to RAID or not to RAID

N

nerdpulse

n00b
Joined
Jan 21, 2007
Messages
14
In a new system built mostly for gaming, would I be better off getting one huge hard drive or two mediums and setting up raid striping? Which would perform better?
 
Are you asking, would i be better off with two 80gb sata2 hdds in raid 0 vs a Raptor?

I dont really care much for raid unless it is being used in a server application.
 
Actually I was thinking of 2 160g sata2 drives (on sale right now for like $40 after rebates at compusa) or if I should get a higher performance single drive
 
Michael Daly said:
RAID0 will reduce your reliability. Better to go with a single drive.
Click to expand...

Ummm, RAID 0 does not reduce reliability. It does increase the risk of losing all that data. Many people on this forum run RAID 0 and have not lost data on their RAID 0 (the one's who know the risk have their data backed up too).
 
I it were me, I would look at if my game is constantly accessing the HD, if you have a lot of HD activity while playing I would think about the raid 0, if it loads and the HD doesnt do much, I play DAOC and usually only when I change zones is the HD working hard. then one drive.
 
Ive never used a 10000 RPM HDD, but I have 2x160 HDDs in RAID 0 and definitely see a difference than just running w/ 1 in non-RAID.

For the price, its a good deal. Just back up your data, then you don't have to worry about losing any data. This applies to non-RAID tho too. I have had no stability or reliabiliy issues ever w/ RAID.
 
Standpoint said:
Ive never used a 10000 RPM HDD, but I have 2x160 HDDs in RAID 0 and definitely see a difference than just running w/ 1 in non-RAID.
Click to expand...

I'm in the same boat. Had 1 250GB SATAII, and when I added a second and striped it, I saw a noticeable increase in performance. Never had a problem with running RAID 0, but I also back up about every two weeks at most so I don't have to worry about it.
 
protias said:
Ummm, RAID 0 does not reduce reliability. It does increase the risk of losing all that data.
Click to expand...

The probability of one of two drives failing is higher than that of one drive failing. If the two drives are storing one collection of data, that means your probability of losing it is higher. RAID0 does reduce reliability.
 
Michael Daly said:
The probability of one of two drives failing is higher than that of one drive failing. If the two drives are storing one collection of data, that means your probability of losing it is higher. RAID0 does reduce reliability.
Click to expand...

Well, no, not really. It increases probability, yes, but that does not necessarily correllate to reduced reliability. Just because there is a higher probability that an array will fail does not mean that it's going to.
 
You guys are arguing over semantics instead of sticking to the points. RAID0 has a much higher chance of data-loss...whatever fancy name you want to call it. The performance gains, especially in gaming or pretty much nill anyway.

My suggestion is to go in the middle of your choices. Get two drives, but don't RAID them at all. Use one for your OS, apps, and games, and then use the second as data storage. You could stick the ghost.exe file on a bootable thumb drive, and make full system backups an image on your D drive. You have performance and data reliability in one package.
 
EvilTwinSkippy said:
Well, no, not really. It increases probability, yes, but that does not necessarily correllate to reduced reliability. Just because there is a higher probability that an array will fail does not mean that it's going to.
Click to expand...

Maybe you should study probability and statistics and then study reliability. You haven't got a clue what you're saying.
 
ugh
raid 0 has a greater chance of overall data loss due soley to there being two drives instead of one
while reliability may not be the correct word(perhaps failure is better), people are still trying to say the same thing

that being said, ive had one drive die on me in 10 years, a 300gig maxtor. shit happens, Backup your documents and pictures often, regardless if running raid

im building 2x320gig 7200.10 raid 0 raid tomorrow :)
 
Michael Daly said:
Maybe you should study probability and statistics and then study reliability. You haven't got a clue what you're saying.
Click to expand...
You're wasting your time. Besides the simple math of it that goes ignored, othes ignore the fact it doesn't really offer anything in terms of real world performance. People have been banned on here for arguing about it, instead of just accepting that it was hype. Just let people run RAID0...since it's their system.

Then you can hop over to the OS forum and see the people asking how to recover data from a corrupted RAID. Many people forget it's not just the drives dying, but the array becoming corrupted as well that can cause problems.
 
People have been banned on here for arguing about it, instead of just accepting that it was hype.
Click to expand...
People who argue are the ones who've actually tested RAID0 on their systems and seen the improvement instead of listening to people like you who only see their opinion/experience as accurate................or reference the same article over and over justifying their opinion as gospel.

To the OP, if you don't have a way to backup a RAID0 array, I'd suggest not using it.
 
tuskenraider said:
People who argue are the ones who've actually tested RAID0 on their systems and seen the improvement instead of listening to people like you who only see their opinion/experience as accurate................or reference the same article over and over justifying their opinion as gospel.
Click to expand...

Funny you should post here...you were one who was reported by several people. Nice. I actually was one who ran the tests when we had about 40 people on here try it out. I did my tests with dual Raptor 36GB drives. With real world testing, my game levels loaded about a second or two faster, but my seek times when up almost 20%. For well over a year, I had the results posted on my own site as my proof, along with posted in a thread where quite a few people did the same. Most of the 40 people did, and we all came to one agreement. RAID0 was an over-hyped techology. The problem is, you can't seem to understand that, and you have to degrade anyone who disagrees with you. Time and time again it's been proven, and yet all you ever do is flame people who disagree, until someone reports you and the thread is locked. Do you ever give proof? Absolutely not...you just go right on disagreeing with several respected, popular websites. Have you efver given proof? No. Have you ever taken my advice and e-mail Anand and Anandtech telling him he's flat wrong? No. Have you ever contacted the owners of StorageReview to tell them they are wrong? No, all you do is troll through the RAID threads and flame people who think differently than you.
 
If I may borrow a line from my buddy GreNME, the burden of proof is on him. All these times, I've referred back to the testing done on here, and given several links. All we've ever gotten in return is flames and trolling. I haven't seen one real world test give any validation to the argument. I know emotion doesn't come across in text, but I'm actually finding this quite humorous. I often wonder if this is how people reacted to the first scientists who said the Earth wasn't flat, and wasn't the center of the universe. I understand it's a paradigm shift, but plenty of time for acceptance has gone by.
 
you know by nowthe debate will never end dj, nothing was ever definitive IMO,too many variables.....ANYHOW....did ya check out that DVD template soft. yet?
 
PaHick said:
you know by nowthe debate will never end dj, nothing was ever definitive IMO,too many variables.....ANYHOW....did ya check out that DVD template soft. yet?
Click to expand...
Yes, I actually did download it, and it might do what I want, thank you. I've never heard of it, and have been using some klunky Fellowes MediaFace crap up until this point.

By the way, I'm all for a debate about the subject. What kills it, is the flaming and trolling, with the closed mindedness. I used to be one extolling the graces of RAID0, without really doing any testing. I took everyone else's words. Then when I started doing some actual testing on my own, I was disappointed to find a wasted my money buying 2 Raptors, when I could have bought one and a larger drive.
 
i used to use it alot,very configureable. takes some testing, but output is very nice with most printers. good luck!
 
djnes said:
Funny you should post here...you were one who was reported by several people. Nice. I actually was one who ran the tests when we had about 40 people on here try it out. I did my tests with dual Raptor 36GB drives. With real world testing, my game levels loaded about a second or two faster, but my seek times when up almost 20%. For well over a year, I had the results posted on my own site as my proof, along with posted in a thread where quite a few people did the same. Most of the 40 people did, and we all came to one agreement. RAID0 was an over-hyped techology. The problem is, you can't seem to understand that, and you have to degrade anyone who disagrees with you. Time and time again it's been proven, and yet all you ever do is flame people who disagree, until someone reports you and the thread is locked. Do you ever give proof? Absolutely not...you just go right on disagreeing with several respected, popular websites. Have you efver given proof? No. Have you ever taken my advice and e-mail Anand and Anandtech telling him he's flat wrong? No. Have you ever contacted the owners of StorageReview to tell them they are wrong? No, all you do is troll through the RAID threads and flame people who think differently than you.
Click to expand...
If I was reported, as I've heard from a couple people that don't like to be disagreed with, there was never any communication to me about it so it must have seen as petty by the admins? Who was degraded here? And if you find my post here inflamatory, then you really have thin skin and it's not my problem. As I've stated in various threads here, I've recorded decreases in boot time, the loading and execution of various programs, etc. As I've stated before, most increases were about 15-20%, some less. 20% of 8 seconds might not be much to you, but it's there and measurable for me and it counts to me. Now as I've stated before, I don't care what Anandtech and Storagereview tell me about my own experience with RAID0 on my own PC's. I could put together a pretty little graph and whatnot explaining my experiences, but why would it be worth my time? For credibility in a handful of people's eyes? Is your opinion gonna change? Do I care? FACT is the benefit is real to me and I certainly don't get compensation for sharing my positive experiences with RAID while you supposedly don't have any, but claimed otherwise before doing your own testing as you just posted above. You stated your opinion in this thread and I respectfully stated mine, you have the problem, not me.
 
tuskenraider said:
As I've stated in various threads here, I've recorded decreases in boot time, the loading and execution of various programs, etc. As I've stated before, most increases were about 15-20%, some less. 20% of 8 seconds might not be much to you, but it's there and measurable for me and it counts to me. Now as I've stated before, I don't care what Anandtech and Storagereview tell me about my own experience with RAID0 on my own PC's. FACT is the benefit is real to me and I certainly don't get compensation for sharing my positive experiences with RAID while you supposedly don't have any, but claimed otherwise before doing your own testing as you just posted above.
Click to expand...

1. Good thing most people DO care more about objective tests run by Anandtech, StorageReview, and TrustedReviews than one person's personal experience.

2. If I hadn't read half a dozen respected and well-designed objective tests on striping but only read the theory behind it and tested it myself, I guarantee I would think my hard drive performance was improved whether or not it was true.

The benefit may be real to you and it may count for you which is good, but placebo pain killers also give real benefits and counts for many people although in fact sugar pills have never been shown to kill pain when you tell the person in advance that it's a sugar pill :) Expected improvements alter perceived improvements in many different areas outside of hard drive performance, which is why we spend all these resources doing objective testing.

3. There are a limited number of things that have been shown to improve slightly, so if these things are the most important things to you and worth the increased risk of data-loss then it's a personal choice and no one should be crucified for this choice. Lots of people use RAID 0 and like it, but to claim specific percentages and detailed values with no objective evidence is another thing especially when there is abundant evidence to the contrary.
 
nerdpulse said:
Actually I was thinking of 2 160g sata2 drives (on sale right now for like $40 after rebates at compusa) or if I should get a higher performance single drive
Click to expand...

If your single and only intention is to get every last ounce out of gaming and storage/backup mean absolutely nothing because you have other systems for that purpose, then a single Raptor is the best choice IMO. Raptor performance is nearly 100% guaranteed vs the open debates on benefits of RAID 0 and guaranteed increased risk of data-loss.

If you want a balance and the couple sec load time difference between a Raptor and a nice 7200RPM SATA-2 drive doesn't bother you, I would also get two 7200 drives and not RAID them as suggested above. A smaller, cheaper OS/Apps drive and a larger storage drive. I also put the pagefile on the storage drive but can't prove that it actually makes real world difference, and the more important files I regularly copy to both drives (documents folder). If cost is a concern, your two SATA-2 160GB drives combined costs far less than even a 74GB Raptor (If you get the rebates)

Yet another option is get to get a used 36GB Raptor OS/Apps drive and add your $40 160GB SATA-2 drive. Can still stay near $100 which is close to your price for the two CompUSA drives including tax.
 
lawrence131 said:
1. Good thing most people DO care more about objective tests run by Anandtech, StorageReview, and TrustedReviews than one person's personal experience.

2. If I hadn't read half a dozen respected and well-designed objective tests on striping but only read the theory behind it and tested it myself, I guarantee I would think my hard drive performance was improved whether or not it was true.

The benefit may be real to you and it may count for you which is good, but placebo pain killers also give real benefits and counts for many people although in fact sugar pills have never been shown to kill pain when you tell the person in advance that it's a sugar pill :) Expected improvements alter perceived improvements in many different areas outside of hard drive performance, which is why we spend all these resources doing objective testing.

3. There are a limited number of things that have been shown to improve slightly, so if these things are the most important things to you and worth the increased risk of data-loss then it's a personal choice and no one should be crucified for this choice. Lots of people use RAID 0 and like it, but to claim specific percentages and detailed values with no objective evidence is another thing especially when there is abundant evidence to the contrary.
Click to expand...

1.:confused: If a "respected" site tells you you're supposed to see one thing and you get something else, how can that be objective since it should be repeatable over and over on every PC with a RAID0 setup?? And I can tell you from experience that isn't the case.

2. A stopwatch is not an analogy to a sugar pill. What you test with it, is either faster or slower. I certainly wasn't trying to manipulate my results in favor of a setup that costs more and has more risk to me. I mean, I just can't even fathom why someone would lie to themselves and I use a measely 15GB of 70GB available of two expensive 36GB Raptors.

3. I've NEVER claimed that anyone else will see the benefit I have, but it is possible considering I've had similar improvements across three different systems and there are pleny others here with similar benefits. As far as evidence is concerned with my specific numbers, I post my general results as a simple reply to those asking about what kind of performance they may see and as a counterpoint to those who state there are no noticeable performance benefits of RAID0 in a typical desktop environment.
 
tuskenraider said:
1.:confused: If a "respected" site tells you you're supposed to see one thing and you get something else, how can that be objective since it should be repeatable over and over on every PC with a RAID0 setup?? And I can tell you from experience that isn't the case.
Click to expand...
One thing I've always wanted you to explain is how or why one opinion means more than another. In my experience, these tests are repeatable, and do back up the many major websites that have also debunked RAID0 as this major performance boosting option. You stated your opinion, and I stated mine, both apparently from our experiences. Why does yours hold more weight than mine? I've shown links to websites and also participated in the massive test done on these boards. You keep repeating that in your own little test bed, it seems to give you all kinds of wonderful improvements. Tell us why your one individual system means more and proves more than any number of others who've done the same test and found that it's just not the case? I'm not saying this in a cocky, smug way. I'm pointing out facts of our past discussions on the subject, and I'm curious to know your answer. You have a very valid point with the first statement I quoted from you.....but what if most people running the tests find their results are in agreement with those "famous websites"?
 
tuskenraider said:
1.:confused: If a "respected" site tells you you're supposed to see one thing and you get something else, how can that be objective since it should be repeatable over and over on every PC with a RAID0 setup?? And I can tell you from experience that isn't the case.
Click to expand...
Non sequitur. Different raid controllers behave differently, and that's to be expected!
tuskenraider said:
I mean, I just can't even fathom why someone would lie to themselves
Click to expand...
Really? Denial isn't just a river in Egypt, you know...
 
but what if most people running the tests find their results are in agreement with those "famous websites"?
Click to expand...
Than I'm sorry it didn't prove to be as beneficial as expected to the person considering it. And you have the inverse of that statement to consider as well. Like I said, eliminating 20% of an 8 second app load will thrill some and be nothing to other. Those who with the loudest voice aren't always correct.
People who argue are the ones who've actually tested RAID0 on their systems and seen the improvement instead of listening to people like you who only see their opinion/experience as accurate................or reference the same article over and over justifying their opinion as gospel.
Click to expand...
This is my post in this thread. Am I really saying my opinion is the only one that counts? Or did I explain that there are differing results and the debate will continue? Is the gospel being preached here again?
Non sequitur. Different raid controllers behave differently, and that's to be expected!
Click to expand...
Yes! So the possibility of different results from all the combinations of systems, cards, chips, whathaveyou, doesn't end in one result.
 
djnes said:
Tell us why your one individual system means more and proves more than any number of others who've done the same test and found that it's just not the case? I'm not saying this in a cocky, smug way. I'm pointing out facts of our past discussions on the subject, and I'm curious to know your answer. You have a very valid point with the first statement I quoted from you.....but what if most people running the tests find their results are in agreement with those "famous websites"?
Click to expand...

I dont wanna get into the middle of this , really i dont. But dj , unless some new exciting news/proof comes out on either side, your "proof" is no more benefitial than his. We all been there with this discussion before. Nothing was definitive, except the fact that this forum is the most anit-RAID0 out there. And those other websites you tout are to be well respected,as are the website/s who routinely used to point out their flaws. They've all moved on in their "fight". My question is why do you continue to try to change others mind? Most times to the point of pissing someone off? Wouldnt it be better to voice a simple opinion, a couple facts, then agree to disagree? I love a good debate, and by posting this reply do not want to get into one...lol..but at the same time, it tires me to keep reading your posts to RAID0 time after time. Sorry if this offends you dj, I really do like reading most of what you post.
 
Mainly to get the right information out there. Quite often, I'll see a thread on the OS forum asking hwo to recover data from a failed array. I'll so a search on the username of the OP for fun, and find a thread in here or GenHardware where the same user is building a system, or setting up an array. The advice given to the OP touts how amazing this performance increase is with RAID0, that it can double your drive's speed, etc. I have even, on a number of occasions, seen people ask about gaming and FPS, and were told on this very subforum that a RAID0 array will greatly improve FPS in a game. These forums are filled with people seeking answers, and some of the misinformation given is downright scary.

I'll give me official response again, for the record. If you are doing some raw video editing, yeah RAID0 may give you a small performance increase. Will your game levels load faster...maybe, by a few seconds or so. Does RAID0 give you anything close to it's touted performance? No, absolutely not. Now, as others have tried to reason with logic and common sense, the negatives far outweigh the small gains that some attain. We all can say, so what, just keep good backups. What percentage of the people actually do that on here. 20%? And this is a computer enthusiast board. How about the general population? 5%...maybe? There just isn't enough of a reason to consider using RAID0 for the vast majority of people. There also certainly isn't enough of a gain to go so far as to bitch, argue, and flame people who say they can't ever recommend RAID0, and that's been my point all along. It's just not worth it.
 
tuskenraider said:
People who argue are the ones who've actually tested RAID0 on their systems and seen the improvement instead of listening to people like you who only see their opinion/experience as accurate....
Click to expand...

And then there are those of us that see computers used in situations other than just gaming and modding and know the real cost of data and downtime.

Disk drives are so cheap these days that if you want both reliability and performance, get a multi-disk setup that provides both. RAID5 or better or just use a big RAID1 (or no RAID) for important stuff that is relatively static and then a couple of small drives in RAID0 for low value stuff and performance. Add a UPS and backup regularly as you suggest
 
Michael Daly said:
And then there are those of us that see computers used in situations other than just gaming and modding and know the real cost of data and downtime.
Click to expand...
No disagreement there. I have a degree from DeVry and happen to have worked(suprise) within computer networking environments so that side isn't lost on me. Since this is a computer "enthusiast" site, the more "outer fringe" technologies of computers will certainly get more attention and won't be considered as "outer fringe".
There also certainly isn't enough of a gain to go so far as to bitch, argue, and flame people who say they can't ever recommend RAID0, and that's been my point all along. It's just not worth it.
Click to expand...
I was the one who was called out here, again, by you like I made some inflametory statement here. What was that? Things may have got heated in a thread or two awhile back, but you apparantly still have the chip on your shoulder. Using the search function can certainly find those few posts as well as the fact that it won't find any grand claims about RAID0 from me, just basic percentages of some of my improvements and my noted positive experience with it. Everyone must be getting bored again with the holidays gone.
 
We all been there with this discussion before. Nothing was definitive, except the fact that this forum is the most anit-RAID0 out there.
Click to expand...
Well if you want to get involved in the "replacement" debate of RAID0 benefit, go over to Overclockers Forum and start telling people why Intel's Matrix RAID0 isn't making their hard drives faster than a standard RAID0. They love talking how their Seagate 7200.10's in Matrix RAID0 are killing Raptors in performance.
 
I'm sure it has been covered but I don't feel like reading everything since I know the answer. I tried different forms of raid 0 using 2 320Gb seagate 7200.10's and the performance increase was non existant for any tasks I do on my 'gaming' computer.

It is kinda harsh to label a forum anti-raid0 as mentioned above. Clearly from my experience it would indicate that people here know more about what they are talking about. [H]ardcore computers used by 95% of members on this forum just don't benefit from it.
 
We went the rounds on this not too long ago. That was a long and involved thread, let's not do that again :)

I don't have the hardware or the time to do the testing right now, but my experience has been that there are better ways to utilize multiple drives on a desktop than RAID 0. RAID 0 only benefits you when you're reading large files that aren't fragmented, and then only with linear reads--that is, start to finish. Games and desktop applications do this rarely.

A better way to get more speed out of your system is to divide up you disk access tasks. My recommendation is to do this across three drives, but even two should work better than RAID 0. I'll be switching to the three-drive arrangement myself once I can afford the Raptor drives I want.

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.

If I only had two drives, I would leave the paging file on the OS drive and put the apps on a separate drive. When gaming, the vast majority of your disk access will be either paging or reading game files.
 
tuskenraider said:
Well if you want to get involved in the "replacement" debate of RAID0 benefit, go over to Overclockers Forum and start telling people why Intel's Matrix RAID0 isn't making their hard drives faster than a standard RAID0. They love talking how their Seagate 7200.10's in Matrix RAID0 are killing Raptors in performance.
Click to expand...

My older Seagates will outpace a buddy's Raptor on raw speed, but there is no way they can compete with the Raptor on seek performance.

The problem is that most people really don't understand hard drive performance and what affects it. I'm no expert on the subject, but it's clear to me that even on these forums there are a lot of people that still just don't get it.

Remember when megahertz was everything in a processor? Remember when the industry abandoned Mhz because it was no long an accurate indicator of performance? I recall a lot of people in the tech world poking fun at the "but it's got more Mhz" crowd for being ignorant. I see a lot of parellels here with RAID 0.

Read speed is important and RAID0 will drastically improve it, but it's not the most important aspect of hard drive performance for your average gaming desktop. Like 3DMark scores and clock cycles, HDTach scores and read speeds are not indications of real world performance.
 
PopeKevinI said:
Install the OS on one drive
Set the paging file to a second drive
Install my applications to the third drive
Click to expand...

I did that on a 386 many years ago. The paging disk was a small, fast SCSI (recycled from an older computer) and the other two were PATA33 (66 was still a future hope). It ran noticibly faster than the single drive version I started out with.

CPUs, GPUs, RAM and a whole lot of stuff is enomously faster than those days, but disks are only a few times faster. Optimizing the use of the drives makes more sense than putting all your eggs in a RAID0 basket.

PS - Why don't they call RAID0 what it is? ADD - array of dependent disks.
 
My older Seagates will outpace a buddy's Raptor on raw speed, but there is no way they can compete with the Raptor on seek performance.
Click to expand...
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.
 
You must log in or register to reply here.
Back
Top