SATA Optical Drive/RAID Question

[TheGeek]

I just built a new Core i7 Extreme system with an ASUS P6T Deluxe MB, 6GB RAM, Nvidia GTX280 and 2x750GB 7200RPM hard drives. I'm having a problem with my drives in RAID. I have them set up as a RAID0 array. On the same controller I have my SATA optical drive. Vista grants every part of my computer a 5.9 in the Windows Experience Index thing, but my Hard drive as only a 3.0. I agree with that. My hard drive does seem slower. Its slower than it is in single drive, non-RAID mode. Why is that? Is it because my optical drive is on the same controller as the raid array so it gets confused or has compatibility issues? Right now I'm back to running as 2 separate hard drives because it's significantly faster. Can anyone help?

Thanks.

 

[mpilchfamily]

Are you sure its in a Raid 0? If the raid is slower then seperate drives then something is either wrong with the drives or the controler. The optical drive being on the same controler shouldn't effect the Raid performance.

 

[Slugbait]

Originally posted by: TheGeek
Right now I'm back to running as 2 separate hard drives because it's significantly faster.

Increase your system performance further by moving your pagefile to the second hard drive, and make it permanent at 12 gigs. Even if you can get RAID to be performant, the difference will be slight.

 

[betasub]

A proper HDD benchmark app might be more reliable and informative than "the Windows Experience Index thing".

 

[Crusty]

Originally posted by: Slugbait

Originally posted by: TheGeek
Right now I'm back to running as 2 separate hard drives because it's significantly faster.

Increase your system performance further by moving your pagefile to the second hard drive, and make it permanent at 12 gigs. Even if you can get RAID to be performant, the difference will be slight.

Same with moving the pagefile and setting a static size, it's not enough to notice IMO.

 

[Slugbait]

Originally posted by: Crusty
Same with moving the pagefile and setting a static size, it's not enough to notice IMO.

You misunderstood.

I meant that the performance difference between moving the pagefile or going with RAID on a two-drive machine is slight.

However, the performance difference between moving the pagefile and not moving it is considerable. Depending on the hardware, you'll see between a 5 and 10 percent increase in system perf by moving the pagefile and making it static. The reason is because the heads on the primary drive does not do double-duty: it's no longer reading/writing to the pagefile while simultaneously reading/writing system files, so it doesn't stop doing one thing in order to accomplish something else. Your proc is just burning cycles waiting for the only true bottleneck in your system to complete tasks. It's like going to Home Depot with your wife...is it faster to shop together the entire trip, or have her get the light bulbs while you get the bag of planter soil?

With RAID, both drives will be responsible reading/writing both. If your machine is primarily for gaming, RAID is not going to impress anybody...but for many other things, RAID will give a nice perf boost. But it's about the same boost as a relocated pagefile...

 

[RebateMonger]

Originally posted by: TheGeek
Vista grants every part of my computer a 5.9 in the Windows Experience Index thing, but my Hard drive as only a 3.0.

Yeah, I get 5.4 out of my single 200 GB Seagate 7200.7 IDE drive. Something sounds seriously wrong. Mine's on the same IDE channel as my Samsung IDE DVD/RW drive on my Intel-chipset ICH8/R drive controller.

 

[Crusty]

Originally posted by: Slugbait

Originally posted by: Crusty
Same with moving the pagefile and setting a static size, it's not enough to notice IMO.

You misunderstood.

I meant that the performance difference between moving the pagefile or going with RAID on a two-drive machine is slight.

However, the performance difference between moving the pagefile and not moving it is considerable. Depending on the hardware, you'll see between a 5 and 10 percent increase in system perf by moving the pagefile and making it static. The reason is because the heads on the primary drive does not do double-duty: it's no longer reading/writing to the pagefile while simultaneously reading/writing system files, so it doesn't stop doing one thing in order to accomplish something else. Your proc is just burning cycles waiting for the only true bottleneck in your system to complete tasks. It's like going to Home Depot with your wife...is it faster to shop together the entire trip, or have her get the light bulbs while you get the bag of planter soil?

With RAID, both drives will be responsible reading/writing both. If your machine is primarily for gaming, RAID is not going to impress anybody...but for many other things, RAID will give a nice perf boost. But it's about the same boost as a relocated pagefile...

I'd like to believe you but my experiences have not indicated that. Do you have any benchmarks to back up the performance increases?

 

[Slugbait]

Originally posted by: Crusty
I'd like to believe you but my experiences have not indicated that. Do you have any benchmarks to back up the performance increases?

No. Do you read Maximum PC Magazine? The Holiday issue calls out this tweak specifically as one of the best...if it was myth, they probably would have have panned it like other well-known tweaks that don't do squat. Myself, I've been using this tweak for a couple of years now.

I'm not going to get into an argument over this. If it doesn't give you a 5 to 10 percent boost, fine. Every machine and config and specific usage is different.

 

[Crusty]

Originally posted by: Slugbait

Originally posted by: Crusty
I'd like to believe you but my experiences have not indicated that. Do you have any benchmarks to back up the performance increases?

No. Do you read Maximum PC Magazine? The Holiday issue calls out this tweak specifically as one of the best...if it was myth, they probably would have have panned it like other well-known tweaks that don't do squat. Myself, I've been using this tweak for a couple of years now.

I'm not going to get into an argument over this. If it doesn't give you a 5 to 10 percent boost, fine. Every machine and config and specific usage is different.

Just like the article says

It won?t work for all computers, so the jury?s still out on this one

Seems to me like they aren't sure, especially after they said they noticed the performance increase "particularly on older machines". The OP's rig is far from older.

I basically agreed with your first post saying that difference will be slight, but then you come out quoting these 5-10% numbers(which aren't slight) so which one is it? It's a huge peeve of mine to see people recommending tweaks for computers when they have nothing but anecdotal evidence to support the claim. I don't even see the magazine offering up any benchmark results !

 

[postmortemIA]

I get 5.9 for RAID0 2xRaptor 74GB

 

[RebateMonger]

Originally posted by: Crusty
I basically agreed with your first post saying that difference will be slight, but then you come out quoting these 5-10% numbers(which aren't slight) so which one is it?

I have no idea what's the "truth" in the case of some of these questions. What I will point out is that, many years ago (like 1990), folks interested in such things reported that it took about a 20% CPU speed increase before users could notice any speed improvement. And this was in the days of 16 MHz processors. While this isn't the end-all of information, I'll conjecture that many folks wouldn't be able to detect a 5% or 10% increase in speed, especially since the difference would only be apparent under certain circumstances.

Again, I'm not arguing about it. Just suggesting it might take careful examination (or a benchmark program) for users to be able to "tell the difference" unless told.