• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

RAID

M

Mister T

Diamond Member
So I got some money to spend.

I was thinking a couple 75gxp's 30gb.
Do I need a RAID card? I heard of something called
software RAID - what is that? Also do I use Raid 0, 1, 2? Striped? I don't know what anything means. My needs is that I want everything to be really FAST. Redundancy is really a non issue.
Would the optimal RAID setup with 2 identical
drives be faster than just one of the drives solo?

I know its alot of questions, but please educate me.
 

If you want speed and only speed i say get a RAID IDE card and RAID 0 your two IBM drives. You will get a speed increase if you can put more drives on that RAID controller.

Or, you can use NT or Win2k and software RAID your drives. Hardware RAID is a lot more faster.

RAID 0 is a striped set of harddrives.
RAID 1 is a mirrored set of harddrives.
RAID 0/1 is a striped mirrored set of harddrives
 
Ok. Lets say I get a Raid Card and run in Mode 0 under W2k.

The more drives I have the better?

So would I be better on getting 2 30gb or 4 15's?

Any other issues?
 
Quick question to anyone who knows, if you have 4 IDE drives in a raid 0 configuration, does that fully tie up the two IDE channels on most motherboards? It sounds like a speed demon, but if you couldnt add a plextor or dvd drive, it wouldnt be worth it.
 
Those four drives would run off the Promise controller, and not the controller on your mobo.
 
I had a feeling that was the case, but wanted to be sure. How many IDE devices can the Promise controller handle?
 
wow, I just red the gamepc review.

A Cheetach @ 15,000 rpm got smacked up by (2) IBM 75gxp ATA100 drives
in Raid 0.

Sounds like a killer setup.

Ayone have a ata100 RAID controller card?
Where do I get one?
 
Wow, that GamePC review was horrid. How many spelling and grammar errors can you make in a 2 page article? This has no bearing on the validity of the review, but a little proofreading goes a long way in making an article look like an 8 year old didn't write it.

Any storage review that uses Sandra as one of its benchmarks should automatically be considered useless. I don't see how any review can be taken seriously that contains the statement: "So to conclude, when speed is an issue, the X15 barely passes." Excuse me? The fastest drive on the planet barely passes? Did this guy even use the drive?

"Our reasoning for the RAID wining is that with RAID 0 the actual throughput of the IBM drives ran into the 200mb/s limit where as the X15's was running into the 160mb/s limit."

Does this guy have any idea what he is talking about? "Winning" has 2 n's in it. One GXP75 is capable of 300mb/s. The theoretical limit of the IDE RAID setup is 200MB/s, not mb/s. Ok, what about running into the limit. The IBM RAID setup topped out at 71MB/s. That's not even close to half the theoretical 200MB/s limit. The interfaces had absolutely no bearing on the performance of either setup.

"So in Sandra, the RAID is the clear winner being about ten thousand points faster in both FAT32 and NTFS tests."

Points? This guy doesn't even know what Sandra is testing.

Now let's look at how this guy rates the setups in relation to each other of the final 2 paragraphs.

"the X15 didn't fall too far behind" ok, RAID is a bit faster.
"So in Sandra, the RAID is the clear winner" ok, RAID was a lot better
"RAID setup was spiking like crazy on the same test but still did obtain higher transfer rates" ok, RAID had higher transfer rates, next sentence?

"Though hands down, the X15 blew the RAID 0 setup away" huh? You just said it was slower than the RAID setup. Next sentence.

"the X15's access times helped it in keeping up in our tests." Wait a minute, it blew the RAID setup away, now it is trying to keep up? Which is it?

"So to conclude, when speed is an issue, the X15 barely passes." I don't even need to explain how this is a joke.

I wouldn't base hardware purchases on any of the reviews on this site.
 
I'm looking at a Promise Fasttrak ATA100 PCI myself as a RAID controller. That is, unless an AMD 760MP mobo has a controller onboard. 🙂

-SUO, almost excited at the idea
 
Seems like there's alot of discussion about RAID setup suddenly. Yeah article is a joke Pariah. Too much contradiction to be taken seriously. Anyway I love my RAID setup.
 
BossDog: The Promise FastTrak66 can use up to 4 hard drives. And ONLY hard drives, your motherboard controller has to do any other devices you want.

Mister T:

For performance 4x15GB 7200RPM RAID0 would be the best.

Adding drives to a RAID0 isn't linear performance increase though (that is to say 2 drives aren't twice as fast as one, 3 drives aren't 3 times as fast as 1 and so on) because all the drives have to seek, you can't improve seek times with a RAID. 2x RAID0 is probably the best for price to performance. Wait wtf am I talking about 🙂 This is RAID, no price/performance pansies here hehehe.

Yes 4 drives will be better than 2. Watch out though, I read somewhere that the max RAID Array size in Win98 (for the Promise card) is 64GB (not that this matters with 4x15 drives, just a general warnning).

There are lots of exotic RAID schemes, by far the most common are RAID1 and RAID0 though, RAID1 duplicates data and increases read performance (you have two copies to read from), but tends to hinder write performance (you have to write to both drives).

RAID0 is what you want for performance, this reads/writes alternating chunks to each drive and boosts performance quite nicely*. Of course if one drive in the Array crashes, everything on the whole Array is lost.

Take it easy.

*: (Fine details if you want them) The only thing that is actually improved is the data transfer speed, so for large files RAID0 will show huge improvements, however the seek speed does not improve on a RAID because all the drives have to seek together (the same length of time as single drive seek). So if you are dealing with lots of files strewn all over the drive a RAID0 array won't perform much better than a single drive.
However, since most disk activity while running programs is virtual memory, if you set up your swap file so it's continuous on the disk (well the array, but it looks like a single disk once set up) this will minimize the amount of seeking that needs to be done, so you should get a respectable performance boost out of a RAID0 set up.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top