RAID Fellows

[bytemyfsb]

Does anyone out there have an ATA133 RAID setup?

I just bought 2 Maxtor D740x ATA 133 drives to setup up on my Shuttle AK35GTR and I was just wondering if anybody out there has a similar setup. If so you how do you like it? Does it seem faster than ATA100 IDE. I just set this up and I am installing XP right now. Before I had one IBM 40 gig (going bad) so I hope to see an improvement.

Give me your thoughts and comments

Thanx

 

[rbV5]

I have those drives in RAID 0 on my FIC board. I'm using a fasttrak 100 though, so its only ATA100. There is a huge difference in transfer rate on my rig, but my FIC only supports ATA66 onboard so my ide transfer rate is a bit slow anyway, and this FIC is stable but won't win any IDE speed awards.

I ordered my XP333R tonight, so I'll let you know in a week how they work on the HP controller.

 

[ChrisIsBored]

Heh... ATA66, ATA100, ATA133... you guys seem to think there's a performance difference between the three. You're WRONG! The bus deals with HDD size limitations, not speed. The only difference in speed you'll see if from ATA33 to ATA66 and there's not a whole lot there either....

 

[rbV5]

<< you guys seem to think there's a performance difference between the three. You're WRONG! >>



Bah, Apparently, you seem to know what I'm thinking, and you're WRONG. I am fully aware of the concepts of transfer rates, and how they relate to real performance. If you don't think 2) ATA133 drives in a RAID 0 array are faster than a single ATA100 drive, you are WRONG there too (isn't that what he's asking?). 2) ATA100 drives are faster in RAID 0 than a single ATA100 drive, do you think ATA133 is slower?

Asking how a couple new drives perform in a RAID setup is a completely valid question. I have those particular drives currently in a RAID 0 array on an ATA100 PCI controller, and will soon have them in a RAID 0 array on onboard ATA133 controllers.

While there may not be much real world performance difference between ATA66/100/133, there most definately a difference in real performance in their IMPLEMENTATION. For instance, My FIC board has poor IDE performance in Win2k/XP period, no combination of busmaster drivers will give me decent real world performance. This is verified with Sandra, HD Tach and video capture performance. My old MSI, had much better IDE performance, not great, but much better. With my Fasttrak controller, performance is much better, even with a single drive. Why? The FIC supports ATA100 on a PCI card, but that shouldn't make much difference, right? Its the implementation that makes the difference. My drives in the RAID array perform well enough that I have decent video capture performance now.

Now when I move them over to my new xp333r, I can add 2 more 40GB D740x to the RAID controller and have a single capture partition over 128MB, and great IDE performance.

 

[wars]

<< ATA133 drives in a RAID 0 array are faster than a single ATA100 drive, you are WRONG there too (isn't that what he's asking?). >>



No, i don't think that's what he's asking. I believe his question (and i might be wrong as well ) is: will ata133 raid 0 array be faster than ata100 raid 0 array, and in this case, not really. A disk can't (individually) outperform itself based in a RAID setup than it can by itself. Since it's pretty well understood that there isn't a gain in going from ata100 to ata133 (much less ata66 to ata100) there shouldn't be much or any gain from going ata100 RAID 0 to ata133 RAID 0. Of course if you compare a single drive vs. a RAID setup there'd be a noticeable performance gain.

wars

 

[ChrisIsBored]

<< and you're WRONG. I am fully aware of the concepts of transfer rates, and how they relate to real performance. If you don't think 2) ATA133 drives in a RAID 0 array are faster than a single ATA100 drive, you are WRONG there too (isn't that what he's asking?). 2) ATA100 drives are faster in RAID 0 than a single ATA100 drive, do you think ATA133 is slower? >>




Hey moron, before you start trying to flame me and act like you actually have a brain cell... the only difference between ATA100 & ATA133 is the allowed hdd size difference. If you'll notice there's no such thing as an 180G ATA100 IDE drive because the ATA100 bus can't handle drives of that size. Thus the implementation of ATA133 is born. So you're WRONG, next time do your reading before you start with the holier than thou attitude.

I suggested sticking with the single WD drive because of its 8mb cache, and it's a superior drive to the maxtors. Even though you might see a few added advantages to going RAID 0, you won't see a lot, perhaps if this system were being used for multimedia purposes(we were never told what the system is for) or a server, there might be an actual performance difference... but if you think spending extra cash on a desktop for RAID 0 is going to be useful... well... you're wrong again. Likewise, if this system were to be used to one of the aformentioned reasons, then a nice SCSI setup would do you MUCH better.

One last thing, you might want to try moving away from FIC if you're going to start wanting performance. Your IDE bus speeds may be slow because 1, VIA is crap, and 2, FIC is crap. You want to brag about setups? Maybe you'd like to play with my dual 1Ghz P3 with 7 SCSI drives in a RAID 5 array. You want to brag about storage space, go search my thread labeled "Holy crap, ever formatted 300Gigs?"

Good day....

 

[Leokor]

<< Heh... ATA66, ATA100, ATA133... you guys seem to think there's a performance difference between the three. You're WRONG! The bus deals with HDD size limitations, not speed. The only difference in speed you'll see if from ATA33 to ATA66 and there's not a whole lot there either.... >>


Actually, this is not quite so simple. I agree with you that, in some situations, the protocol's maximum transfer speed is more like a speed limit on a highway that most of the cars can't even break. But in some situations, it does matter. Namely, the burst transfer rates from the cache are affected.

Now, I used to think that it effects only a minor contribution to the overall performance. But the experience with the WD "Special Edition" drives that have 8 MB buffer (as opposed to the typical 2 MB) proved me wrong. The SE drives perform better than most 10K SCSI drives in desktop applications and are trailing even such performers as X15-36LP, so the cache does matter there. In server applications, of course, they are not much different than other IDE drives, and are outclassed by their SCSI cousins.

And if the cache size does make that much of a difference, then it means the burst rates do too. And so does the protocol.

Leo

 

[rbV5]

<< Hey moron, before you start trying to flame me and act like you actually have a brain cell... the only difference between ATA100 & ATA133 is the allowed hdd size difference. >>



Nice one, brilliant personal attack, now Einstein, perhaps you show me where I stated that ATA133 is faster than ATA100. Nowhere pal, infact, I have stated many times on this forum that ATA133 was developed to to use 48-bit addressing that will enable HD larger than 128GB.



<< but if you think spending extra cash on a desktop for RAID 0 is going to be useful... well... you're wrong again >>



Well, thats your opinion. For my purpose, as I explained in my post. The extra performance RAID 0 gives me for video capture, is well worth the money. I don't drop frames now even at the highest Resolutions, even on the inside tracks of the Platters. These Maxtors are fast and cheap, I got my RAID card used which I will sell when I get my IWILL MB, the RAID costs, what $9 extra? Why pay more for SCSI when my cheap IDE does what I want?




<< One last thing, you might want to try moving away from FIC if you're going to start wanting performance. Your IDE bus speeds may be slow because 1, VIA is crap, and 2, FIC is crap. You want to brag about setups? Maybe you'd like to play with my dual 1Ghz P3 with 7 SCSI drives in a RAID 5 array. You want to brag about storage space, go search my thread labeled "Holy crap, ever formatted 300Gigs?" >>



Are you even reading these posts? I clearly state that "I ordered my XP333R tonight, so I'll let you know in a week how they work on the HP controller." You know, in my first post where I'm answering the poster's question with MY personal experience with the same drives in my RAID setup.



<< You want to brag about setups? >>



Actually, I just wanted to post a response to the original poster's question. I could give a Rat's A$$ what kind of a rig you use.

 

[Leokor]

<< One last thing, you might want to try moving away from FIC if you're going to start wanting performance. Your IDE bus speeds may be slow because 1, VIA is crap, and 2, FIC is crap. You want to brag about setups? Maybe you'd like to play with my dual 1Ghz P3 with 7 SCSI drives in a RAID 5 array. You want to brag about storage space, go search my thread labeled "Holy crap, ever formatted 300Gigs?" >>


Actually, RAID 5 is fairly slow at writes. RAID 1+0 always beats RAID 5 in all aspects but the cost/storage ratio. And let's stop that kind of bragging. It leads nowhere.

Leo

 

[ChrisIsBored]

bah... bytemyfsb still hasn't mentioned exactly what his setup is going to be used for... so i'm through with this thread. Debate it out all you want. For an average home setup I still say the single WD drive will be more than plentiful.