• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

to RAID or not to RAID

Here's what I'll do:

Install the OS on one drive
Set the paging file to a second drive
Install my applications to the third drive

By putting the paging file on its own drive, virtual RAM access will no long interfere with other disk reads. Putting games and applications on their own drives means that OS and application files aren't competing for disk access, and the application drive will defragment better because it won't have immovable sectors.
Click to expand...
I wonder how the controller(s) affect this. If I use a SATA controller with two drives, one for OS and one for apps, will the controller cause a delay trying to access two drives at once if it doesn't have separate controller for each port? Like my VIA that has only 2 ports. I also have a SI3112, but that is on the PCI bus and performs crappy compared to the VIA. I know with 4 SATA ports on an NF4 board the SATA controller is actaully comprised of two. It has one each for port 1&2 and 3&4. So having a RAID0 setup on that would have a drive ideally on port 1&3. Of course one being PATA and one being SATA would negate a controller conflict.
 
tuskenraider said:
Tested on the same system? From what I've seen and experienced, the old 36GB does about 57-8MB/s in STR and the 74GB does 64-66MB/s, all new ones I've seen are 72-78MB/s. Considering the new .10 drive seems to be averaging about 64-66MB/s, and my .8 Seagate does 54MB/s. This is with either HD Tach or HD Tune. I'd be curious how you compared them and what the actual results were.
Click to expand...

No, not tested on the same system. Mine are PATA Seagates in a RAID0 that I can't break until I get enough storage to move everything. They can't even use the same controllers, so the same system would be pointless.

We tested using timed file transfers off a file server using a large RAID 5. Mine showed to about 10-15% faster on straight reads and writes.
 
tuskenraider said:
I wonder how the controller(s) affect this. If I use a SATA controller with two drives, one for OS and one for apps, will the controller cause a delay trying to access two drives at once if it doesn't have separate controller for each port? Like my VIA that has only 2 ports. I also have a SI3112, but that is on the PCI bus and performs crappy compared to the VIA. I know with 4 SATA ports on an NF4 board the SATA controller is actaully comprised of two. It has one each for port 1&2 and 3&4. So having a RAID0 setup on that would have a drive ideally on port 1&3. Of course one being PATA and one being SATA would negate a controller conflict.
Click to expand...

Of course independent controllers would be ideal, but the controller arrangement is immaterial when contrasting RAID vs distributed tasks. So long as it's apples to apples (same controller/drive configuration) your performance differences should be apparent.
 
what do you guys think of this
http://www.overclockers.com/articles1297/index03.asp

i also have little experience with raid.

i have raid 0 with software raid(yes, windows xp raid! only with some hard drive space)
with 2 120gb ide 2mb cache drives

extracting a 300mbytes iso takes 1m14s in on of those drives out of the raid0
in raid0 it takes 46s

loading battlefield2 maps its the same speed!
but theres speed increases in some situations

and I'm using the worst raid0 you can have.
 
the_killer said:
what do you guys think of this
http://www.overclockers.com/articles1297/index03.asp
Click to expand...
He gets some unbelievably good results. His bootup time decreases by more than 50%, for example.
the_killer said:
extracting a 300mbytes iso takes 1m14s in on of those drives out of the raid0
in raid0 it takes 46s
Click to expand...
How long does it take to extract from one drive to the other one? Raid 0 for extracting archives is usually a waste, because one could reduce the number of seeks required by quite a bit simply by using two separate disks.
 
the_killer said:
what do you guys think of this
http://www.overclockers.com/articles1297/index03.asp
Click to expand...

As previously noted, he got better than a 50% increase in Windows load speed. Something tells me the test loads weren't equal; the single drive was doing more reads than the RAID. It shouldn't be possible to get a 160% speed increase. He had similar results with Photoshop CS.

The other application load times are more realistic. There are decreases in time anywhere from 5-40%, which fall well within expected performance. This depends on drive fragmentation and the application's disk access patterns.

As I've said, the big problem here is that none of these guys are stopping to ask whether there's a better way to use the two drives.

the_killer said:
i have raid 0 with software raid(yes, windows xp raid! only with some hard drive space)
with 2 120gb ide 2mb cache drives
Click to expand...

Your poor CPU :(

the_killer said:
extracting a 300mbytes iso takes 1m14s in on of those drives out of the raid0
in raid0 it takes 46s
Click to expand...

I've got a similar array to yours, but with a RAID controller. I'll try extracting a file a few different ways and see which is the fastest way.

the_killer said:
loading battlefield2 maps its the same speed!
but theres speed increases in some situations
Click to expand...

Game map loads might see a substantial increase in speed, depending on how the game handles them.
 
i don't feel a lot of cpu usage by using software raid, but it should be higher than any other solution of course

(it's a pentium d 805)

"How long does it take to extract from one drive to the other one? Raid 0 for extracting archives is usually a waste, because one could reduce the number of seeks required by quite a bit simply by using two separate disks."

i'll test it and post it here, i also have partitions on both discs without raid so I can benchmark that on the same disks
 
well
decompressing the file from one disk to other took 1m30s
these drives have both of them normal partition (the ones i now tested) and raid0 partitions
in raid 0 it still takes about 43s

so it's really a lot faster!

anyway,
one of the non raid partition is in the beginning of the disk and it's fast and the other partition is in the end of the disk and it's not that fast, so that will also influence the result
 
I've got 2x Raptors 10k RPM hard drives in RAID 0 and it's very fast, everything from installing windows, defragging, loading large content in games like BF2 which require loads of data, it's all significantly faster.

The only main downside is the greater chance of data loss, some people say you're twice as likely to lose the data, in truth if you've bought decent drives it just means the array will die when the one with the shortest lifespan dies, if you treat your disks well they should last ages anyhow.

The problem I've found is doing things like booting windows off a RAID 0 drive, with my old motherboard with an nforec4 chipset and the NVRAID drivers simply didnt work under Vista. Even when you have full driver support like I have for my new Intel mobo, there can be issues where if one drive becomes disconnected because of a fault then the whole array gets knackered, I also dont know of a way to update the system drivers (if they include RAID drivers) when booting from RAID, during the restart to put the new drivers on, windows wont boot and just wont load resulting in having to rebuild the array.

Best advice I can give you is store all your important data on a totaly seperate drive and only run windows and your application/game installs from your RAID array, then it doesn't matter if it all goes wrong. Thats why I only went for the 36Gb Raptors, because 72gb is plenty for a primary drive where all I will use it for is the OS and the app installs. I got a 200Gb IDE backup to complement it which doesn't need to be fast (7,200 RPM)
 
the_killer said:
well
decompressing the file from one disk to other took 1m30s
these drives have both of them normal partition (the ones i now tested) and raid0 partitions
in raid 0 it still takes about 43s
Click to expand...

That's really strange. How compressed is the file you're extracting? How big are the input files compared to the output file? Does it hit 100% CPU usage going from one drive to the other? With raid 0?
 
damn,
i'm sorry
I made a big mistake
it's not a compressed file, it's a iso file
(winrar is associated with it so it fooled me)
so compressed size = uncompressed size
its symantec 10 iso file

even tought,
i go from one iso file to 1785 files and it's faster in software raid0,
i'll try with a compressed file now
 
Again, that's really strange. The two-drive raid0 should force both drives to do a lot of seeking, and the independent disks should read and write in order with no seeking. Can you try copying a file from drive A to drive B compared to duplicating a file on the raid 0 array? I did this test a while back, and found as I expected that copying A->B was many times faster than duplicating the file.
 
You must log in or register to reply here.
Back
Top