is this ok for 1tb wd blacks in raid 0?

[exodusprime1337]

Just got these drives installed and running.. performance feels good so far but just want to see what you guys think. The drives are 2 western digital caviar black 1tb drives in raid 0 using amd 790fx and sb750 onboard raid from m4a79t deluxe mobo.

[angelkiller]

That's CRAZY FAST!!

[Duffman]

Seems pretty nice to me. Much better than my 0+1 setup

[exodusprime1337]

phew thank you guys.. i have 2 7200.11 drives for seagate, the barracuda's and they sucked, have sucked and will forever suck.. happened to be looking for a way to spend some money and picked these us at bestbuy(don't knock it, they're local) and so far everything seems to be pretty good.. just wondering what your thoughts were.

one question i happen to have though is about stripe size?? i got it at 64k.. is that good or should i have gone elsewhere with the size.. i just let the raid too choose to be honest with you. But if there was a better setting please let me know?

[valtopps]

yeah the wd blacks are nice drives if you want them faster short stoke them

[angelkiller]

Originally Posted by valtopps

yeah the wd blacks are nice drives if you want them faster short stoke them

Actually, that's not completely true.

Short stroking drives means that you cut the size of the drive through software. For example you could cut both drives down to 500GB drives, so the OS will think you ahve a 500GB drive.

This will improve benchmark scores. Why? Simple. You're only benchmarking the fastest half of the disk. As you move inwards on the disk, (as it fills up), you lose perfomrance because the inside of the disk is not moving as fast as the outside of the disk. So the innermost part of the disk is the slowest and the outermost part of the disk is the fastest. If you cut the drive to 500GB, you're only using half of the physical disk. So the 'slowest' part will actually be in the middle of the disk, rather than the inside of the disk. And because the middle of the disk is faster than the innermost part of the disk, benchmarks will improve. HOWEVER, this DOES NOT improve performance. Why? Again, simple. The only performance gain you'll see is your average performance will improve. Your maximum transfer rate will not improve. The disk has not physically gotten faster because you used half of its space. But the average transfer speed has gone up becsue you've eliminated the slower half of the disk. And that doesn't speed up the faster half, just the average.

Think of it this way. You want to average 4 numbers, 1, 2, 3 and 4. You're looking for the highest average possible. If you average all four numbers, you get 2.5. Cool. But if you want a higher average, why not cut out the smaller numbers, which bring the average down? So now, only average 3 and 4. You get 3.5 as an average. Awesome! You increased your average by removing the slower numbers. BUT your numbers didn't get bigger. Your highest numbers is still 3 and 4. This is exactly what happns on the hard drive. 4 is the outside of the disk, which is the fastest, and 1 is the inside of the disk which is the slowest. If you short stroke a drive, you take out 1 and 2. Your average goes up, but you still have the same numbers, which means that you still have the same performance levels. Hopefully that makes sense.

In sum, short stroking a drive improves benchmark scores, but does not make the drives perform faster.

[brandonwh64]

that is some fast drives. i have two 500gb blue drives in Raid 0 and im loving it

[Loosenut]

Wow, mine sucks

Mine's on the right...

[angelkiller]

Originally Posted by Loosenut

Wow, mine sucks

Mine's on the right...

Are you using 500GB drives?

[Loosenut]

yes, 2 WD blacks 500Gb

[angelkiller]

That's really interesting.

I remember when I got my WD3200AAKS, it was much slower than the WD6400AAKS, even though the only (obvious) difference is that the 640GB drive had two 320GB platters instead of one. You would think speeds would be similar. I think I found out that WD optimizes the firmware for different drives, even when in the same series. So my 320GB drive was optimized more so for noise than performance, so my access times are higher than the WD640AAKS. My personal opinion of why WD did this is because of it's intended target. Most people that buy a 320GB hard drive are gonna be typical home users, not on the quest to performance. People that buy a 640GB drive were most likely looking for a large, high performance drive. (Remember this was back when 1TB was the biggest) So maybe the same thing is going on here? Maybe the 500GB Black is tuned to be a little slower? Or better yet, maybe the 1TB Black is tuned for the highest performance possible? Hm.

[Loosenut]

Originally Posted by angelkiller

That's really interesting.

I remember when I got my WD3200AAKS, it was much slower than the WD6400AAKS, even though the only (obvious) difference is that the 640GB drive had two 320GB platters instead of one. You would think speeds would be similar. I think I found out that WD optimizes the firmware for different drives, even when in the same series. So my 320GB drive was optimized more so for noise than performance, so my access times are higher than the WD640AAKS. My personal opinion of why WD did this is because of it's intended target. Most people that buy a 320GB hard drive are gonna be typical home users, not on the quest to performance. People that buy a 640GB drive were most likely looking for a large, high performance drive. (Remember this was back when 1TB was the biggest) So maybe the same thing is going on here? Maybe the 500GB Black is tuned to be a little slower? Or better yet, maybe the 1TB Black is tuned for the highest performance possible? Hm.

Interesting indeed. I know my 500's have 2 platters/each, is it the same for the 1Tb or do they have 4? As you seem to be the expert (heck of a lot more than me anyways ), do you think my stripe size has anything to do with performance?

Edit: My bad, I mistook sector size for stripe size. I'm not really sure how big my stripe size is. I forgot

[slyfox2151]

here are both of my Raid 0's for comparison
Running ICH10R normal Raid 0.

2x WD 640 AAKS with 128 stripe (left)
2x WD 1TB Green Drives 128 stripe (right)

the 640's have a higher avarage and lower access time, both with 3000mbps burst :P thx write back cache enabed.

http://www.techpowerup.org/uploaded....0221/speed.jpg

[angelkiller]

Originally Posted by Loosenut

Interesting indeed. I know my 500's have 2 platters/each, is it the same for the 1Tb or do they have 4? As you seem to be the expert (heck of a lot more than me anyways ), do you think my stripe size has anything to do with performance?

Edit: My bad, I mistook sector size for stripe size. I'm not really sure how big my stripe size is. I forgot

Ah, platter count. That also would explain the performance difference. WD is certainly able to fit 500GB on a single platter. (2TB drive) I'm nearly certain that the 1TB Black has 2 500GB platters. If your 500GB drives have 2 250GB platters, that would explain the difference. I don't know if 500GB/platter has made its way into the 500GB drives yet. How do you know that your drives have 2 250GB platters. I've heard that the 500GB Blacks at one point had two 333GB platters. Maybe they've changed it over time? (250GB initially, then 333GB, now 500GB?) You see, WD doesn't give out this kind of info like platter size and number of platters. So you're always guessing as to what you're getting or have. I would be really happy if HDD manufacturers published this info. At the very least, hardware review sites should review as many hard drives as they can and determine platter density and platter count.

[copenhagen69]

very nice ...

better transfer rates than velociraptors ... REVIEW HERE

even though it was not in raid ... still your setup is haulin!

[esberelias]

i'm using 2 OCZ Vertex's (they are SSD's) but they are lighting fast on RAID 0

[brandonwh64]

i have a WD 1tb black drive coming in tomarrow. i will post it by its self for comparison