DougLite's post.

As far as the risk....if you can't see how there's more of a risk, you need to go back through elementary school math. In a normal config, if one drive dies, you lose data on one drive. In RAID0 if one drive dies, you lose data on both drives.

As far as your results, you are experiencing the placebo effect, as did most of us when we bought into the hype. Once I split my array, I couldn't tell a difference in performance.

masher

01-17-2006, 09:34 AM

it it no more risky to run raid over a single drive, if a drive is gunna die its gunna go weather there is 1 drive or 50.
Your elementary school math teacher should be hauled before a firing squad and shot for incompetence.

In the meantime, lets examine your statement: If a drive is "gunna go, its gunna go". Very true. But that also implies if a drive is not going to die, then it won't. Now imagine you're at your local Drives-R-Us store, grubbing through a bin of 50 drives. Only one of those is "gunna go", the rest are willing to happily spin for you for the next decade. With me so far?

Now, that one evil-minded drive doesn't have a special sticker on it...its in hiding. Incognito. So if you're buying a single drive, what are your chances of choosing that one? 1 in 50 obviously.

If you're buying two for a RAID 0 array, your chances of getting stuck with the loser drive are therefore 2 in 50. And if you're building the 50-drive array you mentioned in your post...your chances, dear sir, are 50 in 50. 100%.

Get it now?

drizzt81

01-17-2006, 11:33 AM

pretty god damn clear difference to me. windows boots much faster, games and heavy transfer task suck as moving large files like movies get done much quicker.
While I do not doubt your benchmark results, I have a question for you: Where did you do a real-world benchmark? I mean, you surely do not play HDTach all day, do you?

Did you install windblows on both the single and the dual drive config and measure the bootup time? How much of a difference did it make?

rodsfree

01-17-2006, 01:27 PM

Your elementary school math teacher should be hauled before a firing squad and shot for incompetence.

In the meantime, lets examine your statement: If a drive is "gunna go, its gunna go". Very true. But that also implies if a drive is not going to die, then it won't. Now imagine you're at your local Drives-R-Us store, grubbing through a bin of 50 drives. Only one of those is "gunna go", the rest are willing to happily spin for you for the next decade. With me so far?

Now, that one evil-minded drive doesn't have a special sticker on it...its in hiding. Incognito. So if you're buying a single drive, what are your chances of choosing that one? 1 in 50 obviously.

If you're buying two for a RAID 0 array, your chances of getting stuck with the loser drive are therefore 2 in 50. And if you're building the 50-drive array you mentioned in your post...your chances, dear sir, are 50 in 50. 100%.

Get it now?

One of the best analysis's of probability I've ever seen. ;)

BTW mrbigshot Can I interest you in a game of 3 card monte????

http://www.hardfolding.com/ftag1.php/mem/2172.png (http://www.hardfolding.com?go=38&id=2172)

masher

01-17-2006, 01:55 PM

One of the best analysis's of probability I've ever seen. ;)

Thanks. Bear in mind though that I assumed mutually exclusive failures (one and only one drive can fail). The "real" calculations, where each drive has a 1/n failure rate, are slightly different.

Yoshiyuki Blade

01-17-2006, 04:54 PM

Thanks. Bear in mind though that I assumed mutually exclusive failures (one and only one drive can fail). The "real" calculations, where each drive has a 1/n failure rate, are slightly different.

Oh yes, I remember all of that crazy statistics stuff I learned in HS. Losta things to factor in. You can even find the percentage of something failing a certain number of times :eek:. (such as the probability that exactly 17 out of 50 hard drives will fail or something, I forgot :D)

USMC2Hard4U

01-17-2006, 05:19 PM

masher

01-17-2006, 10:02 PM

I would rather have little gain over no gain.
What about all the cases where Raid 0 is *slower* than independent drives, hrm?

DougLite

01-18-2006, 12:12 AM

Like in your all important games?

Ockie

01-18-2006, 07:49 AM

What about all the cases where Raid 0 is *slower* than independent drives, hrm?

All the cases? Seriously, you guys need to grow up a little. He never mentioned what his machine was being used for. Stop generalizing because you have facts for a few tests on a handfull of games as a desktop home machine.

Mashie and Douglite, not everyone will ever agree with you on anything or everything. The world works like this... Douglite should know it better than ever from judging his sig.

masher

01-18-2006, 09:09 AM

Seriously, you guys need to grow up a little.
Ah, the old argument ad hominem gets trotted out. Do you really think this helps build your case?

He never mentioned what his machine was being used for. Stop generalizing...
I have not done so. He didn't mention what his machine was used for, but he stated RAID 0 is categorically faster. A false statement, which you yourself admit. Maybe its faster for his particular usage, maybe its slower-- but without more information, we can't assume either.

Mashie and Douglite, not everyone will ever agree with you on anything or everything.
And you should learn that concurrence does not constitute fact. How many people agree or disagree is irrelevant. What matters are the facts.

djnes

01-18-2006, 09:56 AM

I like my raptors in raid 0. I wouldnt have it any other way. I have never had problems either.

Their may be very little gain, but I dont care. I would rather have little gain over no gain.
It may not be worth spending all the money on 2 raptors just for raid 0. But i was going to buy multiple raptors anyway, so I figured what the heck
This is the kind of closed-minded thinking that results in these threads turning into flame wars. As masher said, facts are facts. As long as a person believes they are getting better performance, they will not listen to anything stating stating otherwise. The placebo effect, once again.

DougLite

01-18-2006, 10:10 AM

All the cases? Seriously, you guys need to grow up a little. He never mentioned what his machine was being used for. Stop generalizing because you have facts for a few tests on a handfull of games as a desktop home machine.USMC acknowledges that the gain may be very little. Masher and I are pointing out that it may be non existant, and that the responsivness of his system may actually be decreased by running RAID-0.

Also, just so you know, I was present at the creation of USMC2Hard4U's rig, when he was planning his uber rig. It was a debate over whether or not he should get a SCSI drive that made my bones on this forum, and I ended up being right there too. There were a lot of people posting "SCSI, SCSI!" but I was here claiming that the lower cost, solid performance, and simplified connectivity of the WD740GD puts it out of SCSI's reach for power user hot rod rigs. Now that WD1500ADFD has raised the bar on performance again, while holding the line on environmental performance and retaining the simplified connectivity, the case for SCSI is weaker than ever. USMC and I also had quite a few PMs go back and forth on this matter.

http://www.hardforum.com/showthread.php?t=871327&highlight=SCSI+WD740GDDon't get me wrong. I love SCSI - I have a 36GB Fujitsu and 2 73GB Seagates in my server. But WD740GD sits at a convergence of performance, reliability (5 year warranty) and ease of integration (cooler, quieter, and easier to setup than SCSI sounterparts) that SCSI will never reach, at any price.USMC has since switched back to Intel (He had an X2 4800+, and that X2 4800+ rig was the eventual result of the planning discussed in the thread that the quote was pulled from), and I hear that his dual core water cooled Extreme Edition rig is a screamer. Unless he's picked up a Presler since then :eek:

The same thing that was happening then is happening now. People are looking at synthetic benchmarks and on paper stats and automatically equating them with superior application level performance, a claim that is simply not true. It may be true sometimes, but not nearly enough to be generalized as "true." Indeed, at least in my judgment, you are more correct to generalize it as "false," especially when you factor in the increased cost and other drawbacks associated with RAID.

mrbigshot

01-18-2006, 11:11 AM

i am also sorry for chiming in with syntetic benchmark numbers. i never said it was a cost effective way of doing things. those benchmarks were done on fresh installs with only bare drivers and a defrag. bootup time did increase by 7 seconds with a single drive. i run a single drive for os and another for swap/ storage to this day.

as for generalizing we can safley assume he will be gaming as he is looking at raptors. raptors dont open firefox any faster than a good wd se drive. why would anyone spend the money on a drive of this type if your not taking advantage of it.

Ockie

01-18-2006, 11:36 AM

USMC acknowledges that the gain may be very little. Masher and I are pointing out that it may be non existant, and that the responsivness of his system may actually be decreased by running RAID-0.

Also, just so you know, I was present at the creation of USMC2Hard4U's rig, when he was planning his uber rig. It was a debate over whether or not he should get a SCSI drive that made my bones on this forum, and I ended up being right there too. There were a lot of people posting "SCSI, SCSI!" but I was here claiming that the lower cost, solid performance, and simplified connectivity of the WD740GD puts it out of SCSI's reach for power user hot rod rigs. Now that WD1500ADFD has raised the bar on performance again, while holding the line on environmental performance and retaining the simplified connectivity, the case for SCSI is weaker than ever. USMC and I also had quite a few PMs go back and forth on this matter.

http://www.hardforum.com/showthread.php?t=871327&highlight=SCSI+WD740GDUSMC has since switched back to Intel (He had an X2 4800+, and that X2 4800+ rig was the eventual result of the planning discussed in the thread that the quote was pulled from), and I hear that his dual core water cooled Extreme Edition rig is a screamer. Unless he's picked up a Presler since then :eek:

The same thing that was happening then is happening now. People are looking at synthetic benchmarks and on paper stats and automatically equating them with superior application level performance, a claim that is simply not true. It may be true sometimes, but not nearly enough to be generalized as "true." Indeed, at least in my judgment, you are more correct to generalize it as "false," especially when you factor in the increased cost and other drawbacks associated with RAID.

Double talk much? LOL

Who is this USMC guy, he doesn't really impress me. I didn't know to be always right you have to prove him right :confused:

Anyways, these r0 threads are getting old, audios.

unhappy_mage

01-18-2006, 06:30 PM

Who is this USMC guy, he doesn't really impress me. I didn't know to be always right you have to prove him right :confused:
He built an uber machine, that's all that's relevant. He's the exception to the "it's better spent elsewhere" rule, because he pretty much spent as much as possible elsewhere ;)

http://www.hardfolding.com/ftag1.php/mem/150072 (http://www.hardfolding.com?go=38&id=150072&tm=33)http://www.hardfolding.com/utag.php/mem/1392.png (http://www.hardfolding.com?go=36&id=1392)

rodsfree

01-19-2006, 10:21 AM

He built an uber machine, that's all that's relevant. He's the exception to the "it's better spent elsewhere" rule, because he pretty much spent as much as possible elsewhere ;)

http://www.hardfolding.com/ftag1.php/mem/150072 (http://www.hardfolding.com?go=38&id=150072&tm=33)http://www.hardfolding.com/utag.php/mem/1392.png (http://www.hardfolding.com?go=36&id=1392)

It's one of the advantages of being on Active Duty. - Which I think he is from some of his posts?

When I was in the Army, Uncle Sugar provided "3 hots(sometimes) and a cot(again, sometimes)" so my pay could be spent where I wanted to. And Unlike most of the guys stationed at Ft. Bragg, I didn't spend a lot of time on Bragg Blvd. and at Rick's Lounge. - The Dragon Club and The Pope AFB EM Club are a different story however. ;)
So I had some nice stuff.
He does too!

You should have seen the stuff I brought home from Korea.

http://www.hardfolding.com/ftag1.php/mem/2172.png (http://www.hardfolding.com?go=38&id=2172)

natermeister

01-20-2006, 03:47 AM

Thanks guys, I wansn't trying to stir up shit.

If I ever run two Raptors in RAID 0, I'd keep an SE16 400GB as a backup drive for all my important files.

spiroh

01-20-2006, 01:33 PM

http://www.hothardware.com/viewarticle.aspx?articleid=776&cid=4

Did you guys see this? 5 second improvement in load times for Quake 4 using raptors in raid 0.

Justintoxicated

01-20-2006, 06:34 PM

Anyone who says raid-0 is not faster needs to get a REAL 4 Channel Raid Controller, not use the integrated crap on their MB's...

The difference is there...but it's not Double the gain due to seek times and multiple drives having to spin up.

but for anyone who has not used a seperate raid controller card with 4 HD's each on their own channel, should not be bashing raid arrays.

If I was going to do Raid again, it would be a Raid - 5 configuration, then if one drive dies, I can get a ne wone pop it in, let it be re-constructed, and lose 0 data, without losing 1/2 my storage capacity.

masher

01-21-2006, 06:27 AM

Anyone who says raid-0 is not faster...
Look, its pretty simple. Raid 0 increases bandwidth but also increases seek latency. For a desktop user, that means some things are faster, some slower. Categorical statements such as "Raid 0 is faster" are wrong. It depends on the situation.

If I was going to do Raid again, it would be a Raid - 5 configuration
Of course, Raid 5 is far slower still, regardless of your controller choice. No one choses Raid 5 looking for performance boosts.

DougLite

01-21-2006, 09:13 AM

Anyone who says raid-0 is not faster needs to get a REAL 4 Channel Raid Controller, not use the integrated crap on their MB's...

The difference is there...but it's not Double the gain due to seek times and multiple drives having to spin up.And the impact of increased rotational latency when running RAID-0 (Which doesn't show up in a seek time test BTW) can dramatically reduce localized seek performance, along with buffer strategy the two dominant factors in desktop application level storage performance. BTW, running RAID-0 also hurts the buffer strategy of the drives, as the makers go to a great deal of trouble optimizing read ahead algorithms - running RAID-0 may put that request on a different physical disk, meaning that the time spent reading ahead may be wasted.but for anyone who has not used a seperate raid controller card with 4 HD's each on their own channel, should not be bashing raid arrays.StorageReview used top flight enteprise grade PCI-X storage adapters, with a minimum of 533MB/sec of bus bandwidth, in addition to point to point connections for each drive to the host, yet they note gains of less less than 10% in their productivity measures and a net decrease in performance/responsiveness in their game measures. Actually, "integrated crap on their MB's" has more throughput available than your RocketRAID 404, as "integrated crap" is not limited to the 133MB/sec of the 32-bit/33MHz PCI bus :rolleyes:If I was going to do Raid again, it would be a Raid - 5 configuration, then if one drive dies, I can get a ne wone pop it in, let it be re-constructed, and lose 0 data, without losing 1/2 my storage capacity.RAID protects uptime, not data. Ask anyone who has had a RAID array corrupted, been hit by a virus, accidentally deleted a file, or had a fire, or had a motherboard fail/PSU blow up. All of these causes - even if the physical disks that make up the RAID array are still mechanically operational - can totally trash the data on a RAID array. Running RAID-0, which lacks fault tolerance, is merely an additional risk.

DougLite

01-21-2006, 09:26 AM

http://www.hothardware.com/viewarticle.aspx?articleid=776&cid=4

Did you guys see this? 5 second improvement in load times for Quake 4 using raptors in raid 0.Quake 4 may very well scale with RAID-0. If it is coded like Doom3, Quake4 when it loads, unRAR's a handful of very large archives, then caches them in memory for more efficient execution. As MikeBlas pointed out in his analysis of access in FarCry, it is much more effective to make small access to stuff in memory than on a hard disk. If Id software actually put some thought into how the disk access in Quake 4 is planned out, then it may very well notice some improvement in throughput from using disk striping. Once again, I never said that running RAID-0 would not be a total loss performance wise. I have maintained all along that running RAID-0 is not cost effective. The RAID-0 array gets the job done 13% faster, for double the cost. Even SLI has a return on investment of at least 50%, as much as 70%. Dual core CPUs and other forms of SMP have a return on investment of 90% or more, if you are running multiple CPU intentive applications simultaneously. Disk striping is by far the least cost effective form of parallelization on the desktop.

From SR's results, FarCry and WoW certainly do not scale well at all when running RAID-0. The number of poorly coded games out there that simply hog all of the physical resources they can are legion. If you use a RAID-0 array looking for increased game loading performance, you are likely to be disappointed unless you have money to burn.

cooter

01-21-2006, 11:18 AM

Ok, I get so sick of this arguement. I wish i could sit down everyone who thinks that RAID0 is going to provide some form of UBER leet performance give them a stopwatch and tell them to Time load times of things they do normally. They will then see that there is a minimal difference. But because I can't do this here are my results from a while ago. These were done on my own systems. I have RAID0 tests with both SCSI and IDE vs single drive. I ran each test 3 times and took the average, I also did some synthetic Benchmarks and they show how bad synthetic HD benchmarks can be.

Keep in mind that this is for what I use my comptuer for which is gaming. You will notice the Synthetic Benchmarks show a huge performance gain where as the real world applications show very little gain.

Also after you read my little rant check out this article which i think is pretty good as well:
http://www.overclockers.com/articles1063/

Keep in Mind I have nothing againsts RAID0, I ran 2 SCSI drives on it for 2 years and 2 IDE drives on it for some time as well. I do not care about losing data cuz my computer is set up to where I can reformat at any time cuz nothing important is saved on my C: drive. I am simply giving results to some tests I did, which was mainly to figure out what strip size to use on my SCSI RAID0 Array before I got my Fujitsu Drive.

SCSI RAID TESTS (Seagate Cheetah 15k RPM 36.7 GB Drives, more info on first pic):

Game & Program Load Times:
http://img28.imageshack.us/img28/162/scsitest2pb.th.jpg (http://img28.imageshack.us/my.php?image=scsitest2pb.jpg)

Synthetic Benchmarks:
HD Tach:
RAID0 16k Strip:
http://img28.imageshack.us/img28/2275/16mbstriphdtach1xx.th.jpg (http://img28.imageshack.us/my.php?image=16mbstriphdtach1xx.jpg)
RAID0 32k Strip:
http://img28.imageshack.us/img28/2470/hdtach32kstrip5vv.th.jpg (http://img28.imageshack.us/my.php?image=hdtach32kstrip5vv.jpg)
Single Drive Seagate Cheetah:
http://img28.imageshack.us/img28/5789/singledriveu3201sw.th.jpg (http://img28.imageshack.us/my.php?image=singledriveu3201sw.jpg)
Single Drive Fujitsu MAS:
http://img28.imageshack.us/img28/7943/hdtachmas1rk.th.jpg (http://img28.imageshack.us/my.php?image=hdtachmas1rk.jpg)

ATTO:
RAID0 16k Strip:
http://img373.imageshack.us/img373/589/scsiattoraid016kstrip3sv.th.jpg (http://img373.imageshack.us/my.php?image=scsiattoraid016kstrip3sv.jpg)
RAID0 32k Strip:
http://img373.imageshack.us/img373/9330/scsiattoraid032kstrip6tt.th.jpg (http://img373.imageshack.us/my.php?image=scsiattoraid032kstrip6tt.jpg)
Single Drive Seagate Cheetah:
http://img482.imageshack.us/img482/9181/scsiattosingleseagate1km.th.jpg (http://img482.imageshack.us/my.php?image=scsiattosingleseagate1km.jpg)
Single Drive Fujitsu MAS:
http://img482.imageshack.us/img482/8783/scsiattosinglefujitsu9pb.th.jpg (http://img482.imageshack.us/my.php?image=scsiattosinglefujitsu9pb.jpg)

cooter

01-21-2006, 11:20 AM

IDE RAID Tests:

Maxtor ATA133 80 GB drives:

Game and Program Load Times:
http://img478.imageshack.us/img478/8292/idetest9ck.th.jpg (http://img478.imageshack.us/my.php?image=idetest9ck.jpg)

ATTO:
RAID0:
http://img26.imageshack.us/img26/4357/ideattoriad05lg.th.jpg (http://img26.imageshack.us/my.php?image=ideattoriad05lg.jpg)

Single Drive:
http://img26.imageshack.us/img26/7987/ideattosingle6bu.th.jpg (http://img26.imageshack.us/my.php?image=ideattosingle6bu.jpg)

HD Tach:
RAID0:
http://img26.imageshack.us/img26/3089/hdtachideraid08pt.th.jpg (http://img26.imageshack.us/my.php?image=hdtachideraid08pt.jpg)

Single Drive:
http://img26.imageshack.us/img26/3789/hdtachidesingle1pg.th.jpg (http://img26.imageshack.us/my.php?image=hdtachidesingle1pg.jpg)

PCMark:
RAID0:
http://img26.imageshack.us/img26/8549/pcmarkraid0ide4we.th.png (http://img26.imageshack.us/my.php?image=pcmarkraid0ide4we.png)

Single Drive:
http://img26.imageshack.us/img26/1893/pcmarksingledriveide8lx.th.png (http://img26.imageshack.us/my.php?image=pcmarksingledriveide8lx.png)