Raid 0

It increases read/write speeds and decreases loading times, however it doesn't increase access times.
 
A few years back I had a pair of Raptors in my machine. Yes, it was noticeably faster in load times and such, but not enough to make me wish to keep it :P (hence the single 160GB now ;) )
 
i am currently using 2 raptors in RAID 0 as well, and I would not recommend it. it is not worth the extra speed. 1 raptor, or even 1 500GB 7200RPM drive is fast enough (i say 500GB bc the data density is so high on those that they are fastor than most other 7200RPM drives)
 
robina_80 said:
does raid 0 increase your performance ie load times, startup times
Click to expand...

Actually, RAID0 is one of the biggest myths going.

The answer is a big, hearty "NO!!!!". It does *nothing* for real life performance, unless you are moving huge files around. Not only does it do bugger all for performance, you cut your MTBF (mean time between failures) in half. When they do fail, you lose *everything*. Been there, done that. As RAID 0 is not a RAID array at all, you have zero redundancy. Which means that when it fails, which it certainly will, there's no way to rebuild it and preserve your data. You're screwed.

And for what? So a pointless, useless benchmark can say "yes, it's faster" when in actuality it isn't? Ask yourself if it's worth it... The answer should be obvious.

Do *not* believe the drive benchmarks you see. I've seen them posted and they do not mirror real world results. In fact, proper reviews totally trash the likes of HDD Tach or whatever because the results they give are false. In other words, they are crap. If you see a review based on those results, you know the reviewer doesn't have a clue what they are doing... In other words, they are an amateur and you don't give credence to their results.


Ok.. so in summary.. Unless you are consistently moving large database files, striping is useless. And if you are a true geek, or a true professional IT, you never, ever, EVER stripe without building in redundancy... Or you're an idiot and you'll get what's coming to you.

lol.

Now, let me tell you how I really feel. :eek:


edit: I will give you an example of where you use striping. As mentioned in another post on here, I have built a couple of Supermicro rack servers for a manufacturing client. The one is SQL Server 2005, hosting a CMS database. They are huge files... The array is a RAID 5 array, with 3 - 137 GB hot swap Seagate 10,000 RPM SCSI drives. RAID 5 does striping, but it also has redundancy built in. Should one drive fail, the server will stay running until I can get to the site to swap out the drive. Then, while the server is running, you pull the drive, throw in another one and it rebuilds on the fly.

I have also done the same thing in my office with a file server. It has six 146GB Fujitsu 10,000 RPM hot swap SCSI drives in two RAID 5 arrays. Because it's used locally and remotely, multiuser, that's when you benefit from the striping.

Anyone who tells you it has any practical application on a home desktop (and I'm sure you'll see it pop up) doesn't know what they are talking about, as they are using synthetic benches to back them up, when in reality the only real world gains they are seeing are between their ears. :)
 
Last edited:
SirKenin said:
Actually, RAID0 is one of the biggest myths going.

The answer is a big, hearty "NO!!!!". It does *nothing* for real life performance, unless you are moving huge files around. Not only does it do bugger all for performance, you cut your MTBF (mean time between failures) in half. When they do fail, you lose *everything*. Been there, done that. As RAID 0 is not a RAID array at all, you have zero redundancy. Which means that when it fails, which it certainly will, there's no way to rebuild it and preserve your data. You're screwed.

And for what? So a pointless, useless benchmark can say "yes, it's faster" when in actuality it isn't? Ask yourself if it's worth it... The answer should be obvious.

Do *not* believe the drive benchmarks you see. I've seen them posted and they do not mirror real world results. In fact, proper reviews totally trash the likes of HDD Tach or whatever because the results they give are false. In other words, they are crap. If you see a review based on those results, you know the reviewer doesn't have a clue what they are doing... In other words, they are an amateur and you don't give credence to their results.


Ok.. so in summary.. Unless you are consistently moving large database files, striping is useless. And if you are a true geek, or a true professional IT, you never, ever, EVER stripe without building in redundancy... Or you're an idiot and you'll get what's coming to you.

lol.

Now, let me tell you how I really feel. :eek:
Click to expand...
RAID 0 does work, it increases the average transfer rate, which drastically cuts down the time it takes to copy say a 1GB video file from one folder to another, or in games it takes slightly less time to load the level because of the faster read/write speeds.

Granted it's over exaggerated, however there is a performance gain.

I average about 100-105MBps when copying files on my RAID 0 array.
 
[-0MEGA-];848907 said:
RAID 0 does work, it increases the average transfer rate, which drastically cuts down the time it takes to copy say a 1GB video file from one folder to another, or in games it takes slightly less time to load the level because of the faster read/write speeds.

Granted it's over exaggerated, however there is a performance gain.

I average about 100-105MBps when copying files on my RAID 0 array.
Click to expand...

No, there isn't... and no, you don't. You might on a synthetic benchmark, but I guarantee you you won't in the real world.

The 1GB video file, however, is the exception to the rule.

On loading games, Windows and the like, you will not see a gain because the files are too small. There are certain games where the level files are huge... And you might see a fraction of a second difference.

However, because RAID 0 setups are prone to failure (and I LMAO when someone tells me they did it with WD Caviars... Because I know *exactly* what's coming, considering those things just LOVE corrupting the MFTs).. What is your data worth to you? When you lose everything and have to start from scratch, was that extra second worth it? Add them up... And then figure out how much money and time you have to spend rebuilding.... If you figure out a way to convince me that it pays for itself, I'll be *very* surprised.
 
Last edited:
SirKenin said:
No, there isn't... and no, you don't. You might on a synthetic benchmark, but I guarantee you you won't in the real world.

On loading games, Windows and the like, you will not see a gain because the files are too small. There are certain games where the level files are huge... And you might see a fraction of a second difference.
Click to expand...
Hmmm... these files do look small:

CoD4_Files.jpg


As for failures, granted it is much more likely, however the chance of drives failing is pretty slim (although the chance increases the more drives you have). Now unless you have extremely important information then theres no reason to have redundant backups in place. A simple backup drive is more then enough.
 
For a home desktop, the best thing you could ever do for yourself is forget that some moron even invented "RAID0"... If *anything* you're looking for RAID0+1. You *have* to have redundancy when you're using striping, or you're asking for trouble... and seeing as how most people on this board insist on advocating Caviars, despite the fact that on a weekly basis every drive I repair is a Caviar that's toasted it's MFT... You have to prepare for the inevitable.

If your data is important to you, and you're not one to backup daily... do NOT use RAID 0. You'll regret it, I promise. Get a file server, a RAID 5 array and a gigabit LAN.. Or worst case scenario, if the money situation prohibits that, a RAID 1 array... And dump all your important stuffs on there.

You'd be much better off investing in a Raptor (and do *not* enable NCQ, as that actually hampers performance in a single user configuration).
 
Like I said, there are *certain* games where the level files are huge... COD is one of them (probably at least partially due to the fact that EA sux0rz at teh game coding thing.. lol).
 
On loading games, Windows and the like, you will not see a gain because the files are too small.
Click to expand...
Most new games do load faster, and anything sequential will be faster as well. Granted, random access has no improvement.

You're right, the most noticable area of improvement is in synthetic benchmarking.

Is it something for an office machine? Of course not. But to say no load times get even a little faster is silly. I can't post up a fancy graph to say the loading time for BF2 or something is faster because as far as I know there is no accurate way to test that.
 
That's the point. Get a stopwatch. The gain is *so* little, if any, that you won't accurately be able to measure the difference. And for what? Are you a masochist? hehe.

It's not something for a home desktop at all. It's not something where data integrity is critical. If you don't mind losing everything for that extra second here and there, yeah, ok... Fine. Personally, though, it's not worth it for me as my time is worth more than that. I'll sacrifice a second here and there for reliability, or I'll spend the money on the RAID 5. For me, failures are not an option. Not for myself, not for my clients.

I'm the kind of person that will spend an extra thousand dollars, just so I don't have to screw with something later. Rogers totally pissed me off not long ago. I went to Telus and dropped almost $2000, just so I could get away from them. I do not have the time for failures, for down times, for broken promises, for mistakes. These things are not an option. I am willing to pay good money to avoid them.

About 3 weeks ago I had a rather disturbing reality check while standing in line at Walmart. It suddenly occured to me, out of the blue, that my time was worth $2 a minute. When that hits you, you'll totally rethink how you do things. It was one of those paradigm shifts that has totally, permanently altered the way I think.

Figure $2 a minute times the number of hours you have to rebuild. That's how much that extra second here and there will cost you. Well, me anyways.

And thus my stance for saying "there's no place for RAID 0 on a home desktop".
 
Last edited:
I currently have a 500GB external drive which I use for my backup. I use SyncToy to synchronize all of the data in the "Users" folder, as well as saved games, which I do at least once a day.

BTW, I'm using Seagate drives. I've had very bad experiences with WD's portable hard drives in the past.
 
It's very good that you do a regular backup. That's awesome. At least you have a "plan b". Most people don't do that (one of the biggest reasons why I'm called in to homes... To save them from disasters because they haven't done backups).
 
And thus my stance for saying "there's no place for RAID 0 on a home desktop".
Click to expand...
I agree completely but some people think they really extra few seconds of play time it will get them and to them it's worth having to rebuild completely more often due to the increased failure rate. Of course, just because the failure rate is increased that's just on average, it may not fail more often or in the entire time you use it.
 
Cromewell said:
I agree completely but some people think they really extra few seconds of play time it will get them and to them it's worth having to rebuild completely more often due to the increased failure rate. Of course, just because the failure rate is increased that's just on average, it may not fail more often or in the entire time you use it.
Click to expand...
And on the reverse side of that, someone may have a single drive and think they are safe, to then have that drive fail on them causing data loss.
 
You must log in or register to reply here.
Back
Top