So, who's getting a WD Raptor?

[Adcadet]

Anybody?

 

[Jake the Dog]

I know I am, it's the when that I don't know.

 

[Mickey]

I plan to; it's just a question of when they become available. My sources say they are due any day now in distribution.

 

[Adcadet]

what's up with the lack of command queing? If WD is aiming at the server market, don't you think they would have included it. And if they're aimin for the server market, why bother with 8 MB cache? My guess is that they're aiming for the enthusiast market first, then will add command queing later when it looks like they'll be more accepted into the server market.

 

[Mercutio]

Not interested. Not enough space. Plus not enough Maxtor or Samsung.

 

[blakerwry]

I'm intersted. I'll not be getting it "just for the hell of it". But if I need a new boot drive, it's at the top of my list.

 

[CityK]

Blake,

Given your recent frustration with your WD80JB, you might want to take note of what Eugene wrote here.

I have a feeling that its noise level will make it one annoying SOB.

CK

 

[blakerwry]

yes, before i buy another drive i will make sure i listen to a sample...

This is why the 800JB now lives in a box in the closet (when I say box, I mean a cardboard thing with 6 sides and no CPU) and my 180GXP is being happily used in my main PC.

 

[Tannin]

Can't say I'm all that excited. Guess I ought to be, I've waited for 10K IDE drives for long enough, but now that one is on the radar .... shrug.

Why the underwhelming boredom factor?

Well, for myself, there is no way it's going to rival my X-15s, and when I need more space I just add cheap, quiet, ultra-reliable, Samsung 7200s, which are fast enough for storage purposes.

For sale? Maybe.

We have been sending too many Western Digital drives back faulty this last couple of years, and the replacements are slow to arrive, refurbished, and all too often have crap quality control. Not as bad as the appalling Seagate U Series replacements, which I don't even use anymore - I flog them off to the trade, cheap, with 0 days warranty - but bad enough.

So I'm reluctant to trust the Raptor. It's certainly not something I'd feel comfortable putting my own data on, so how can I justify recommending that people buy it? Still, the 5 year warranty is attractive.

Maybe.

 

[LunarMist]

Anybody?

No. Not until a few more platters are added.

 

[blakerwry]

a few more platters means this thing falls out of my price range... at 36Gb it's plenty for my OS/App drive... which is right where I want it.

 

[Buck]

They're on my pricelist and readily available at the moment. But I'm not overly excited about the drive. I would need to get a SATA controller card too. By the way, who says the drives don't do Command Queueing?

 

[Fushigi]

Like Tannin said, the drive is no replacement for my X-15 and I already have a WD 160GB data drive.

If I build a new box, I may give it consideration. But I doubt I'll have a use for a drive like this anytime soon.

- Fushigi

 

[LunarMist]

I would need to get a SATA controller card too.

Is there not a parallel ATA version of this drive? If the only interface is SATA, then it is totally useless as an "upgrade" in many systems.

 

[e_dawg]

If it is available in pATA format, I would definitely get one in the summer. However, I think WD said it will be sATA only, so I will likely get one next year when I get a new mobo (with onboard sATA controllers).

 

[Pradeep]

Given I already have a SCSI controller, I'd be more likely to either get a 3rd gen X15 or some kind of rather large 5400rpm IDE mass storage drive.

 

[Buck]

I would need to get a SATA controller card too.

Is there not a parallel ATA version of this drive? If the only interface is SATA, then it is totally useless as an "upgrade" in many systems.

It will only come in the SATA flavor. The connector end has a SATA power connector, SATA data connector, jumper block (which is of no use and doesn't do anything), and a molex power connector.

 

[Mickey]

I haven't found a local shop that carries SATA controller cards, though. I don't want to get a new motherboard, since my P3V4 is working well for me. Plus, in all honesty, I don't need the speed; I'm getting it for the geek factor.

 

[Dïscfärm]

The WD Raptor product seems rather obviously a rush job -- a rush to get a product out because WD probably caught wind of someone else preparing to release a 10kRPM ATA drive. And, if it really is a "rush job" product, it could be obviated by its replacement in as little as 6 months.

 

[blakerwry]

Sure, power on causes wear. But starting and stopping and heating and cooling cause a hell of a lot more wear. We build computers every day and usually look after them for their entire working lives. And while I think we build pretty good systems, we have our share of failures: drives that die, main boards that go kaput, and so on. But in our own systems, the five systems that run the office and the one I have at home, we have had the following parts fail in the last ten years. (Yes, I can list them exactly - there are not so many that I can't remember them.)

• The Micropolis 2217 (my own fault)

• An Intel ISA network card in my main server, which just died one day. It was maybe four years old

• The CPU fan in my Thunderbird 1400, which then cooked the chip but not, astonishingly enough, the motherboard.

• The power supply in my DOS box, which died about 5 years ago.

• The replacement power supply in my DOS box, which has gone intermittent this last week or two. I should be changing it now - but here I am wasting time on the web again.

• The CPU in my DOS box, an AMD 5x86-133 which has been in there since near-new and has never had a fan on it - yeah, yeah, I know. You are supposed to use a fan on those things, and we would never dream of leaving the fan out of a customer's machine, but it was only a 5x86 and I was curious to see how long it would stand it for. Ans: about five years. It used to get very hot, too hot to touch, but it never used to crash. The crappy old Pentium-100 board I replaced it with (properly cooled this time) is not as stable, and no faster than the old thing was for the tasks this box usually does - formatting floppy discs and stuff like that..

• Various assorted keyboards, mice, floppy drives, and CD-ROM drives that just wore out or got tea poured on them or had other nasty accidents..

• The Sony 40-speed CD-ROM drive in the spare back office machine, which started playing up not long ago while still under warranty. Actually, come to think of it, that machine does get turned on and off a bit. Sometimes I get sick of the sound of two computers under my desk and switch it off for a while. (It has two of the oldest SCSI drives, by the way: a 1GB IBM and a 2.2GB Quantum.)

Most of the time, most of those machines have had IDE drives in them.

The secret is: turn off all that power management crap. Don't let your drive spin down, because if you spin it down you have to spin it up again, and spinning a drive up is very hard work. And if you let it cool down you have to warm it up again, and that heating-cooling cycle is what kills electronic circuit boards. They expand, they contract, the traces become brittle and eventually they develop hair-line cracks in them. That's what kills components.

Active cooling does no harm, but all you need is decent airflow in the case. I use active cooling for one of my two X15s because it shares a fairly small case with a very large CPU and the case lives in a sealed cupboard to keep the noise down - which is why my XP 1700 runs only just cool enough.

Oh, and don't be too hard on my DOS box with its habit of blowing power supplies. Sure, it's about to get its second new PSU, but they were only cheap ones and it started life as a 12MHz 286. I had more hair then.

I had an ISA intel NIC fail on me for no reason too... it just suddenly one day decided not to work.. if i remember correctly it would freeze the PC when windows was trying to load.. or maybe just freeze the PC altogether...

Never worked that good anyway... ping times within the LAN were very high (~20ms) and data rate wasn't that great either... replacing it with another ISA 10baseT NIC solved that problem (1ms pings) and.. well.. still not that great data rate of course.

 

[jtr1962]

To put things in perspective, the access times are less than a factor of two improvement over even run-of-the-mill 5400 RPM drives. Even an X15 is only about 3 times faster than 5400 RPM. Unless it's about the same price per GB, just as cool, and just as quiet, I simply couldn't justify it. Right now it's none of those, so that means no Raptor for me anytime soon. Thinking about it, most of my programs load in less than 5 seconds anyway, so other than defragging, I can't think of any applications that will benefit much from a drive like this. So what if I shave 20 seconds off my boot time, or a program loads 1.5 seconds faster?

I can't wait for solid state. Not so much just for the extra speed but much more for the greater reliablity it should deliver compared to mechanical disks and the complete lack of noise. Also the fact that I would be able to move a running computer without worrying about head crashes.

 

[blakerwry]

is there *any* advantage to mechanical storage when compared to sold-state storage (besides cost)?

 

[time]

How about a 100:1 advantage in capacity?

BTW, I'm planning to construct a 20 drive cluster of Raptors.

 

[Tea]

Yup. Solid-state storage is boring.

 

[blakerwry]

Yup. Solid-state storage is boring.

aww, you didnt mentino the old post i dug up tea... how old where you then?

 

[blakerwry]

Yup. Solid-state storage is boring.

aww, you didnt mentino the old post i dug up tea... how old where you then?

As already mentioned earlier tonight.. my spelling goes to pot after 12:00am... maybe I need to clean this keyboard...

 

[Tea]

I was 7. Or possibly 5. I forget.

That's the nice thing about being imaginary. You stay the zame age all the time. So I was 7 last year, and I'm going to be 7 again next year. Which is good, because I like being 7. (Or maybe 5.) (I get confuzed sometimez.)

 

[Buck]

The WD Raptor product seems rather obviously a rush job -- a rush to get a product out because WD probably caught wind of someone else preparing to release a 10kRPM ATA drive. And, if it really is a "rush job" product, it could be obviated by its replacement in as little as 6 months.

Interesting observation. From the information that I received, WD had this drive on their roadmap at the beginning of 2002. That is much longer then their other products stay on their roadmap before release.

 

[Jan Kivar]

Yup. Solid-state storage is boring.

aww, you didnt mentino the old post i dug up tea... how old where you then?

As already mentioned earlier tonight.. my spelling goes to pot after 12:00am... maybe I need to clean this keyboard...

No no Blake, You just think faster than You can type... :lol:

Jan

 

[Mickey]

Interesting observation. From the information that I received, WD had this drive on their roadmap at the beginning of 2002. That is much longer then their other products stay on their roadmap before release.

That's what I heard, too. I've never seen a new platform come out after only a few months of work; it usually takes closer to a year at best for a new platform.

 

[Dïscfärm]

Interesting observation. From the information that I received, WD had this drive on their roadmap at the beginning of 2002. That is much longer then their other products stay on their roadmap before release.

If you've heard that it's been in development since Q1 / 2002, then at least the mechanicals aren't a rush job. Otherwise, the performance of the drive (as first published) as well as the connectors used pretty much scream out Rush Job.

 

[mubs]

Over at SR, MaxtorScsi siad in a thread that it takes two years from the time a drive is dreamed up to actual production. Mickey read that and responded like he did here (1 year). Maybe Mickey works for WD :lol: ?

 

[Mercutio]

Mickey refuses to comment whenever I accuse him of that. But then, he also knows more about WD drives than even Buck, and occasionally mentions designing drives for a living, so I think we can draw our own conclusions.

 

[Dïscfärm]

Behind the hardened steel doors of every hard drive manufacturer, guarded by rabid German shepherds nonetheless, are the laboratories where various "candidate" drives spin the hours and days away in burn-in chambers. And, within these burn-in chambers simulating years worth of "normal" operational time in weeks -- where the ambient temperatures are high and the I/O loads are heavy -- are the hard drives that will become the next new model. A group of each of these drive mechanisms could, presumably, have an assortment of interfaces attached (PATA, SATA, SCSI, SAS, F-C), with each interface receiving firmware updates, when needed, by the engineers evaluating them.

Realising all of the above, someone might then suspect that suddenly one day late in 2002, a Product yay-hoo at WD ordered a 10kRPM SATA drive to be delivered ASAP. March 2003, it arrives -- with firmware still being updated as the cardboard boxes arrive at the reviewers' doorsteps.

 

[Mickey]

Behind the hardened steel doors of every hard drive manufacturer, guarded by rabid German shepherds nonetheless, are the laboratories where various "candidate" drives spin the hours and days away in burn-in chambers. And, within these burn-in chambers simulating years worth of "normal" operational time in weeks -- where the ambient temperatures are high and the I/O loads are heavy -- are the hard drives that will become the next new model. A group of each of these drive mechanisms could, presumably, have an assortment of interfaces attached (PATA, SATA, SCSI, SAS, F-C), with each interface receiving firmware updates, when needed, by the engineers evaluating them.

Realising all of the above, someone might then suspect that suddenly one day late in 2002, a Product yay-hoo at WD ordered a 10kRPM SATA drive to be delivered ASAP. March 2003, it arrives -- with firmware still being updated as the cardboard boxes arrive at the reviewers' doorsteps.

:lol: That was great! A disk drive hatchery!

That could explain the teeth marks on my rear. Forgot my security badge in the car when I tried to get to work. (Sorry, I've had a stressful day at work so I'm being silly now)

I mentioned 1 year design cycles because that's standard for IDE/desktop drives, especially new platforms. Since I work in that area, it's what I'm most familiar with. HDD designers are a rather small group of folks, so people tend to cycle between companies, hopping between SCSI and IDE or even laptop design groups.

In talking to folks that *have* done SCSI drive designs in the past, they specifically mention having longer design cycles (18 months to 2 years). By design cycle, I mean from initial product definition (spin speed, interface, areal density, and other performance characteristics) until it's in high volume production.

If it's a "make-from" design, or one that is an evolution of an earlier design/platform, the design cycle is 3 months to 6 months, typically. Again, I'm basing this on what I've seen firsthand for IDE drive designs, not SCSI designs. No doubt MaxtorSCSI could give the inside view on SCSI drive designs.

Lastly, I've never stated where I work because I'm not here to represent my firm and would really rather keep my job. I'm here because I like talking about storage and exchanging general information. I'll leave the marketing department to formally speak for the firm.

 

[Ekaf-Ami]

Forgive me for pointing out obvious but possibly hurtful personally fact, Mr Mickey, but (with all respectful to fine gentleman that you be) you not be having anywhere near enough stupidities and non-talents for being marketing department person. Best you stay humble engineer.

 

[Prof.Wizard]

Yes, if I also add a SATA controller.

 

[CougTek]

It appeared yesterday on my pricelist (although still back ordered). 259$CDN for 36GB. About twice the price of a 40GB 7200rpm ATA drive, which typically cost 125$CDN.

 

[LiamC]

Well Coug, that's made up my mind.

I have bought just about every performance upgrade drive since I started to care about performance

850MB Seagate Decathlons from various Maxtor/Seagate drives
4.3GB Fireball ST (I think) -> 850MB Seagate Decathlons
13.6GB Fireball KA+ -> from 4.3GB Fireball
75GXP from Fireball KA+
WDBB, 60GXP from 75GXP
JB series in 33,40 & 60GB platter models

At twice the price, and I've only ever seen incremental improvements in performance, I'm jumping off of the train

 

[blakerwry]

i allready jumped off the train when i didnt notice any direct improvement from going to 2x75gxp's to a single 800JB..

The 800JB is basically 3 minor generations(15->20->33->40) ahead and yet didn't make a noticeable improvement.

One thing that I'm hoping will make an improvement is the decreased access time of the raptor while maintaining top notch STR speeds.