Fastest drive setup for Windows XP Pro

[kenham40]

I am looking for the fastest drive setup for Windows XP Pro (my main rig). I have tons of drives for storage and backup, just need the fastest solution for my system drive. I was looking into SAS but I am not sure if that is the fastest available right now. Currently I am only running 2 36 gig raptors (10k rpm) in raid0.

 

[tuskenraider]

If you have the 8MB cache drives, upgrade to two new 36GB 16MB cache drives as they are much quicker. I have done this myself and recommend it. Striping 3 or 4 will give you higher transfer rates, but not quicker seek times, so that isn't a solution. Otherwise, you're probably looking at a costly SCSI setup.

 

[IanG]

A single 150 raptor will likely be faster than a 2x36 raid setup in almost every way except for a few benchmarks. I think even some of the gigantic drives (Hitatchi 400GB, Seagate 750gb) are almost at the level of raptors thanks to data density.

SCSI 15k drives would be the absolute performace kings, and will work with WindowsXP, but would likely require you to get a new motherboard for a pci-x slot.

 

[kenham40]

A single 150 raptor will likely be faster than a 2x36 raid setup in almost every way except for a few benchmarks. I think even some of the gigantic drives (Hitatchi 400GB, Seagate 750gb) are almost at the level of raptors thanks to data density.

SCSI 15k drives would be the absolute performace kings, and will work with WindowsXP, but would likely require you to get a new motherboard for a pci-x slot.

How about this 74gb raptor? It has 16mb cache.

http://shop2.outpost.com/product/5008955?JSESSIONID=F6TpB4GP3R1KvjJhjNPrRFrTk5qh3ZL8sNyDMJcFNMT1LGHyJ2Rw!1676565143!-497049064&site=sr:SEARCH:MAIN_RSLT_PG

 

[tuskenraider]

A single 150 raptor will likely be faster than a 2x36 raid setup in almost every way except for a few benchmarks.

Not likely. The 150GB Raptor has low to mid 8ms seek times compared to high 7's of the 36GB model and transfer rates in the high 70MB/s(84 theoretical). With the old drives in RAID0 you should be pushing 90-100MB/s and 8.0ms seek times. Only question is how much better the new firmware and electronics algorithms of the IDE section of the 150 is to say whether it's better than his current setup.

Now the new 16MB cache drives are all the same in every way except for the two platters in the 150GB drive vs. the one in the 36 and 74GB model, but the smaller drives still seem to show a touch better seek time. So with the new drives, you just pick the one to meet your size needs.

I think even some of the gigantic drives (Hitatchi 400GB, Seagate 750gb) are almost at the level of raptors thanks to data density.

Again, they're improving, but not that close in performance. The Raptors clearly superior seek times and still superior sustained transfer rates, give it a big edge in real life usage and show up in the typical HD Tach benches, etc. Check out storagereview.com for comparisons.

 

[kenham40]

Not likely. The 150GB Raptor has low to mid 8ms seek times compared to high 7's of the 36GB model and transfer rates in the high 70MB/s(84 theoretical). With the old drives in RAID0 you should be pushing 90-100MB/s and 8.0ms seek times. Only question is how much better the new firmware and electronics algorithms of the IDE section of the 150 is to say whether it's better than his current setup.

Now the new 16MB cache drives are all the same in every way except for the two platters in the 150GB drive vs. the one in the 36 and 74GB model, but the smaller drives still seem to show a touch better seek time. So with the new drives, you just pick the one to meet your size needs.

Again, they're improving, but not that close in performance. The Raptors clearly superior seek times and still superior sustained transfer rates, give it a big edge in real life usage and show up in the typical HD Tach benches, etc. Check out storagereview.com for comparisons.

Thanks for the clarification, I am going to run HDTach and see how I do.

 

[tuskenraider]

Since I have them, here's some benches with my new 36'ers, don't have any old 36 benches unfortunately:

Single:

RAID0:

 

[kenham40]

Any idea why mine looks so weird??

 

[tuskenraider]

What's your stripe size? I usually see that happen with 64K or larger, it's not a problem with the drives so don't freak about that. Anyway, do the Long Bench and it should smooth out. Also, what mobo do you have?

 

[kenham40]

I picked up the e6300 core2duo and ECS p4m800pro-m combo from frys for 159.99. The board has sata raid and the two together were cheaper than an e6300 by itself.

What's your stripe size? I usually see that happen with 64K or larger, it's not a problem with the drives so don't freak about that. Anyway, do the Long Bench and it should smooth out. Also, what mobo do you have?

 

[unhappy_mage]

Not likely. The 150GB Raptor has low to mid 8ms seek times compared to high 7's of the 36GB model and transfer rates in the high 70MB/s(84 theoretical). With the old drives in RAID0 you should be pushing 90-100MB/s and 8.0ms seek times. Only question is how much better the new firmware and electronics algorithms of the IDE section of the 150 is to say whether it's better than his current setup.

Synthetic banchmarks don't directly correspond to real-world performance. The closest thing I've been able to find to real-world benchmarks is StorageReview, but I don't think they've done raid 0 tests. Too bad, it'd be nice to see the head-to-head.

A single 150 will cost less than 2 36s, have more capacity, take less power, and be less complex than two disks. It seems like a simple decision to me.

 

[kenham40]

When I purchased my 2 36gbs though, it's when they first came out, there was no 74 or 150gb raptor I can see your point though. I would rather get 2 150's in raid 0 though

Synthetic banchmarks don't directly correspond to real-world performance. The closest thing I've been able to find to real-world benchmarks is StorageReview, but I don't think they've done raid 0 tests. Too bad, it'd be nice to see the head-to-head.

A single 150 will cost less than 2 36s, have more capacity, take less power, and be less complex than two disks. It seems like a simple decision to me.

 

[EnderW]

Not likely. The 150GB Raptor has low to mid 8ms seek times compared to high 7's of the 36GB model and transfer rates in the high 70MB/s(84 theoretical). With the old drives in RAID0 you should be pushing 90-100MB/s and 8.0ms seek times. Only question is how much better the new firmware and electronics algorithms of the IDE section of the 150 is to say whether it's better than his current setup.

Now the new 16MB cache drives are all the same in every way except for the two platters in the 150GB drive vs. the one in the 36 and 74GB model, but the smaller drives still seem to show a touch better seek time. So with the new drives, you just pick the one to meet your size needs.

Again, they're improving, but not that close in performance. The Raptors clearly superior seek times and still superior sustained transfer rates, give it a big edge in real life usage and show up in the typical HD Tach benches, etc. Check out storagereview.com for comparisons.

I think you should take your own advice.
http://www.storagereview.com/php/be...&numDrives=1&devID_0=309&devID_1=249&devCnt=2

Synthetic banchmarks don't directly correspond to real-world performance. The closest thing I've been able to find to real-world benchmarks is StorageReview, but I don't think they've done raid 0 tests. Too bad, it'd be nice to see the head-to-head.

A single 150 will cost less than 2 36s, have more capacity, take less power, and be less complex than two disks. It seems like a simple decision to me.

Like this?
http://forums.storagereview.net/index.php?s=&showtopic=21621&view=findpost&p=221874

 

[tuskenraider]

Synthetic banchmarks don't directly correspond to real-world performance. The closest thing I've been able to find to real-world benchmarks is StorageReview, but I don't think they've done raid 0 tests. Too bad, it'd be nice to see the head-to-head. A single 150 will cost less than 2 36s, have more capacity, take less power, and be less complex than two disks. It seems like a simple decision to me.

He already has 2x36GB drives that should perform better than a single 150 in the basic benchmarks we typically use. Based on the benchmark he posted, the controller on his mobo seems craptastic at best, so it could very well be a single newer 16MB cache drive may perform better. Since it'll be on the very same controller, I doubt it. Of course that's a separate issue altogether. And if the OP goes with a single drive it doesn't have to be the 150GB model. ALL the new drives perform basically the same, and like I noted, the seek times seem to be a touch better on the 36 and 74 drive based on user reports in other forums I frequent. As far as synthetic vs. real world, read below.

I think you should take your own advice.

I noted storagereiw.com to compare the Barracuda to the Raptor, we already know where the Seagate will end up. But since you found that nice chart and thread, I'll comment further. First, there is no RAID comparison at all of a 150 Raptor against two new 16MB cache drives. So that makes the comparison useless for new Raptor buyers, but is an arguement against my opinion of his current Raptors against the new ones, noted. Second is there an official reveiw yet so methodology can be examined?, a link I missed?, though it looks like methodology is already being questioned in the thread. Third since the OP does have the older Raptors that is certainly representative of his situation and shows my opinion of his current setup possibly being able to outperform a single new Raptor may be wrong. But I don't believe everything that's written. Since I have two new 16MB cache drives(which perform equal to the 150) tested as single and in RAID0, have had a pair of old 36GB drives in single and RAID0 configurations and a single 74GB, I can say his current or any Raptor RAID setup should outperform any new single Raptor in game loading, driver and program installation, and file transfers, etc. So, everyday computer use. I've timed bootup, game loading, etc with a simple stopwatch and witnessed anywhere from 4-10 seconds drop in all of these scenario's. So with that said, I take knowledgeable sources and get information and then if possible, test it myself, since all that matters is how it affects me(the end user) in the real world. I think storagereview's benchmarks are certainly way more representative and thorough than HD Tach, etc and of course reference them, but they aren't the end all be all of testing because what they show doesn't always jive with real world use. Result, my old 2x36GB RAID0 config was faster than my new 16MB cache single drive in timed performance and feel and is why my original post is what it is.

If you have the 8MB cache drives, upgrade to two new 36GB 16MB cache drives as they are much quicker. I have done this myself and recommend it. Striping 3 or 4 will give you higher transfer rates, but not quicker seek times, so that isn't a solution. Otherwise, you're probably looking at a costly SCSI setup.

To the OP, I'd look into performance expectations with the RAID controller on your mobo(what is it?), since your about 20MB/s off compared to what mine were on a VIA VT8237 southbridge controller. The seek times aren't gonna be affected, so they are what they are, not the best, but good.

I would rather get 2 150's in raid 0 though

If you need the space, sure. I guess I'll constantly repeat that the 36 or 74GB drives perform the same.

 

[TeeJayHoward]

Say, "screw the budget". Pick up a couple iRAMs and drop your OS on a RAID of those! Make it go zoom!

 

[tuskenraider]

I picked up the e6300 core2duo and ECS p4m800pro-m combo from frys for 159.99.

I see this has basically the same southbridge chip as my AMD setup. Are you overclocking? I have found out with my mobo that when overclocking past a certain point(262FSB in my case) and using VIA RAID, the disk performance slows down by a good 15%. Another member here on the boards can't even install a single SATA drive on this controller in Windows OC'ed at 245FSB as he would get detection errors. He put everything back to stock and bingo, install went fine. I wonder if this can happen across platforms with the same southbridge. If you are overclocking, try putting everything back to default and run HD Tach again, as you might see a very nice improvement.

 

[EnderW]

He already has 2x36GB drives that should perform better than a single 150 in the basic benchmarks we typically use. Based on the benchmark he posted, the controller on his mobo seems craptastic at best, so it could very well be a single newer 16MB cache drive may perform better. Since it'll be on the very same controller, I doubt it. Of course that's a separate issue altogether. And if the OP goes with a single drive it doesn't have to be the 150GB model. ALL the new drives perform basically the same, and like I noted, the seek times seem to be a touch better on the 36 and 74 drive based on user reports in other forums I frequent. As far as synthetic vs. real world, read below.

I noted storagereiw.com to compare the Barracuda to the Raptor, we already know where the Seagate will end up. But since you found that nice chart and thread, I'll comment further. First, there is no RAID comparison at all of a 150 Raptor against two new 16MB cache drives. So that makes the comparison useless for new Raptor buyers, but is an arguement against my opinion of his current Raptors against the new ones, noted. Second is there an official reveiw yet so methodology can be examined?, a link I missed?, though it looks like methodology is already being questioned in the thread. Third since the OP does have the older Raptors that is certainly representative of his situation and shows my opinion of his current setup possibly being able to outperform a single new Raptor may be wrong. But I don't believe everything that's written. Since I have two new 16MB cache drives(which perform equal to the 150) tested as single and in RAID0, have had a pair of old 36GB drives in single and RAID0 configurations and a single 74GB, I can say his current or any Raptor RAID setup should outperform any new single Raptor in game loading, driver and program installation, and file transfers, etc. So, everyday computer use. I've timed bootup, game loading, etc with a simple stopwatch and witnessed anywhere from 4-10 seconds drop in all of these scenario's. So with that said, I take knowledgeable sources and get information and then if possible, test it myself, since all that matters is how it affects me(the end user) in the real world. I think storagereview's benchmarks are certainly way more representative and thorough than HD Tach, etc and of course reference them, but they aren't the end all be all of testing because what they show doesn't always jive with real world use. Result, my old 2x36GB RAID0 config was faster than my new 16MB cache single drive in timed performance and feel and is why my original post is what it is.

To the OP, I'd look into performance expectations with the RAID controller on your mobo(what is it?), since your about 20MB/s off compared to what mine were on a VIA VT8237 southbridge controller. The seek times aren't gonna be affected, so they are what they are, not the best, but good.

If you need the space, sure. I guess I'll constantly repeat that the 36 or 74GB drives perform the same.

Are you talking about the WD360GD or WD360ADFD?

 

[tuskenraider]

I'm talking about both throughout my post. I hoped I was clear enough to distinquish the two, maybe not. The most important point I wanted to make was that replacing the OP's old Raptors with new ones, will be a nice improvement. I've done it, tested it. Could it be system dependant? Sure, as I think the OP's HD Tach shows poor performance of his RAID setup. Anyway, the new ones(moreso with the 36GB drives) are faster, as single or in RAID0. Getting into side arguements was not my intention, but I challenge anyone to buy two drives and test them as singles or RAID. Time boot-up, program load and installs, etc. and tell me the RAID0 came out slower, won't happen. So I don't agree with the conclusion one would get reading the thread/chart you linked comparing the old 74GB Raptors to a new 150GB model. Even as the old 36GB drives are the lowest performing out of all the Raptors, I believe a pair in RAID0 will still proove speedier against any single new ADFD/16MB cache model. Again based on benchmarks, timed tests and "feel" of my system. At best, going to a single ADFD model, would be a lateral move for the OP and you couldn't convince me otherwise.

 

[kenham40]

I do have this: http://shop2.outpost.com/product/4968871?site=sr:SEARCH:MAIN_RSLT_PG 4 port sata raid card also, it's in my bbs machine. It has 4 SATAII 250gb drives hooked to it. Would that controller card help at all compared to the one built on to this motherboard?

I'm talking about both throughout my post. I hoped I was clear enough to distinquish the two, maybe not. The most important point I wanted to make was that replacing the OP's old Raptors with new ones, will be a nice improvement. I've done it, tested it. Could it be system dependant? Sure, as I think the OP's HD Tach shows poor performance of his RAID setup. Anyway, the new ones(moreso with the 36GB drives) are faster, as single or in RAID0. Getting into side arguements was not my intention, but I challenge anyone to buy two drives and test them as singles or RAID. Time boot-up, program load and installs, etc. and tell me the RAID0 came out slower, won't happen. So I don't agree with the conclusion one would get reading the thread/chart you linked comparing the old 74GB Raptors to a new 150GB model. Even as the old 36GB drives are the lowest performing out of all the Raptors, I believe a pair in RAID0 will still proove speedier against any single new ADFD/16MB cache model. Again based on benchmarks, timed tests and "feel" of my system. At best, going to a single ADFD model, would be a lateral move for the OP and you couldn't convince me otherwise.

 

[tuskenraider]

I don't know if it would be any help since it'll be on a PCI bus controlled by VIA's chipset. You'd have to experiment. I'd still like to know if you are currently overclocking or not?

 

[kenham40]

Ah, forgot to mention, no, I am not overclocking. Everything is stock.

I don't know if it would be any help since it'll be on a PCI bus controlled by VIA's chipset. You'd have to experiment. I'd still like to know if you are currently overclocking or not?

 

[kenham40]

Ah, forgot to mention, no, I am not overclocking. Everything is stock. Here is a pic of the via raid tool. Shows my settings.

I don't know if it would be any help since it'll be on a PCI bus controlled by VIA's chipset. You'd have to experiment. I'd still like to know if you are currently overclocking or not?

 

[tuskenraider]

Bummer, was hoping there was a chance for another cause for the low benches. Other than that, it just may very well be RAID performance with that mobo isn't that good. Chaning to a 32K or 16K stripe might help your benchmarks a little, but nothing earthshattering. Only thing left that I can recommend is try your RAID card and eventually moving to an Intel chipset mobo the has the nice ICH8 RAID controller.

 

[EnderW]

So I don't agree with the conclusion one would get reading the thread/chart you linked comparing the old 74GB Raptors to a new 150GB model. Even as the old 36GB drives are the lowest performing out of all the Raptors, I believe a pair in RAID0 will still proove speedier against any single new ADFD/16MB cache model. Again based on benchmarks, timed tests and "feel" of my system. At best, going to a single ADFD model, would be a lateral move for the OP and you couldn't convince me otherwise.

You disagree with results from a widely respected expert, yet offer no evidence to back up your ludicrous statement?

 

[tuskenraider]

You disagree with results from a widely respected expert, yet offer no evidence to back up your ludicrous statement?

You have a comprehension problem so I'll leave it at that.

Second is there an official reveiw yet so methodology can be examined?, a link I missed?, though it looks like methodology is already being questioned in the thread. Third since the OP does have the older Raptors that is certainly representative of his situation and shows my opinion of his current setup possibly being able to outperform a single new Raptor may be wrong. But I don't believe everything that's written. Since I have two new 16MB cache drives(which perform equal to the 150) tested as single and in RAID0, have had a pair of old 36GB drives in single and RAID0 configurations and a single 74GB, I can say his current or any Raptor RAID setup should outperform any new single Raptor in game loading, driver and program installation, and file transfers, etc. So, everyday computer use. I've timed bootup, game loading, etc with a simple stopwatch and witnessed anywhere from 4-10 seconds drop in all of these scenario's. So with that said, I take knowledgeable sources and get information and then if possible, test it myself, since all that matters is how it affects me(the end user) in the real world.

 

[EnderW]

You have a comprehension problem so I'll leave it at that.

You have a problem providing evidence.
You claim decreases in game loading times, then let's see the details of your testbed, testing methods, and results.

Something like this.

You can see their results show two drives in RAID0 are slower in Level Loading in 2 out the 4 games compared to a single drive.

In the games where RAID0 is faster, it's around 10% faster.

That's using drives of the same spec for the single drive and RAID0 tests.

So, a 10% increase half the time, slower the other half. That small increase is supposed to overcome the 57% increase in speed from a WD360GD to a WD1500ADFD seen here?

 

[tuskenraider]

You have a problem providing evidence. You claim decreases in game loading times, then let's see the details of your testbed, testing methods, and results.

I offered up a synopsis a couple of posts ago. Testbeds have been past AMD setups with various chipsets I've had over the past three years. I've ran all the drives in my current rig in my sig. Have I recorded all the data(boot, load, install and execution times) to disprove someone who can't provide any of his own data. No. Will I? No, as I don't have anything but a pair of current drives to provide results and my ego won't suffer to disprove you or some of the charts you linked, as I enjoy the benefits of my testing. But just a little something from my memory is my boot times decreased from 44 to 37 seconds and loading America's Army went from 29 to 24 seconds, going from single drive to RAID0. Reboot, test, reboot test, yada, yada. May not be pretty as a chart right now, but the benefit exists.

You can see their results show two drives in RAID0 are slower in Level Loading in 2 out the 4 games compared to a single drive. In the games where RAID0 is faster, it's around 10% faster.

I've had completely different results, all with noticeable gains in RAID0. Did they even try different stripe sizes? Try it yourself sometime.............oh that's right, you don't have any to offer up to help the OP.

So, a 10% increase half the time, slower the other half. That small increase is supposed to overcome the 57% increase in speed from a WD360GD to a WD1500ADFD seen here?

Wow, you have badly interpreted those results to the "real" world. Do you really think you're gonna swap out and old 36GB Raptor for a new ADFD drive and load Windows 57% faster, install a driver 57% faster, start a program 57% faster?

To the OP, try a single drive against your RAID0 when you get a chance and feel free to enlight our dear friend of all the benefits your switch to a single drive begat you. Over and out.

 

[unhappy_mage]

Raid 0 isn't cost effective. Two drives cost twice as much as one. So unless you're getting a better increase in performance out of that money than you would by buying a better video card or another gig of ram or two, it's not the best place to start.

And if you're going to buy raid 0 now and a better video card later, reconsider. Hard drives are evolving, too - by the time the new video card is bought, the raptors will be outdated once more.

 

[tuskenraider]

Raid 0 isn't cost effective.

Of course, paying for that last 10% of performance never is. But since it is basically the slowest subsystem of a PC, it doesn't hurt to get as much performance as possible from it.

 

[tuskenraider]

Two drives cost twice as much as one.

Two 36GB Raptors are about the same as one 150GB. Yes you're getting less space, but more performance, though some will disagree. Which I say, try it yourself.

 

[EnderW]

Two 36GB Raptors are about the same as one 150GB. Yes you're getting less space, but more performance, though some will disagree. Which I say, try it yourself.

The WD360ADFD and WD360GD are two very different drives.

I will agree that 2 x WD360ADFD drives will likely be as fast or faster than a single WD1500ADFD, but there is no way I'm going to believe 2 x WD360GD drives will beat any ADFD without some concrete evidence.

 

[EnderW]

If you compare the graph I posted above with the performance database results for a single WD740GD, you can see exactly how much of an increase RAID0 gives you.

Single WD740GD
SR Office DriveMark 2006 862 IO/sec
SR High-End DriveMark 2006 641 IO/sec
FarCry Performance 833 IO/sec
The Sims 2 Performance 833 IO/sec
World of Warcraft Performance 671 IO/sec

2 x WD740GD RAID0
SR Office DriveMark 2006 920 IO/sec
SR High-End DriveMark 2006 650 IO/sec
FarCry Performance 750 IO/sec
The Sims 2 Performance 850 IO/sec
World of Warcraft Performance 500 IO/sec

Of course, these are estimates since the values for the 2 x WD740GD RAID0 are from the graph, but as you can see there is little gain.
Also, if you look closely, these results correlate very well to the AT article which also showed a small decrease (0.9 seconds) in Level Load time for Sims 2.

 

[Phoenix '97]

You can get some pretty decent deals on used 15K SCSI drives and cards on eBay now. Some of the things I've purchased recently (prices in Canadian dollars, including shipping):
2x Fujitsu MAM3367NC - $35 each - 36GB U160 80 pin
1x Fujitsu MAS3367NC - $63 - 36GB U320 80 pin
1x Adaptec 29160 U160 SCSI card - $35
1x 80 pin to 68 pin drive bay adapter, U160 cable + terminator - $20

For $90, I get performance like this:

 

[EnderW]

You can get some pretty decent deals on used 15K SCSI drives and cards on eBay now. Some of the things I've purchased recently (prices in Canadian dollars, including shipping):
2x Fujitsu MAM3367NC - $35 each - 36GB U160 80 pin
1x Fujitsu MAS3367NC - $63 - 36GB U320 80 pin
1x Adaptec 29160 U160 SCSI card - $35
1x 80 pin to 68 pin drive bay adapter, U160 cable + terminator - $20

For $90, I get performance like this:
http://mercury.made2own.com/~eiiisol/15K.png[img][/QUOTE]
you could have picked up a brand new WD360ADFD that would be faster, quieter, and not need a SCSI card for $110

 

[Phoenix '97]

you could have picked up a brand new WD360ADFD that would be faster, quieter, and not need a SCSI card for $110

It would've been a bit more than $110 for me; the prices are listed in Canadian dollars after shipping. NCIX.com for example has that drive for $140 Cdn. Even considering free shipping, the price after tax would be something like $148.

Having an additional SCSI card really isn't that much of a problem as long as you have a board with available PCI slots and don't mind the 10-15 seconds it takes for the card to do its thing on boot.

Lastly, not even Raptors can approach a 6.0 ms average access time, this is measured with one of my MAMs, the MAS I ordered should be faster still. One of the better programs that could take advanatge of this is MATLAB 7, which a full installation requires some 50,000+ files. That said, a lot of them are small .m files in various toolboxes, it's still a lot to have on an OS drive, even worse, if it's fragmented.

 

[drizzt81]

I am looking for the fastest drive setup for Windows XP Pro (my main rig). I have tons of drives for storage and backup, just need the fastest solution for my system drive. I was looking into SAS but I am not sure if that is the fastest available right now. Currently I am only running 2 36 gig raptors (10k rpm) in raid0.

Well, I am unfortunately not familiar with what enterprise solid-state drives there are available, but I would say that 4 iRAMs in RAID-0 ought to be the fastest consumer solution for a boot disk at the moment, considering that you'll have a hard time finding a mobo with more than 4 PCI slots and 4+ southbridge integrated SATA ports.

 

[tuskenraider]

Here's a useful thread that disproves many of the review articles that claim RAID0 is not beneficial for a desktop user:

http://www.ocforums.com/showthread.php?t=445800

 

[EnderW]

Here's a useful thread that disproves many of the review articles that claim RAID0 is not beneficial for a desktop user:

http://www.ocforums.com/showthread.php?t=445800

A couple of funny things about that article:

1) it contradicts every other RAID0 vs single drive comparison, including those from Eugene at SR
2) On the Photoshop CS2 benchmark, performance is supposedly increased by 164% from a 100% increase in the number of disks - explain that

at least it doesn't claim a FPS or internet speed increase like I've seen some people do before

 

[tuskenraider]

Hey, like everything else, you gotta sort through the info and use some logic when reading. I certainly don't believe the author lied about his results. I don't claim to have seen the same kind of vast improvements the author did, and of course I wouldn't with a different system, but he did a pretty good job and would have been like something I would have done. It is "useful", I think you would agree.

 

[EnderW]

Hey, like everything else, you gotta sort through the info and use some logic when reading. I certainly don't believe the author lied about his results. I don't claim to have seen the same kind of vast improvements the author did, and of course I wouldn't with a different system, but he did a pretty good job and would have been like something I would have done. It is "useful", I think you would agree.

Well when I look at that, logic tells me something obviously went wrong somewhere and it makes me question the validity of his other results, especially since they are contradicted by major review sites like StorageReview and AnandTech.

I've decided to do my own testing - I'm gonna buy 2 WD360ADFD drives and run the same tests and every other benchmark I can think of to compare RAID0 vs a single drive.
I'll start a new thread later asking for input on what benchmarks to run and testing methods.

This probably won't get done any time soon - maybe around Thanksgiving, but I am really eager to put this issue to rest once and for all.