RAID BABY!!!

[Pariah]

Originally posted by: GOREGRINDER
so they used default stripe size and what?,..the default cluster size?,..a 64k stripe size and a default 4k cluster size = standard disk performance =poor raid0 performance

You don't ever want to change the cluster size for NTFS partitions unless you have very specific needs for a different size. An increase in performance should not be a reason, as there won't be one. You should not be changing the default stripe size either. For obvious reasons, RAID controllers are optimized for certain configurations based on what the intended use for the product is and those will be the default settings the card ships with. You're most likely hurting performance by changing the defaults. That said, changing the stripe size will have very little impact on performance unless you go from like 8KB to 512KB. It's best just to leave the card at the settings the manufacturer shipped it with.

 

[Googer]

Originally posted by: Pariah

Originally posted by: GOREGRINDER
so they used default stripe size and what?,..the default cluster size?,..a 64k stripe size and a default 4k cluster size = standard disk performance =poor raid0 performance

You don't ever want to change the cluster size for NTFS partitions unless you have very specific needs for a different size. An increase in performance should not be a reason, as there won't be one. You should not be changing the default stripe size either. For obvious reasons, RAID controllers are optimized for certain configurations based on what the intended use for the product is and those will be the default settings the card ships with. You're most likely hurting performance by changing the defaults. That said, changing the stripe size will have very little impact on performance unless you go from like 8KB to 512KB. It's best just to leave the card at the settings the manufacturer shipped it with.

Most of the time that is 64k. 256k provides the best performance balance.

 

[Pariah]

I don't think I've ever seen a card default to 256KB. They're usually 32KB to 128KB. And the fact that basically every card has its default in that range should tell you that 256 is not the optimal performance setting. General rule of thumb is that the greater the throughput you want, the smaller you should make the stripe size, for an increase in positional performance (minimal, could be none at all depending on how the controller was designed), choose a larger stripe size.

 

[Googer]

Every raid card i have used had a default stripe of 64k. Find the article in Maximum PC from a few years back. I think it was tomshardware that came up with a similar conclusion but I could be wrong on toms though.

 

[Pariah]

Adaptec has cards that default to 128KB.

 

[Googer]

Originally posted by: Pariah
Adaptec has cards that default to 128KB.

I have been usinag LSI. So maybe my statement is too broad.

 

[KoolDrew]

Here is another quote from StorageReview:

The enthusiasm of the power user community combined with the marketing apparatus of firms catering to such crowds has led to an extraordinarily erroneous belief that striping data across two or more drives yields significant performance benefits for the majority of non-server uses. This could not be farther from the truth! Non-server use, even in heavy multitasking situations, generates lower-depth, highly-localized access patterns where read-ahead and write-back strategies dominate. Theory has told those willing to listen that striping does not yield significant performance benefits. Some time ago, a controlled, empirical test backed what theory suggested. Doubts still lingered- irrationally, many believed that results would somehow be different if the array was based off of an SATA or SCSI interface. As shown above, the results are the same. Save your time, money and data- leave RAID for the servers!

-- Source

This also does a good job explaining everything.

 

[HendrixFan]

http://www.tweakers.net/reviews/515/1

The Business Winstone 2004 programs Internet Explorer, Outlook, PowerPoint, Norton Anti-Virus and WinZip generated just a little I/O activity, with an average load of eight percent. A safe conclusion would be that a Business Winstone 2004-benchmark alone is not a good starting point when testing RAID 0 performance. On the contrary: to have some reliable tests, we will need to put heavy loads on the array.

The third part of AnandTech's benchmark suite, measuring loading times in games, can also be questioned. Just like the previous tests, these tests show a slight improvement in performance with the RAID configurations, with highly diverging performance. Our gaming traces show that most games have very little impact on modern hard drives and generate a rather low load. Even the heavyweight champion Battlefield Vietnam was able to burden a Raptor WD360GD for only 17,1 percent average with a short peak of 70 percent. Other games had even lower averages and didn't rise above 80 percent peak usage. Therefore we conclude that loading game levels is mostly cpu intensive and does not rely on storage devices.

Storage Review, unlike AnandTech, does give some information about its traces in the description of their testing methods. It seems that the Office DriveMark 2002 trace had an average queue depth of 1,34 I/O's and an average of 23KB transfer size. The Content Creation Winstone 2001 trace on average contains 1,4 I/O's with a transfer size of 69,5KB. Compared to statistics of our more modern Winstone 2004 traces, both Business Winstone 2004 and Multimedia Content Creation 2004 seem to have higher loads than Storage Review's corresponding tests. BSW2004, with some extra Winamp and eMule activity, gives us a queue with an average of 3,22 I/O's and MCCW2004 even has 8,82 I/O's in its queue. The transfer sizes seem to be in line with those of Storage Review: 25,7KB in the BSW2004 trace and 62,1KB in MCCW2004's trace.

This higher queue-load in Tweakers.net's traces is positive for RAID 0 configurations, which can easily distribute commands to different disks when facing a large queue. Storage Review's tests on the contrary offer almost no possibility for simultaneous transactions, which puts a limit on performance gains with a RAID 0 array, since all improvements should come from higher transfer rates with sequential access. The fact that traces made with Business Winstone 2004 and Multimedia Content Creation 2004 provide quite different results with Storage Review's older traces, indicate clearly that our colleagues should actualize their benchmarks. BSW2004 and MCCW2004 are both realistic simulations of multi-tasking users in some current software-editions. Disk activities generated by the Winstone 2004 suite are certainly no exception for target group.

Most striking is the fact that neither AnandTech nor Storage Review even bothered to test RAID in situations where the performance of the storage system really matters - for example when a backup application, anti-virus tool or compression program is active in the background while demanding content creation applications are running in the foreground. These are conditions that quickly give rise to noticeable latencies and in which single drive setups don't suffice anymore.

AnandTech and Storage Review should be wise to investigate matters more thoroughy before jumping to quick conclusions. You don't judge a Porsche on its capabilities to carry groceries. A car like that serves a different purpose, and it should be judged on that instead - even if it will never fulfill its true purpose in real life. Power users, tweakers and hardware enthusiasts, the target audience AnandTech, Storage Review and Tweakers.net try to please, use their desktop systems in a different way than the pretty blonde next door who only uses it to check her Hotmail account. What we're trying to say is that you shouldn't assess the performance of RAID 0 with benchmarks that are not made to test the performance of the storage subset. AnandTech and Storage Review's negative verdict on RAID 0 in the desktop environment will likely have a profound influence on the opinions of uninformed users for years to come. A sure loss, since their verdict couldn't stand up to trial.

The article has plenty of benchmarks to support their claims. First, they claim the testing done by Anandtech and SR aren't sufficient. Then they show that with their tests show the advantages of RAID, specifically for advanced RAID hardware.

Ive used RAID myself. Ive used onboard RAID, a nice hardware (Adaptec 2400A) ATA RAID, and a nice hardware (LSI Elite 1600) SCSI RAID. I can tell you from experience that the better hardware you use, the better performance you will get out of the RAID. Using onboard RAID for testing is absurd as is denouncing RAID because onboard RAID sucks.

You can clearly see that when using a real RAID controller, and adding 2, 3, 4 or more drives, the performance scales almost linearly. Instead of claiming all RAID is useless because onboard RAID doesnt do much, you should really see what is going on. It would be just as stupid to claim that video cards dont matter because onboard video doesn't do much.

 

[KoolDrew]

Here is a topic at Ars discussing a little about that article.
http://episteme.arstechnica.com/groupee/forums/a/tpc/f/24609792/m/535005954731

 

[HendrixFan]

Thanks for that link. I can tell you from my RAID experience that losing a drive is a concern (RAID 0). I have lost data on a RAID array before (Ive lost data on a regular HD before) and it sucks. I do regular back ups, even of things not on a RAID array.

What I am trying to refute here, isnt that RAID is risky, but rather that RAID offers no performance gains. RAID brings performance and it brings risks. Whether or not one outweighs the other should be a person's concern. Not whether or not it is faster.

BTW- I would never suggest to anyone to run RAID off any software controller.

 

[KoolDrew]

What I am trying to refute here, isnt that RAID is risky, but rather that RAID offers no performance gains.

I and plenty of others are not saying there are no performance gains. What we are saying is that in typical desktop applications there will be no gains. There is a huge difference.

 

[Pariah]

Originally posted by: HendrixFan
Thanks for that link. I can tell you from my RAID experience that losing a drive is a concern (RAID 0). I have lost data on a RAID array before (Ive lost data on a regular HD before) and it sucks. I do regular back ups, even of things not on a RAID array.

What I am trying to refute here, isnt that RAID is risky, but rather that RAID offers no performance gains. RAID brings performance and it brings risks. Whether or not one outweighs the other should be a person's concern. Not whether or not it is faster.

BTW- I would never suggest to anyone to run RAID off any software controller.

If you try hard enough you could probably create a benchmark to show that a 5.25" floppy drive is faster than a 15k RPM SCSI drive. The question is though, is there any chance at all that a typical user would ever come across the necessary set of conditions to duplicate what that benchmark was testing? Highly, highly unlikely.

Same thing goes for RAID. The reviews that show RAID not increasing performance at all are the ones that usually test the setup under application and conditions that user will be doing the vast majority of the time. The Tweakers link you provided above is of little use or relevance, because they set out with an agenda to try and show that RAID is faster. As such, they went and looked for benchmarks that would show performance increases despite the fact that no one would typically use their computers in such situations.

RAID can improve system performance, but only under very specific situations which most people would rarely come across. For home users the negatives of RAID far outweigh the positives.

 

[KoolDrew]

If you try hard enough you could probably create a benchmark to show that a 5.25" floppy drive is faster than a 15k RPM SCSI drive. The question is though, is there any chance at all that a typical user would ever come across the necessary set of conditions to duplicate what that benchmark was testing? Highly, highly unlikely.

Same thing goes for RAID. The reviews that show RAID not increasing performance at all are the ones that usually test the setup under application and conditions that user will be doing the vast majority of the time. The Tweakers link you provided above is of little use or relevance, because they set out with an agenda to try and show that RAID is faster. As such, they went and looked for benchmarks that would show performance increases despite the fact that no one would typically use their computers in such situations.

RAID can improve system performance, but only under very specific situations which most people would rarely come across. For home users the negatives of RAID far outweigh the positives.

:thumbsup:

 

[Varun]

Originally posted by: HendrixFan
http://www.tweakers.net/reviews/515/1

The article has plenty of benchmarks to support their claims. First, they claim the testing done by Anandtech and SR aren't sufficient. Then they show that with their tests show the advantages of RAID, specifically for advanced RAID hardware.

Ive used RAID myself. Ive used onboard RAID, a nice hardware (Adaptec 2400A) ATA RAID, and a nice hardware (LSI Elite 1600) SCSI RAID. I can tell you from experience that the better hardware you use, the better performance you will get out of the RAID. Using onboard RAID for testing is absurd as is denouncing RAID because onboard RAID sucks.

You can clearly see that when using a real RAID controller, and adding 2, 3, 4 or more drives, the performance scales almost linearly. Instead of claiming all RAID is useless because onboard RAID doesnt do much, you should really see what is going on. It would be just as stupid to claim that video cards dont matter because onboard video doesn't do much.

I think your article questioning Anandtech and Storage Review are really missing the point. The point is they are trying to determine if RAID is worth it for typical desktop use. Obviously typical user level applications do not stress the storage system enough to warrant the increased chance of failure. I'm sure that at times RAID0 would make a noticeable improvement, but as the article stated, things like game loading are mostly CPU limited.

There is no doubt that once you step up to servers and the like, RAID0 would have a huge increase in performance. Of course most servers run RAID5 for redundency + performance mostly because of the risks of RAID0.

There are going to be times when someone wants RAID0 and it would be worth it for that person. If they are editing large files, etc, the performance increases would be noticeable and worth the risk in hard drive failure (with backups of course). Most of the time the small increase in performance is not worth the risk though. The original poster does not need RAID0 if he just wants a storage drive. He would be far better off to just leave them as single drives.

 

[Rubycon]

The only way a floppy is going to be faster than a 15k drive is if you pull the string on the clay pigeon thrower holding both a few moments before the 15k one if you know what I mean.

 

[Fullmetal Chocobo]

Damn I hate these conversations. They never seem to take a few things into consideration:
1) Let people fvcking do what they want to do. STFU!
2) How the fvck can someone go and overclock their processor and video card bitch about someone trying to get some extra performance out of their hard drives?
3) Performance isn't the only issue with RAID. I personally like having a 1TB RAID 5 array, as I have 'docs' on a 700gb partition.
4) RAID is definitely a case of you get what you pay for. If you are connecting drives and doing a software RAID, or even using the integrated controller on your mobo, of course your performance isn't going to be stellar. Get a good RAID card, PCI-e or PCI-X, and go from there. The benefits of RAID really start to show when you are using 4+ disks on a good controller.
5) HendrixFan made a good point below me. RAID will not increase your FPS in Solitaire.

SUMMARY: Not everyone needs RAID. Not many people need anything really. But if someone wants to run RAID, give them the information they need (IE if they don't know that one drive failing will kill a RAID 0 array), and let them be.

Okay, just my $.02.

 

[HendrixFan]

It would seem that NONE of you actually read the article I linked to. It shows numbers for things used in desktop systems, especially when multitasking.

Why would you test things that dont stress a hard drive if you are testing RAID? Does anand fail to show the "real world" results of a CPU by not showing us the little improvement we get in "real world" apps like IE, Firefox, Winamp, etc? If you are testing a component, you need to test where it makes a difference. It is then up to a user to decide if those things matter.

In the tweakers article, they show that using newer versions of the programs you can see more benefit. They also show that using a hardware solution provides benefit.

As far as things that no typical user would do, here are things they test and show a near linear improvement as drives are increased (up to 4 drives in RAID 0):

Filecopy
Winzip compression
Virusscan
Defrag
Windows Bootup
Windows User Switch
Software Install
Windows Update
DVD Ripping
Audio Edit
Photoshop load/save

You tell me if users will ever come across these "strange" scenarios on their computer, much less more than one at a time (where RAID really seperates from a single drive).

 

[SnoMunke]

RAID 5 is not the best solution for large files transfers. If you can afford it, RAID 10 is the best. This is how I have my RAID 10 setup... 3x 2-port PCI-E (x1) controller cards....6x Hitachi 250 GB HDDs...each controller has 2 drives (mirror/RAID 1)...then in Windows I stripe across the controllers (mirrors)...presto!...700+ GB of space with the performance of RAID 0 and the redundancy of RAID 1.

And yes, RAID 0 DOES make a big difference with large file transfers.

 

[imported_goku]

Originally posted by: caz67
Yes, RAID5 is good, but it's not an option with only 2 drives..!!!

The OP COULD just combine the free space of both drives you know... A spanned volume. Oh BTW, if anybody knows about spanned volumes, please enlighten me as to whether or not if one drives fails, would you be able to access the data on the other drives with out issue or not. I'd assume you wouldn't because there is no striping of data... The only data I'd assume to be effected would be the data spanned on both volumes.

 

[SnoMunke]

I am only assuming since spanning drives in Windows requires the use of Dynamic Disks, if one drive fails all data is lost since the Dymanic Disk is broken. If data could be recovered, I would think it would only be done using Linux or some third-party software.

 

[Missing Ghost]

Originally posted by: goku
Oh BTW, if anybody knows about spanned volumes, please enlighten me as to whether or not if one drives fails, would you be able to access the data on the other drives with out issue or not.

For *nix systems, in my experience, you can still read the data from the online drive.

 

[fire400]

Originally posted by: Soviet
Ive used RAID for many years, its excellent stuff.

dumb

 

[SnoMunke]

Fire400:

Dude! Can I have dibs on your rig????????

 

[Fullmetal Chocobo]

Originally posted by: SnoMunke
Fire400:

Dude! Can I have dibs on your rig????????

He might trade for your Seagate 74gb Raptor.

 

[DAPUNISHER]

Originally posted by: magomago
Raid has a LOT of hype...especially if you are not working with large files such as intensive DV editing

Fvck speed, RAID5 lets me be lazy about backing up

Some keep bashing dual-core as not being very useful yet, but if you want to setup a nice and cheap software raid, the issues with CPU usage become a minor detail instead of a BFD. Just another benefit of dual-core processors IMO.