What does 10000RPM do for performance?

[Ice Czar]

I do alot of multitasking on my NAS
(which would be the closest box I have to a "typical single")
and I dont run a RAID0

3 single HDDs each on thier own channel (one shares w\ a DVD)
and a RAID5 array (currently changes all the time)

Im a big proponent of multiple drives optimized for what your doing
but I also have a hard time definatively ruling out a RAID0 2 HDD lean box for all possible users,
and my "offical postition" is test

 

[Ice Czar]

test isnt benchmarks
its logging your access pattern and making decisions based on it

(Im currently trying to learn to do that with my Linux boxes, wish I would have started my migration years ago)

 

[KoolDrew]

I would much rather buy a much larger drive for cheaper

 

[Ice Czar]

or lots of small older drives and a RAID5 card

 

[unhappy_mage]

That's five posts in a row with a in them.

Who am I to break a trend

 

[Ice Czar]

Im a happy guy
and Data Storage is widely known as a friendly place

 

[LoneWolf]

you dont get to be a mod without instinctively qualifying your statements
we are as vauge and slippery as the greasist polititian.

Or a retail salesman.

LoneWolf (5 long years in Retail Hell as a youngster)

 

[eggrock]

So your saying you couldnt tell a difference... and I read the tweakers review... and unfortunately while it was good,,, it missed the point entirely....
Too show the advantages of raid-0 they had to process images in photoshop and perform a network backup... Okay here is what I'm saying, that is not desktop usage... so the entire point is moot... we're talking about real life desktop usage...I' have personally never worked in photoshop while performing a network backup... or anykind of backup for that matter.

Apply that to benchmarking bragging rights... But that's okay in the [H] forums. RAID is not and should not be spoken of unless one is criticizing others for whatever reason.

Or perhaps a little more concisely:

Whatever.

 

[Ice Czar]

this whole thing is folliwing a trend that has many parallels
first its the end all and be all
then the backlash
then it normalizes

in phase one the drives where lacking and there really was quite a benefit
then the access backlash with the new drives, I admit to being on that bandwagon
then the re-examination and realization that like most things....

it depends

 

[Archer75]

"RAID 0 implements a striped disk array, the data is broken down into blocks and each block is written to a separate disk drive
I/O performance is greatly improved by spreading the I/O load across many channels and drives
Best performance is achieved when data is striped across multiple controllers with only one drive per controller
No parity calculation overhead is involved"

RAID 0 is not fault-tolerant and will make loading times(if your a gamer, spawning) much faster then your conventional 7200 rpm drive. People install win xp in under 4 min with a 10,000rpm drive and seek times are 2-3x faster then 7200's. The risk of RAID 0 is that if you lose one drive your data is toast. SCSI is better but this is a cheaper solution.

If I am wrong please correct me.

No, it will make it SLOWER. It only speeds up writing large files to disk. NOT reading a bunch of small files like you do for gaming. The benchmarks are out there. Feel free to google for them.

 

[Zepher]

We use a RAID 0 2 disk setup on one of our older editing machines since the drives a few years ago couldn't sustain the write speed required.
All the newer drives out now seem to handle the write speeds just fine now.
We just built another Matrox RTX.100 machine with an 80gig System drive and 2 160 A/V drives, one for capture/editing and one for exporting Native DVD Mpeg-2 video in real-time.

For my personal machines I have 10K rpm SCSI disks for the system drives and 7200rpm IDE for mass storage.

 

[DougLite]

As for 10,000 RPM (and the original topic) seeking faster, that is a function of the actuator and its logic, not spindle speed. Rotational latency (which is only affected by spindle speed) is not even included in manufactuer seek time quotes, although the SR average read service time test does account for rotational latency. With some changes to the logic board, it would be possible to take the actuator mechanism off of a modern 10K drive and put it on an 80GB/platter desktop drive and cut the seek time in half...Why doesn't it happen? Because the swifter seeks require a more powerful actuator, which is louder, and MUCH more expensive to produce to the same reliability standard.

Also, faster seek does not necessarily equal faster performance. The Maxtor MaxLineIII is quoted for 9.3ms seek, yet blows away all but the most recent wave of 10K SCSI drives and the 36GB Raptor in desktop performance, despite seeks taking 50% longer. It is also faster than a Hitachi 7K250 w/ 8.5ms seek, mainly do to the 16MB buffer and some other 'soft' improvements. In this SR comparison you will see that MLIII takes all of the single user tests except for Office DriveMark, despite being 5ms behind the 10,000 RPM drives (WD360GD and an older Seagate SCSI unit) and 1ms behind the 7K250. High spindle speed and low seek time do not a fast hard drive make.

It looks we're Back to Square One on RAID-0 folks. RAID-0 doesn't make anything more than minutely slower, and it is under very rare circumstances that it is slower than single drive. There is a controller overhead with RAID-0, and although it is not as bad as RAID-5, you are still not using full hardware accelerated RAID on a gamer mobo. So yes, there may be very specific situations where it is minutely slower than SLED, but not very often. On the other hand, the assertion that it "will make loading times(if your a gamer, spawning) much faster" is every bit as flawed as thinking it will make some tasks noticeably slower. The answer, like usual, is somewhere in the middle, and depends on each user's goals and methods.

 

[eggrock]

As for 10,000 RPM (and the original topic) seeking faster, that is a function of the actuator and its logic, not spindle speed. Rotational latency (which is only affected by spindle speed) is not even included in manufactuer seek time quotes, although the SR average read service time test does account for rotational latency. With some changes to the logic board, it would be possible to take the actuator mechanism off of a modern 10K drive and put it on an 80GB/platter desktop drive and cut the seek time in half...Why doesn't it happen? Because the swifter seeks require a more powerful actuator, which is louder, and MUCH more expensive to produce to the same reliability standard.

I heard somewhere around here about the Raptors using SCSI spindles, controllers etc. (hence the 5-year warranty.) Dunno enough to know if that's shens/BS/stupidity or not.

 

[Zepher]

I heard somewhere around here about the Raptors using SCSI spindles, controllers etc. (hence the 5-year warranty.) Dunno enough to know if that's shens/BS/stupidity or not.

I think it was the Server Editions of the Raptors that use SCSI technology in them.

 

[xonik]

What server edition? The Raptor is aimed at the "enterprise" buyer, if that's what you mean.

 

[MAngelo]

this whole thing is folliwing a trend that has many parallels
first its the end all and be all
then the backlash
then it normalizes

in phase one the drives where lacking and there really was quite a benefit
then the access backlash with the new drives, I admit to being on that bandwagon
then the re-examination and realization that like most things....

it depends

This also describes our pop culture perfectly

 

[SirKenin]

For workstation work, and heavy desktop productivty work, I can uderstand this truly disk intensive multi tasking model that Ice Czar references in the tweakers.net article. However, this is a gamer website, at least last time I checked. Anyone who is running that kind of disk intensive laod while gaming is an idiot. We're gamers - every FPS, and every eye candy/resolution level counts, and we, at least I hope, are not going to exhibit a usage pattern like Tweakers tests while we're gaming.

Yes, there are workstation users here on the [H] that do the type of work that can benefit from RAID-0, but they are much fewer in number than the gamers. I have no problem recommending RAID-0 to a user who needs a fast scratch disk for content creation, for a user who has a solid backup strategy and needs the highest sequential throughput possible, or a user who beats the drives to death with highly random, high queue work loads that can benefit form having an extra spindle on long stroke seeks. However, none of these scenarios describe a gamer usage model.

See I do a balance on my system. A bit of everything. Even when I am gaming I am multitasking in the background, so I am pushing my system to it's limits. If I have a bias, that's it. I don't have the gamers bias where they measure the size of their pecker by their 3dmark score or their bandwidth in MB/s or how far they can push that AMD before it melts. I'm not cheap. I buy a workhorse CPU. I buy top of the line everything and I push my system for all it's worth in a wide range of applications. It has to be reliable, it has to be on 24/7 and it has to be stable at the speeds I'm running at 365 days a year.

Sooooo... When I post on here about different types of hardware it's how it survives in commercial applications and under persistent loads. That's precisely why I say the Caviar is absolute garbage, yet the Raptor is amazing.

Storage Review means jack. None of the professionals I know refer to them or give two hoots about what they have to say. We all rely on personal experience, and incomplete benchmarks from SR simply don't tell the story. Same with that dork over at Anandtech. Even Tom Pabst *gag*

When you have 30% of your WDs or more go down to catastrophic failure, but SR goes "but they're great" or some punk on a message board goes "I've had four in my AMD gaming system and their great", you laugh at them. Trust me. Then you get pissed because you are getting crappy RMA service for your WD HDDs. Heh. grrrrr.

 

[DougLite]

See I do a balance on my system. A bit of everything. Even when I am gaming I am multitasking in the background, so I am pushing my system to it's limits. If I have a bias, that's it. I don't have the gamers bias where they measure the size of their pecker by their 3dmark score or their bandwidth in MB/s or how far they can push that AMD before it melts. I'm not cheap. I buy a workhorse CPU. I buy top of the line everything and I push my system for all it's worth in a wide range of applications. It has to be reliable, it has to be on 24/7 and it has to be stable at the speeds I'm running at 365 days a year.

Most people here on the [H] are also on a budget, usually $2000 or less. There is a price difference of over $350 between a P4 660 and a P4 630, $350 most people on the [H] can't afford. They make up the difference overclocking, and either save the bucks or allocate them elsewhere. I am not criticizing you, as there are other users like you that use their comps to make money, and therefore can afford to shell out for the top of the line parts, and that is a perfectly valid and justifiable expense, as it does insure the best out-of-the-box performance available.

Sooooo... When I post on here about different types of hardware it's how it survives in commercial applications and under persistent loads. That's precisely why I say the Caviar is absolute garbage, yet the Raptor is amazing.

Storage Review means jack. None of the professionals I know refer to them or give two hoots about what they have to say. We all rely on personal experience, and incomplete benchmarks from SR simply don't tell the story. Same with that dork over at Anandtech. Even Tom Pabst *gag*

Probably because most professionals don't have time to tweak and build every server they use - they have software problems to deal with and users to support, and the hardware is on a commodity basis and they just use the Cheetahs that ship with their server from Dell/HP/Sun/etc. Once again, not a criticism, but fact of life that servers are commodity units that someone else designs/supports, not the Pros using them.

When you have 30% of your WDs or more go down to catastrophic failure, but SR goes "but they're great" or some punk on a message board goes "I've had four in my AMD gaming system and their great", you laugh at them. Trust me. Then you get pissed because you are getting crappy RMA service for your WD HDDs. Heh. grrrrr.

I will completely dismiss this comment, because if WD desktop drives were really this bad, they would not be in business. I have had more WD drives fail than any other brand in my systems - 1 out of nearly 50 that have come through. It was serviced under its warranty promptly and satisfactorily. Hard drives are mechanical devices, they will fail. If you call one drive failure after decades of unit/year service poor reliability, then you are quite unrealistic. No other brand has had a failure in one of my or my customers' systems in the past 5 years. Going back to 1998, the only other member of the club is an 8.4GB IBM. Also, SR does not make reliability claims, only statements direct from the manufacturer about the warranty that backs the drive. Users are free to interpret data from the reliability survey as they wish. I personally do not use it at all, as there is nowhere near enough data, let alone consistent and unifrom data, to make informed conclusions about hard drive reliability. The only readily quantifiable statement about HD reliability is the length of the warranty - that's why you have backups. Once again, you have delivered gross generalizations and exaggerations and resorted to name calling to support your position, and looking at realities will show your exaggerations and generalizations to be far from the truth.

 

[SirKenin]

Most people here on the [H] are also on a budget, usually $2000 or less. There is a price difference of over $350 between a P4 660 and a P4 630, $350 most people on the [H] can't afford. They make up the difference overclocking, and either save the bucks or allocate them elsewhere. I am not criticizing you, as there are other users like you that use their comps to make money, and therefore can afford to shell out for the top of the line parts, and that is a perfectly valid and justifiable expense, as it does insure the best out-of-the-box performance available.

True true true. That is certainly a valid way of approaching the issue.

Probably because most professionals don't have time to tweak and build every server they use - they have software problems to deal with and users to support, and the hardware is on a commodity basis and they just use the Cheetahs that ship with their server from Dell/HP/Sun/etc. Once again, not a criticism, but fact of life that servers are commodity units that someone else designs/supports, not the Pros using them.

I know I don't necessarily have time to fart around and try and "tweak" a client's computer, because that costs me money that I won't necessarily realize when it's time to submit the bill. I expect the thing to just work and that's why WD Caviar isn't the drive for me.

Once again, you have delivered gross generalizations and exaggerations and resorted to name calling to support your position, and looking at realities will show your exaggerations and generalizations to be far from the truth.

It might not be a reality in your circles, but it certainly is in mine. Same with those crappy Maxtor thinline drives (and those infamous 75 GXPs. *gag*). I loved Quantum Fireballs. I was sad to see them get snatched up by Maxtor.

 

[DougLite]

I don't use Caviars, unless they're on rebate special, either. You will notice they come as not recommended in my Buyer's Guide at the top of this forum. I will shell out more than 50 cents per GB on a Hitachi, Seagate, or Samsung, but not WD, and in many cases, Maxtor. WD is fairly close to Hitachi on 7200/8MB performance, but just a little behind Hitachi/Maxtor/Samsung and way behind Seagate on reliability. Maxtor may be equal or a little better on reliability, but lags on performance in the 7200/8MB sector, which is why the 7200/8MB Maxtors come as not recommended in the Buyer's Guide. This takes nothing away from Atlas and MaxLine, just like Caviar doesn't take away from Raptor in your perception.

I will not disagree with your assessment of WD having the least reliable desktop drives out there. Once again, they are the only drives I've had to replace recently. However, to say that they are absolute junk is an exaggeration. The differences in reliability are very small, and the vast majority of hard drives shipped will either lead a full service life or fail due to factors that are not the drive's fault, usually poor quality of power supplied to the drive. I have 7 Caviars, all of them retail rebate specials, humming away right now. I realize you shudder at his prospect, but follow me for a few more sentences. For over a year, I had 10 of them, some of which have been traded/sold, and to my knowledge, are still working. Of course, I do not rely on the drives - we have multiple off site backups for data and custom reinstall CDs to get a box back in action quickly if a drive does out.

Wait a tick - I just said I don't rely on the drives. That means that the whole concept of hard drive reliability is a misnomer. Never trust something to a hard drive that you don't want or can't afford to lose.

I realize that you, as a VAR, lose money everytime you have to replace a hard drive. If you choose to not market Caviars, then that is a business decision that you make, and I respect that. However, even your experience, while much larger than mine, does not account for every Caviar sold. However, you don't need to spitefully spread misinformation about hteir products. WD routinely wins best HD maker polls among enthusiasts, that isn't just because of Raptor. If anything, enthusiasts care more about HD reliability than you do, as many of us don't do backups Also, even WD admits that Caviar is not designed for enterprise duty cycles - that's why they have Raptor and Caviar RE. WD does make capable drives for desktops - their large installed base and continued viability without a high profit SCSI line is proof of that. Once again, if they were as bad as you would have everyone believe, they would have gone out of business long ago.

 

[mikeblas]

Yes. For OS/boot drives, or gaming, all you are really doing is doubling your chance of failure. For multimedia editting of large files, spanned disks have some advantages.

RAID0 is striped, not spanned.

 

[Zepher]

What server edition? The Raptor is aimed at the "enterprise" buyer, if that's what you mean.

That is the what I meant, couldn't remember it off the top of my head