10000rpm ATA?

[Gnerma]

Where are these drives?

I havn't even heard a whisper about them, do you guys think its possible/probably with the size of ATA drives? What do you think is holding the HDD manufacturers back from producing these?

 

[Spec_Ops2087]

I don't think there are ones because ATA would hold it back...that is why there is SCSI...so the RPMs are the bottleneck not the way the drive transfers data....and the manufactors keep trying to make them bigger and BIGGER but don't realize they are producing the slowest product in the computer...




Spec

 

[Jkasmann]

There probably never will be, though not because ATA holds it back. 10k drives dont necessarally transfer data faster, they just have better seek times (find data faster). Once a block of data is found, the rotational speed isnt that important. The fast seek times help more in some areas than others....for a lot of what the average person does it isnt that helpful.

You wont see ATA-10k drives any time soon because there isnt any reason for them, they are quite a bit more expensive and the people that can afford them would probably prefer to pay more for scsi also (since it is definetly superior to ATA).

 

[NookieN]

There are technical limitations due to the size of current ATA drives. When you're talking 20-30-40 Gbytes per platter, the individual bits are so small that they cannot be reliably read at 10k+ in a cheap drive.

However, bigger platter sizes can give better transfer rates. At 7200 rpms, more bits move under the heads each second on 30Gbyte platters than on 20Gbyte platters.

 

[Slogan]

i actually saw 10,000 for sale in a magazine... it was some magazine i was flippin through at school, and this was a year ago.. but VERY expensive

 

[6502kid]

I bet they run way hot.

Remember IBM's little problem with those 60gb GXP60's ?
I think I could fry an egg on mine. (RMA'ed 3 times so far....)

 

[cursor]

What about independent arms for each side of the platter like SCSI? Wouldn't that make a bigger difference, especially on non-sequential reads and writes? 3.9 ms access time

 

[64026402]

10k drives are coming for ATA. Because of the higher densities the 7200s are as fast as all but the most expensive SCSI hard drives. The SCSIs need 15k to get the transfer rates up.
You'll notice the larger new 10k drives coming out are pushing more data than the smaller 15k drives.
10k will probably push the ata 100 and 133 pretty hard so the serial ATA will be out before the 10k IDE.

 

[WildArmSX]

you can possible for see a 10k prm drives for the new IDE SATA standard but i highly doubt yoy will ever see it come out for IDE ATA standard

 

[64026402]

you can possible for see a 10k prm drives for the new IDE SATA standard but i highly doubt yoy will ever see it come out for IDE ATA standard

Probably right unless they make the 10k drives in both ATA and SATA to boost sales with the installed base of IDE owners.

 

[seamadan000]

SCSI in later generations is still way faster than ATA, no matter what rotational speed, look at the bench marks at storagereview.com
________
SHIP SALE

 

[Gnerma]

Okay, so what have we learned so far? Eh not much.. it seems like nobody has reliably heard of 10000 RPM ATA drives in development or any possible time we might see time, if at all.

Also there is also some doubt that its technically possible within the ATA 'construction' and doubt it would help all that much at all.

The questions that just begs to be asked is what's next then? If ATA drives aren't going to get faster/better how will they continue to increase their transfer rates / seek times? Will the HDD companies keep finding ways to jerry rig (increased cache etc) 7200 rpm drives to be a bit faster than the last generation? Will they increase performance by just increasing data density? Will we all just move up to SCSI?

I'm going to do some research on this, I'll let you know what I find out.

 

[donny_paycheck]

ATA will move to 10k sometime, it'll just be a wait. SCSI was at 7200 while ATA was just dreaming of 5400, so SCSI offers a multi-year preview into what is to come for ATA.

What really sucks is that SCSI = bad for overclocking. But that's why we have the WD SE drives with the 8mb cache...

 

[lonewolf1983]

id expect they are holding back speeds a little until pricing can get down. One of the reasons scsi is so expensive is due to the better construction and engineering that goes into building a reliable high RPM drive.

anyone wonder why we are seeing 1 year warranties on 7200rpm hard disks now??? my guess is because the cheap construction of drives now isnt really up to the spindle speed and prolonged use.

my $0.02

 

[PYROMANIAC]

to see how 7200 compares with 5400rpm i benched my 80gig ata133 7200rpm maxtor against my 160gig 5400rpm ata133 maxtor and as follows:

80gig: average transfer rate = 57.2mbps

160gig: average transfer rate = 58.3mbps

then for fun i benched my old ata100 7200rpm 60gig: average transfer rate = 76.3mbps

edit: the point, i see no real dif between 7200/5400 rpm..its not really that noticable and i'd rather have more gigs than faster speeds by 1mbps

 

[FTC]

Pyromaniac,

The transfer rate depends much more on the 'density' (bits per inch) of the disk than on the rotational speed. The density of the data in the disks doubles every 2 years (roughly), and all rotational speed has done in the last 18 years (not counting SCSI) is going from 3600 to 7200.... so, when you compare transfer speeds you will have to make sure the compared drives have the same density to be able to say you are comparing 'apples' to 'apples'.

Also, with the data you are giving you are probably measuring cached data transfers or burst transfer speed, which is even less related to media transfer speed. (unless yours is a RAID setup).

Where you will also notice the difference between 7200 and 5400 or lower is in the 'latency', i.e time it passes to *begin* your data transfer, which is composed of the seek time (time to move the head) plus the latency (time before the desired sector is under the read/write head). For 7200 disks, the mean latency is of about 4 ms, and for 5400 disks it is of about 5.5 ms. This may seem a small difference, but adds up..

So, in summary : If you compare equal technology disks (same density and seek times), then a 7200 disk is about 10-20% faster than a 5400 disk *finding* data, and has the potential to transfer data at about a 35% faster speed from the disk platter to the disk controller.... and belive me, this difference is *quite* noticeable in the real life and/or if you do the proper tests.

Regards
FTC

 

[donny_paycheck]

Well said, FTC. Thanks for a great explanation that I was about to attempt!

 

[PYROMANIAC]

i used HD tach so it did burst read, oh well.. the oldest drive was the 60, then the 80, then the 160..how would the 80 be slower than the 60 when its a lot newer?

 

[FTC]

Pyromaniac,

Burst read speed is not *that* much relevant, and depends basically on your controller (IDE chipset), cable AND disk controller board, not at all on the rotational speed or seek time, so 'how old is the disk technology' will not matter, as long as comparing disks from the last 2-5 years. Also depending on the disk cache size and speed, you can find differences in the order you showed us.

It would be a lot better test that you give us atto figures, or even Sandra drive figures (but of course, with an empty unfragmented disk).

Regards
FTC