Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Hard Drive Overclocking Competition From Secau

timothy posted more than 3 years ago | from the spit-it-out dept.

Data Storage 162

Blittzed writes "We were reminiscing about the good old days of overclocking CPUs and memory, and the subject of hard drive overcloking came up. The discussion / argument we were having in the research lab ended up in a bet which now has to be settled. So, we are putting our money where our mouth is, and putting up $10,000 to anyone who can read a 500GB drive in under an hour. We will also consider other attempts with a smaller amount of money in the event that the one hour is not possible. There are a few rules (e.g. the drive still needs to work afterwards), but otherwise nothing is ruled out. Specific details can be found on the URL. Go let the white smoke out!"

cancel ×

162 comments

Sorry! There are no comments related to the filter you selected.

First Post (2)

Niedi (1335165) | more than 3 years ago | (#36629262)

And the link is dead already? That was quick...

Re:First Post (4, Funny)

wizzy403 (303479) | more than 3 years ago | (#36629276)

Someone let the white smoke out...

Re:First Post (4, Funny)

eharvill (991859) | more than 3 years ago | (#36630410)

We have a new pope already??

Re:First Post (1, Redundant)

cultiv8 (1660093) | more than 3 years ago | (#36629354)

The OP forgot to mention which HDs they were trying to overclock...

Re:First Post (1)

Sulphur (1548251) | more than 3 years ago | (#36630400)

The OP forgot to mention which HDs they were trying to overclock...

How big is your hard disk?

He replies with hand gestures.

Re:First Post (0)

hairyfeet (841228) | more than 3 years ago | (#36631078)

I have a much better question....why? While I don't know if it is the same with Linux (I gave up messing with it around Ubuntu 10.4, but I'll assume it is) with Windows and the frankly insane amounts of cache one gets on a drive nowadays combined with Superfetch one will rarely notice your drive! In fact lately I've been using 5900RPM drives as OS drives because with 64Mb caches and plenty of RAM there just isn't enough of a difference in speed to deal with the extra noise and heat, especially if the customer wants it as an HTPC.

So honestly I don't get it. you OC a CPU (or in the case of AMD unlock cores as well as OC) because you get a more expensive CPU at a cheaper price. Same thing with RAM or flashing an HD4830 into an HD4850 which I have done a couple of times so far and works well. But with HDDs you simply add a second in RAID 0 or if you don't mind dealing with the hot/crazy scale [codinghorror.com] and the possibility you may be spending several hundred a year on replacements and be risking data? Well then go SSD. But it isn't like OCing that EcoDrive is gonna turn the thing into a Velociraptor, it is just gonna wear the thing out quicker for less gain than you would have seen adding a RAM stick. I just don't get it, folks with more time on their hands than they know what to do with I guess.

Re:First Post (1)

Joce640k (829181) | more than 3 years ago | (#36631270)

I'm failing to grok this. My two year old Velociraptor can sustain something close to 138MB/s transfer with no tweaking (the speed needed to read 500GB in an hour).

Is there really no enterprise-level drive that can manage this...?

Re:First Post (0)

Anonymous Coward | more than 3 years ago | (#36629404)

can the original poster repost the rules here so we can get the concept?
sudo hdparm -tT /dev/hda
tests the combined throughput of the processor, cache and memory. This might be more involved than initally appears.

Perhaps instead of such a large prize... (2, Funny)

Aranykai (1053846) | more than 3 years ago | (#36629314)

They should have considered spending some of it to upgrade their hosting.

An hour? (5, Funny)

hawguy (1600213) | more than 3 years ago | (#36629340)

An hour!? I have a 500GB drive on my desk and I can read it in under a minute! The first line says: "Seagate Barracua 7200.11 500 Gbytes" The entire label has only a few dozen words and serial numbers.

Re:An hour? (4, Funny)

Dunbal (464142) | more than 3 years ago | (#36629450)

With a Seagate Barracuda I think the challenge is getting the thing to actually run for over a minute.

Re:An hour? (4, Informative)

0123456 (636235) | more than 3 years ago | (#36629472)

With a Seagate Barracuda I think the challenge is getting the thing to actually run for over a minute.

I just ran smartcl here and the two Seagate Barracudas in this machine have each been running for 29,908 hours.

Re:An hour? (5, Funny)

jamesh (87723) | more than 3 years ago | (#36629492)

With a Seagate Barracuda I think the challenge is getting the thing to actually run for over a minute.

I just ran smartcl here and the two Seagate Barracudas in this machine have each been running for 29,908 hours.

I see what you're saying... even the SMART data is corrupt.

Re:An hour? (1, Redundant)

flappinbooger (574405) | more than 3 years ago | (#36629694)

I see what you did there...

Re:An hour? (1)

TheDarkMaster (1292526) | more than 3 years ago | (#36629780)

This one made my day, thanks :)

Re:An hour? (0)

Anonymous Coward | more than 3 years ago | (#36630030)

I see what you're saying... even the SMART data is corrupt.

When I worked at Dell, the common statement was that SMART successfully finds 10 out of every 6 actual errors.

Re:An hour? (1)

EdIII (1114411) | more than 3 years ago | (#36630202)

I have been laughing for at least 5 minutes hysterically....... you pwned him so good. LOL

Thank you

Re:An hour? (2)

Luckyo (1726890) | more than 3 years ago | (#36629706)

I can top this with my old 2x 200GB seagate barracuda 7200.7 drives. Used to be used together in raid0 on my old machine, now been in use as separate drives (one as system drive).

They're yet to cause problems, unlike several other brands I had to kick into the curb while these two lived. Their power on time is reported as 43092 and 45394 hours respectively by S.M.A.R.T.

You're probably talking about that specific failure in barracuda family, 7200.11. I had one of those, and had the typical problem (logic board dies). Exchanged it on warranty, they sent back a similar size 7200.12. No problems with that one either.

Re:An hour? (1)

PopeRatzo (965947) | more than 3 years ago | (#36629790)

I've got an old Seagate 2.1Gb SCSI Barracuda that's been running since the 1950's.

You start it with a pull-cord, like an old lawnmower. Sounds about the same, too.

I'm not sure I can read that in under an hour, though.

Re:An hour? (1)

Sulphur (1548251) | more than 3 years ago | (#36629982)

I've got an old Seagate 2.1Gb SCSI Barracuda that's been running since the 1950's.

You start it with a pull-cord, like an old lawnmower. Sounds about the same, too.

I'm not sure I can read that in under an hour, though.

Will it run Stuxnet?

Re:An hour? (1)

number11 (129686) | more than 3 years ago | (#36630304)

I've got an old Seagate 2.1Gb SCSI Barracuda that's been running since the 1950's.

Now, that's impressive. Presumably a secret project that IBM stole for their first model, which was introduced in 1956. But IBM's only had fifty 24 inch platters, with a total capacity of 5MB, and it needed 3-phase power and a forklift to move it. Yours is a lot bigger. But is it faster than IBM's (whose access time was close to 1000 ms)?

Re:An hour? (1)

rmaureira (1414691) | more than 3 years ago | (#36630456)

You forgot the "GET OFF MY LAWN" part... *sigh*

Re:An hour? (1)

number11 (129686) | more than 3 years ago | (#36630572)

You forgot the "GET OFF MY LAWN" part... *sigh*

Hey, that's implied by my low UID. It shouldn't have to be explcit.

Obligitory: You must be new here... (1)

rts008 (812749) | more than 3 years ago | (#36631278)

For saving confusion in the future, Grasshopper, you should automagically append "GET OFF MY LAWN!!!" to every comment from a six digit /. UID.

See, sometimes the Geritol hasn't had time to kick in yet, and adding that to our comments gets overlooked in the frenzy of typing the reply.

Oh yeah, and...GET OFF MY LAWN!!!

Re:An hour? (1)

hedley (8715) | more than 3 years ago | (#36630736)

He means it's been running since about 10 to eight.

Re:An hour? (2)

SharpFang (651121) | more than 3 years ago | (#36630988)

The problem is not the old disks. Actually, the older, the more reliable. It's the newest disks that are the worst. When you boast "My disk is running fine for 5 years already" you're talking about a disk from 5 years ago. And it's the disks from 2 years ago that keep dying on us. Tollerances get

Re:An hour? (1)

ian_from_brisbane (596121) | more than 3 years ago | (#36631244)

Tollerances get

Must have been using one of those new disks.

Re:An hour? (0)

Anonymous Coward | more than 3 years ago | (#36631068)

Bah! I'm using an old millstone rotated by a crazy donkey. If I stop the donkey, I can read the whole millstone in half a minute!
Writing, though, needs a quick handyman with his chisel.
And don't you dare, youngsters, bring your newfangled lawnmowers on my lawn!

Re:An hour? (2)

Freultwah (739055) | more than 3 years ago | (#36629532)

You won't be able to push any more than 18 gigabytes in a minute through SATA-II and that's in theory. So theoretically one could read a 500 GB drive in ~28 minutes, but the drives just aren't nowhere near as fast. Then again, maybe your Barracua is many fold faster than Barracudas. I know my Sonny cassette player was faster than that from Sony.

Re:An hour? (1)

Anonymous Coward | more than 3 years ago | (#36629674)

I think you missed the point. He can read all the text that is physically printed on the 500GB Hard Drive in under a minute.

My speed-reading ability has (thus-far) not been limited by SATA-II bandwidth.

Re:An hour? (0)

Anonymous Coward | more than 3 years ago | (#36630034)

The page isn't up, so I can't check the particulars of the contest, but if SATA-II was the limitation, you might be able to get around it by getting a 4K drive and reading only a small part of each sector (sequentially, of course.) This would cause the entire sector to be read by the drive's controller and yet only a small fraction of the data passed on through the SATA-II interface. All the data would still be technically read by the drive.

But, as you said, SATA-II isn't really the limitation...I've only heard of pro SSD models coming close.

Re:An hour? (1)

symbolset (646467) | more than 3 years ago | (#36630486)

The SAS 6G and SATA 3 (6Gbps) models of SSD go up to over 500GB now. Reading that in a few minutes is no big deal. Even the SATA II Intel 320 series [intel.com] does 600GB and sequential reads at 270 MB/s, which would be 600GB in (600000/270 seconds) - 2222 seconds or just over 37 minutes. My laptop has a better data rate, but I use off-brand components :-). This is no problem at all.

A spinning rust platter isn't ever going to dish that, but if this is a job you need done and you're willing to spend ten grand, I'll take your money all day.

And let's not have the Base2 Vs Base10 argument, OK? A 2.4 percent difference isn't going to change the outcome of this one.

Re:An hour? (1)

spire3661 (1038968) | more than 3 years ago | (#36630770)

This is exactly why I limit any working computer with actual unique data to 500 GB or less. Beyond that it really becomes alot harder to maintain your dataset.

Re:An hour? (1)

symbolset (646467) | more than 3 years ago | (#36630960)

That's rather limiting. There are PCIe attached solutions that consistently read/write at more than 6GB/s rather than 6Gb/s - like for example the ioDrive Octal [fusionio.com] . It can have far more storage than your limit - ten times as much on one card. That thing has a serious 48Gbps serial read bandwidth, sustained, and you can configure many PCs with eight or sixteen of them. This is only one of many. There are actually some applications that strain against the limitation of this bandwidth.

The good rule for a thumb rule is that it should be general enough to scope its use. By bragging your rule on slashdot you've only defined the limits of your own vision and use. You're advertising that you're either "old school" or you're limited by some vendor's products, or late. Not a good place to be.

Here's a good rule of thumb: A storage drive can never be so fast, nor so capacious, nor responsive enough to serve every storage customer's need - but on a clear disk you can seek forever.

Do you see what I did there?

Re:An hour? (0)

Anonymous Coward | more than 3 years ago | (#36629580)

An hour!? I have a 500GB drive on my desk and I can read it in under a minute! The first line says: "Seagate Barracua [youtube.com] 7200.11 500 Gbytes" The entire label has only a few dozen words and serial numbers.

Barracua [youtube.com] .... oh yeah .....

Re:An hour? (1)

SharpFang (651121) | more than 3 years ago | (#36631004)

Oh, but did the rules state the data read needs to be the same as the data written? Error free?
Just grab the data from the cache without worry if the cache got to be filled correctly and enjoy superior read speed!

Re:An hour? (1)

MurukeshM (1901690) | more than 3 years ago | (#36630250)

And you still couldn't read it correctly. Epic.

Sarcasm filter broken? (1)

rts008 (812749) | more than 3 years ago | (#36631320)

Whooosh!

I think you need some PF Flyers[1], AND a trampoline to catch everything flying over your head.

[1]From the wiki on PF Flyers [wikipedia.org] :

...were very popular in the 1950s, renowned for helping you "run faster and jump higher..."

emphasis mine

Oh yeah, and 'Get off my lawn!!'

White smoke? (0)

damn_registrars (1103043) | more than 3 years ago | (#36629358)

I think there is some coming from the server that hosts that website.

Re:White smoke? (0)

Anonymous Coward | more than 3 years ago | (#36629702)

There's some white smoke blowing around in here. Wait, what were we talking about?

White smoke, bah! (1)

Rungi (1098221) | more than 3 years ago | (#36629424)

If you're OC'ing an SSD, wouldn't it be blue smoke? http://www.ncat.edu/~dowtin/pc/smoke.html [ncat.edu] Can you effectively OC an SSD?

Is that all? (1)

Anonymous Coward | more than 3 years ago | (#36629434)

Um, if all that has to be done is read the drive in under an hour, what's to stop me from just putting an additional set of heads and control electronics on the other side of the disk? I know there used to be such disks that could only read from one set of heads as a marketing thing for forward facing web services...

Re:Is that all? (1)

spydum (828400) | more than 3 years ago | (#36629482)

Exactly -- I always wondered why this was not done -- is it a limitation of the form factor? Why not have two arms? We already use multiple heads, multiple platters. Seems like you could double the performance or at least allow a minimal cost error checking (single disk-level mirroring?) with such a solution.

Re:Is that all? (4, Interesting)

a_nonamiss (743253) | more than 3 years ago | (#36629536)

According to the rules, it needs to be reversible. They mention forensics, so maybe they're trying to do it undetected. At any rate, I'm pretty sure cracking the seal on the hard drive is verboten.

Re:Is that all? (1)

jones_supa (887896) | more than 3 years ago | (#36631216)

Then how can you do almost anything?

Re:Is that all? (1)

blair1q (305137) | more than 3 years ago | (#36629574)

Dunno what you're talking about. Disks used to have multiple r/w arms. They also used to be the size of your desk. Putting another arm in the housing would only work if it was on the opposite side from the one that's there, but now your housing is 4 cm longer, and you've got extra wire causing latency and skew problems.

Re:Is that all? (2)

tibit (1762298) | more than 3 years ago | (#36629748)

Head preamps are usually somewhere on the arm assembly, and they drive controlled impedance differential pairs, so an extra inch or two shouldn't be that big of a deal. Latency is not an issue at all, each arm would be controlled separately and they don't need to be synchronous at all.

Re:Is that all? (1)

Grishnakh (216268) | more than 3 years ago | (#36629606)

There have been hard drives in the past with two arms. I don't think it helped performance enough to make up for the cost difference.

Re:Is that all? (1)

mlts (1038732) | more than 3 years ago | (#36629614)

I've seen drives with two arms before, configured multiple ways:

One drive had one arm just for reading, one for writing.

One was made so if one arm failed, the other could continue. In the mean time, one arm did the work, while the other one just stayed idle.

One was similar to the previous, except both arms did reading/writing at the same time.

I think it didn't catch on because of cost, but could be wrong.

Re:Is that all? (1)

Z00L00K (682162) | more than 3 years ago | (#36630688)

I would also foresee problems with precision involved. Hard to keep heads on track.

Re:Is that all? (3, Informative)

Sloppy (14984) | more than 3 years ago | (#36629778)

I always wondered why this was not done

If you look at it the right way (translation: I'm about to break a rule) it's done all the time. It's called RAID0.

But seriously, that tells you why it's not done: because if your really care about performance that much, you can get more performance than a multi-head-set drive and spend less money by using commodity parts. If you make a drive that works this way, no one will buy it. (Except for money laundering purposes. ;-)

Re:Is that all? (1)

Anonymous Coward | more than 3 years ago | (#36630398)

Well, it's not strictly equivalent to RAID0... (Yeah, you said break a rule, so I'm not saying "you're wrong!", just indulging in the essential and fun pedantry without which /. is not.)

Say you have a video recording application where you're writing a video stream to disk, and that (perhaps uncompressed?) stream is of such ungodly bandwidth as to take a significant chunk of your drive's throughput. One head's fine if your disk isn't ridiculously fragmented (which it won't be); you have RAM to buffer it while the drive seeks occasionally (e.g. past a file fragment to the next unallocated space), then it'll catch up. But now suppose you want to playback a timeshifted stream of this same bandwidth -- even if you jump to RAID0 with two disks, thus doubling your throughput, you should have plenty of time to read/write. Only now you're seeking back and forth, and that latency really digs into your throughput. Maybe if you add 2 or 3 disks, it'lll work; maybe if you have enough RAM for huge buffers and a properly tuned I/O scheduler, you can read and write in big enough chunks to minimize the seek penalty. But a 2-headed hard drive will do just fine, and (at equivalent quantities) would be even cheaper than 2 identical drives, and similarly priced to 2 half-capacity drives for the RAID0 setup. Of course, since very few applications fit that profile, quantities are not the same, and RAID0 with however many disks you end up needing is still cheaper.

So there is some sense in a dual-headed drive for very specific applications (something like DVR is the most obvious thing, but anywhere you have a real-time writing requirement where bandwidth is a significant fraction of drive throughput AND you can't put it on a separate drive from other IO), but there's nowhere near enough of those applications to justify it.

Re:Is that all? (1)

adolf (21054) | more than 3 years ago | (#36630672)

From the practical aspect: My U-Verse DVR had a not-very-special 2.5" drive in it, and was able to record four things at once while replaying a fifth in my not-special configuration at home. (I believe it can actually do more than that with multiple receivers networked to it, but I just had the single DVR box.)

That said, in the interest of pedantry: Unlike a striped RAID 0, a RAID 1 array of n+1 disks could conceivably perform as a single disk with multiple heads, since a RAID 1 of n+1 has n+1 worth of independent head stacks, all reading identical data. (Also in the interest of pedantry, n is 1 or greater, since otherwise it is perfectly possible to create a RAID 1 consisting of a single disk with none of this potential, even though it is neither redundant nor an array.)

I don't personally know of an implementation that takes advantage of this mode of operation (in fact, ISTR discussion on the Linux md driver that specifically said that there were no plans to do so). Write performance would still be (best case) identical to that of a single head disk, but the potential gain for read performance is obvious.

And in the interest of history, plus a being at least orthogonally related to TFS: I have a few IBM 9ES 7200RPM SCSI drives on a shelf, still. I distinctly recall a bit of verbiage describing the availability of different firmwares for these drives, which would optimize the caching algorithm for differing numbers of threads: For example, firmware n.2 for two threads, n.4 for four, et cetera.

I never explored the idea of looking into it further because the closing sentence of "Please contact your IBM sales representative for more information" was rather off-putting, and sounded expensive. Besides, the drives were very able to keep up with every combination of seemingly-abusive realtime tasks I could shove at them when I was using them without any adjustment -- those fuckers were awesome back then. (Why fix it if it ain't broke? ;)

Re:Is that all? (1)

hawguy (1600213) | more than 3 years ago | (#36630768)

Say you have a video recording application where you're writing a video stream to disk, and that (perhaps uncompressed?) stream is of such ungodly bandwidth as to take a significant chunk of your drive's throughput. One head's fine if your disk isn't ridiculously fragmented (which it won't be); you have RAM to buffer it while the drive seeks occasionally (e.g. past a file fragment to the next unallocated space), then it'll catch up. But now suppose you want to playback a timeshifted stream of this same bandwidth

That said, in the interest of pedantry: Unlike a striped RAID 0, a RAID 1 array of n+1 disks could conceivably perform as a single disk with multiple heads, since a RAID 1 of n+1 has n+1 worth of independent head stacks, all reading identical data. (Also in the interest of pedantry, n is 1 or greater, since otherwise it is perfectly possible to create a RAID 1 consisting of a single disk with none of this potential, even though it is neither redundant nor an array.)

Except that that's not true in the case that the grandparent post described where you have a stream bring written at close to the maximum physical transfer rate of the disk, then you want to read that same stream from the beginning while you continue to write to it.

With two heads on the same disk, this should be possible, one head is busy writing, the other head is busy reading, with little seeking going on for either head and the full bandwidth of each head is available for each of the concurrent streams.

But if you're using 2 singled-head disks in a RAID-1 pair, this would not be possible, since the heads would have to seek as they switched between read and write mode and half the time the head would be reading, the other half of the time it would be writing, cutting the available bandwidth by half.

I think what you're talking about is the ability of the controller to read one piece of data from one disk in a RAID-1 set, and another piece of data from a different disk in that same set at the exact same time, using each disk as a "head". I think this behavior is present in most quality RAID-1 controllers, I'm surprised md doesn't do it too. Of course, this only works for reads, writes still have to go to both mirrored disks at the same time (well, a battery backed caching controller can defer writes for a time, so writes don't have to be flushed to both disks at the same time)

Re:Is that all? (1)

SharpFang (651121) | more than 3 years ago | (#36631048)

I'm afraid the limitation is cost, few people are willing to pay twice the price for same capacity...on obsolete technology. There were few CD-ROM drives that used multiple lasers, then DVD came in and the projects didn't return their own cost. So far the bus was always not fast enough to guarantee doubling the speed of the fastest drives.
If you want faster HDD, get SSD.

And as for home mods, 1) the precisions involved are out of reach of any non-professional, 2) just think about writing the firmware to run that sensibly...

Somewhere up to 80GB drives were openable and even had "rebreather holes". Currently they are filled with protective atmosphere and a membrane protects the inside from atmospheric air while retaining pressure. Opening a disk pretty much destroys it.

Hmm... (4, Informative)

screwzloos (1942336) | more than 3 years ago | (#36629440)

I don't get it. 500GB in an hour would be about 140MB per second (yes, I am rounding up). Most of the enterprise level 15K drives are right in that range without any overclocking, with a couple well above that. Do I win ten grand for buying a Seagate Cheetah 15K.7 for $450 and bringing it in to show that it works?

http://www.tomshardware.com/charts/enterprise-hard-drive-charts-2010/Throughput-Read-Average,2156.html [tomshardware.com]

No, I didn't look at the page. It's Slashdotted.

Re:Hmm... (3, Interesting)

mariushm (1022195) | more than 3 years ago | (#36629510)

It's about 132 MB/s actually - remember, it's multiples of 1000, not 1024 and then some space is used by the file system.

Anyway, it's not clear what they want just from the description here on Slashdot. Read the labels of the drive? But seriously, one could get a 2 TB drive or whatever drive has the most density these days and make it show up as 500GB drive... I believe it's called http://www.tomshardware.com/reviews/short-stroking-hdd,2157.html [tomshardware.com]

Re:Hmm... (1)

SirMasterboy (872152) | more than 3 years ago | (#36629730)

Re:Hmm... (1)

mariushm (1022195) | more than 3 years ago | (#36629976)

http://www.wdc.com/wdproducts/library/SpecSheet/ENG/2879-701277.pdf [wdc.com]

Formatted Capacity 500,107 MB = 488 GB so you need 138 MB/s to get data in an hour.

Re:Hmm... (1)

Lost Race (681080) | more than 3 years ago | (#36630156)

From that same document:

1. As used for storage capacity, one megabyte (MB) = one million bytes, one gigabyte (GB) = one billion bytes, and one terabyte (TB) = one trillion bytes.

So 500,107 MB = 500.107 GB

"Formatted capacity" has nothing to do with file system formatting; it refers to the host-accessible storage capacity of the drive, which is 976,773,168 sectors (also from that same document). The contest is to read all those sectors in under an hour. Sectors are 512 bytes each, so you need to read 500,107,862,016 bytes in 3600 seconds or an average of 138,918,851 bytes per second.

Re:Hmm... (1)

adolf (21054) | more than 3 years ago | (#36630710)

Sweet! After all these years, finally someone has come up with a reason for something better than UDMA 133 [wikipedia.org] !

Re:Hmm... (4, Informative)

a_nonamiss (743253) | more than 3 years ago | (#36629562)

Is that 15k RPM drive a "Western Digital Caviar Black 3.5" SATA 500GB hard drive (WD5002AALX)." It's stated pretty clearly in the rules that it needs to be that model. I don't think they're going for a speed test here, because there are plenty of SSDs that blow that speed away. They're trying to take a "normal" drive and super-speed it, for forensic purposes.

Re:Hmm... (1)

flappinbooger (574405) | more than 3 years ago | (#36629722)

So, upping the RPM obviously, but there must be various actuator settings in firmware that could be tweaked - safeguards, gain settings? What are the possibilities? I've never seen this done or even talked about, most people are "afraid" of hard drives, amazed they even work at all.

Re:Hmm... (0)

Anonymous Coward | more than 3 years ago | (#36630302)

A virtual "Western Digital Caviar Black 3.5" SATA 500GB hard drive (WD5002AALX) using DDR3 ram would certainly be faster than the non-virtual (once one populated it of course -- perhaps via tcpdump of an active internet oc-3 line) && once could certainly read the drive tag line in less than a minute.

Re:Hmm... (1)

Mitsoid (837831) | more than 3 years ago | (#36629586)

They specified a brand and model that had to be used (e.g. the net result would be an X% increase in speed).. I think they also limited the hardware modifications you can do.
So this is a test to make a hard drive 'over clock' and I believe they mean it in the sense like we do for CPU and memory -- Software/voltage/etc.. More cooling would be okay, but not disassembly of the hard drive

Re:Hmm... (1)

mysidia (191772) | more than 3 years ago | (#36629698)

More cooling would be okay, but not disassembly of the hard drive

You could replace the drive firmware with a hacked one that changes error detection behavior, changes the way the buffer/cache is used to optimize the drive for the contest's access pattern, or kills any power saving features.

The other thing would be changing characteristics of the drive's mounting to reduce vibration to insanely perfect vibration dampening for maximal mechanical performance.

Re:Hmm... (2)

Cramer (69040) | more than 3 years ago | (#36629932)

In all modern IDE/SATA drives, the firmware is stored on the plater, not in an eeprom. And for most manufacturers, it's not field accessable. Plus there's zero documentation for the firmware / internal processor(s) outside of the manufacturer's labs. (and maybe the company making the chips.) Hacking the firmware is beyond the reach of anyone who would be wowed by a $10k prize.

Re:Hmm... (0)

Anonymous Coward | more than 3 years ago | (#36630008)

Wasn't this the cause of the infamous Seagate 7200.11 issue?

One of the possible fixes for that issue was to connect to the serial port on the drive and issue commands directly to the processor.

ISTR one of the suggested tools would let you dump and overwrite firmware using that same connection.

Re:Hmm... (1)

Cramer (69040) | more than 3 years ago | (#36630138)

Seagate was one of the few that would field update a drive -- through a special mode program sent to the drive... I'm aware of console ports on tape drives but not any hard drive. (esp. cheap retail drives.)

Re:Hmm... (0)

Anonymous Coward | more than 3 years ago | (#36630750)

True, but there's prestige, too.

If I was an engineer at WD (and therefore had access to the specs and firmware source), I'd definitely try it in private, and push for permission to claim the prize & bragging rights (without undue disclosure of IP, obviously). IMO it would be good publicity for the company, of course if I worked at WD, my boss would probably be the kind of asshole to forbid it.

Re:Hmm... (1)

slinches (1540051) | more than 3 years ago | (#36629980)

Does it have to be a spinning platter drive? If not, some of the PCI-E SSDs can get over 1GB/s sequential reads which would easily put a 500GB read at under 10 minutes. Of course, you'd likely have to spend at least half of the $10k prize on the drive itself.

Re:Hmm... (1)

b0r1s (170449) | more than 3 years ago | (#36630860)

Meh. 15k drives, multi drive arrays, there's no reason to overclock CPUs anymore, and there's no reason to overclock hard drives. More cores scales better than overclocking, and more spindles scales WAY better than overclocking.

Orgrimmar:DATA admin$ df -h | grep disk7
/dev/disk7 18Ti 238Gi 18Ti 2% /Volumes/DATA
Orgrimmar:DATA admin$ date ; dd if=/dev/zero of=test.bin bs=16k count=10240000 && du -sh test.bin && dd if=test.bin of=/dev/null bs=32m ; date
Thu Jun 30 22:47:45 PDT 2011
10240000+0 records in
10240000+0 records out
167772160000 bytes transferred in 240.477549 secs (697662466 bytes/sec)
156G test.bin
5000+0 records in
5000+0 records out
167772160000 bytes transferred in 139.259221 secs (1204747226 bytes/sec)
Thu Jun 30 22:54:06 PDT 2011



Server has 8G ram, so some of that may be in cache, but certainly not all of it. Extrapolate as needed.

Re:Hmm... (0)

Anonymous Coward | more than 3 years ago | (#36630954)

Unless you have a sparse filesystem. http://en.wikipedia.org/wiki/Sparse_file

Try with something that isn't zeros :P

Done already! (5, Funny)

allanw (842185) | more than 3 years ago | (#36629452)

Re:Done already! (1)

ooshna (1654125) | more than 3 years ago | (#36630428)

You sir have the hair of Ariel and the muscles of Schwarzenegger for linking to that awesome site.

Re:Done already! (1)

arth1 (260657) | more than 3 years ago | (#36630532)

You're sure it's not the hair of Schwarzenegger and muscles of Ariel?

Fizzzzzzttt (0)

Anonymous Coward | more than 3 years ago | (#36629480)

Oh dear, hosting 500Gb drive dead already, I guess they not getting $10K... :-)

Re:Fizzzzzzttt (4, Funny)

blair1q (305137) | more than 3 years ago | (#36629590)

Maybe that's what they mean.

"If you can get anything off our 500 GB drive in the next hour, we'll give you $10K."

Seems a bit silly (1)

Anonymous Coward | more than 3 years ago | (#36629570)

If any hard drive is allowed, then there's nothing to stop you using a 15K RPM drive.

If any format is allowed, then have one sector per track (so there's one seek per track and you make absolutely maximum use of buffers, minimizing latency). Structured disks are rather unnecessary with smart controllers and huge memory sizes. Read the whole track, modify in memory, write the whole track. Back when disks were actually much larger than machine memories, it made sense to have formatted disks. These days, you can buy flash that's comparable in size that could fit on a controller easily and a single bank of RAM is more than sufficient to handle track-at-once operations on individual drives.

If any controller is allowed, roll your own. You want the lowest-level operations offloaded from the CPU (which is slow) and you want the buffers written direct to memory rather than via the kernel (so you want a bus controller on there).

In short, there's a multitude of ways of speeding disks up, and some of them are actually interesting.

Re:Seems a bit silly (1)

FlyingGuy (989135) | more than 3 years ago | (#36629632)

I think they are referring to things like we used to do. change a crystal, perhaps a cap or two and add a bigger heat sink, not re-engineer the thing.

Solution (1)

gatkinso (15975) | more than 3 years ago | (#36629746)

Post a bounty on slashdot, watch your drive fry.

SSD? (0)

Anonymous Coward | more than 3 years ago | (#36629960)

I haven't been able to get to the linked article but why not just use an SSD? Specs of a Crucial 512GB SSD:

512GB Crucial m4 2.5" SSD with Data Transfer Kit

Part Number: CT512M4SSD2CCA
The fastest SSD on the planet.
Groundbreaking SATA SSD performance
Read speeds up to 415MB/s
Second-generation SATA 6Gb/s w/ Native Command Queuing (3Gb/s backward compatible)

Re:SSD? (1)

jones_supa (887896) | more than 3 years ago | (#36631260)

Because that would be like putting a sports car in a tractor competition.

do not do this (4, Insightful)

JonySuede (1908576) | more than 3 years ago | (#36630020)

This is an attempt by a forensic company to crowd-source the development of a product on the cheap. I you can do this, you can make a fortune selling to the different LEAs around the world. But please don't do it, we do not need more efficient spooks.

Re:do not do this (0)

Anonymous Coward | more than 3 years ago | (#36630136)

Any modification that's performed on the drive would invalidate the evidence extracted from it in court. Tampering with the disk is verboten. Any good lawyer would know that (IANL).

This will allow them to go on a faster phishing expedition when they do things like seize the laptops of travellers. Not evidence in court, but enough to evict you from the country perhaps.

Re:do not do this (0)

Anonymous Coward | more than 3 years ago | (#36630680)

I agree. This probably will be used at the airports. Good money making scheme. Do not sell yourself for so cheap. 10K for a forensic breakthrough is peanuts.

Re:do not do this (1)

Paska (801395) | more than 3 years ago | (#36631402)

This is another hidden benefit of Apple hardware that people don't readily consider.

Apple hardware is very hard to get in and out quickly, covertly, and without a few red flags being noticeable.

A couple of years (4+ now) ago when I sat in with Apple's instructor led hardware certification labs there were a small team of high tech crime investigators for the Australian Federal Police, and Australian Attorney Generals department attending.

They weren't interesting in passing the test, they had absolutely *no* interest in OS X, and they hated Apple hardware with a passion - why? Because it was damn hard to get in and out to get access to the hard drive. The '06 model iMacs (which at the time, was the current model) caused this guys to be furious with Apple, given you had to rip/tear the heat shield to get to the hard drive.

Being a predominate Apple technician at the time in Canberra, and given my reputation as being half decent I stayed in contact with a few of these departments, and still do to this day.

Even a little under a year ago I was (2010) I was personally requested to attend on-site to an unnamed department, to get a hard drive out and back in. What was normally a 45 minute on-site, turned in a *2 day* on-site as the team I was with (I was heavily supervised) had never worked on an Mac before, the matter was clearly urgent, and after removing every screw, the LCD panel, every cable, photos were taken and I was providing on-site training as to what every step I was taking.

Placing the iMac back to gather took the longest as I was stopped every few moments, asked to leave the room and then asked to come back into the room to continue. I suspect they were comparing the pull-apart photos to how I was placing the machine back together to ensure the machine was being placed together exactly as it was before.

While I never asked questions, a few months later the Australian Federal Police busted one of Australian's largest children pornography rings in Australia. I have the niggling feeling I was somewhat involved in this operation.

Disclaimer: I am now an owner of an Canberra, Australian based Apple Mac managed services business.

Smells like marketing... (1)

znerk (1162519) | more than 3 years ago | (#36630174)

Has anyone considered that this is simply a high-profile device for:

A: Selling amazing amounts of the specified WD 500GB HDD?

B: Giving WD some free (or nearly so, $10k is pennies in the pot) development of a product line, via a third-party agent?

Smells like marketing... for the university itself (0)

Anonymous Coward | more than 3 years ago | (#36630220)

No. It's more likely the Edith Cowan University "Security Research Centre" advertising itself. It's cheaper to pay a $10,000 bounty, than to buy advertising space, in order to put the university's name in front of prospective students.

Re:Smells like marketing...+++ (0)

Anonymous Coward | more than 3 years ago | (#36630544)

Maybe someone knows that HDs are purposely crippled in performance because:

1) they need to last past their warranty
2) overheating
3) to segment the market and offer premium products even though technology employed is the same

Re:Smells like marketing... (0)

Anonymous Coward | more than 3 years ago | (#36630896)

Sometimes Even giving it free, is not enough when... "Educated" people regarding business leadership gets their hands on the stuff... "your saying were getting 10% material savings but ONE added part... can't do! should reform whole production line.."

theres stuff even where patents have expired and still not used widely, despite superiority.
Also in sport's large manufacturers lobby even the olympic rules to keep out new inventions...

Ideas for success (1)

Anonymous Coward | more than 3 years ago | (#36630358)

Posted under anon cause I'm at work

Theoretically
-There are dc electric motors (brushed and brushless) that can run at 30k+ rpm
-The electronics on the drive are running on a clock (you'd assume)
-There is some firmware that governs the parameters of the hdd
-some large forklift motors @ 36VDC can run on 120vDC - meaning that you can in theory put more power in without frying
-Electronic voltage constraints are due to heat/power dissipation.

Quick and dirty Thoughts:
1. Chips should be able to handle data being fed at a quicker rate
2. Remove logic board from drive to help cooling both logic board and motor
3. Measure voltage across the terminals of the motor in use, wire it up to a linear regulator that outputs 20% greater voltage.
4. Main concerns will be heat of the motor (heatsink case, blah)
5. Pray that the stepper motor in the arm and the head can handle the higher spindle rate

Way out there ideas:
1. Replace motor with 30000+ rpm motor
2. Replace hardware in drive with 15k rpm hardware - doubt this will work.

links broken in comments? (0)

Anonymous Coward | more than 3 years ago | (#36630488)

seems once again, the latest slashcode update has broken some fundamental stuff... anyone else seeing the site scripts breaking links inside comments? the links in the story summaries are fine. actually, maybe it's all mouse clicks in the comment area are broken. can't even right click within the border of a comment. wtf guys...!?

Re:links broken in comments? (1)

farseeker (2134818) | more than 3 years ago | (#36630758)

Double-right-click and the menu shows up in FF5

linear array of reading heads (0)

Anonymous Coward | more than 3 years ago | (#36630706)

I've been wondering for awhile now why there still is a mechanical arm with a single reading head. Why not just span a linear array of reading heads over the disk, one for each track.. you'd be able to read out an entire disk in one rotation. that'd be some impressive speed. it'd kick ssd's ass.

my-lptp:~ fred$ ls -l /dev/sda1 (0)

Anonymous Coward | more than 3 years ago | (#36630914)

OK I didn't descend into directories. ...What?

You wanted me to write the data somewhere?

IRONY (2)

roman_mir (125474) | more than 3 years ago | (#36631222)

What is ironic is that this story precedes the one, that gives the actual reason for this one [slashdot.org] .

It's not too saddle, is it:

'Federal Wiretaps On The Rise'
'Hard Drive Overclocking Competition from Secau'

Re:IRONY (1)

roman_mir (125474) | more than 3 years ago | (#36631234)

subtle, I meant subtle. I didn't have a good night rest and am still slippy....

Vary temperature and pressure (1)

sonamchauhan (587356) | more than 3 years ago | (#36631408)

Hmmm... HDD have atmosphere seals.
http://www.pcreview.co.uk/forums/hard-drives-hermetically-sealed-t2014655.html [pcreview.co.uk]

  wonder if varying temperature and pressure will help.

E.g. doing this in a cold, low pressure environment.
Or a cold, high-pressure environment.

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>