Other 10k RPM SATA Drives

R

Relentless3O

Limp Gawd
Joined
Dec 9, 2005
Messages
129
What other 10k RPM SATA HDDs are there out there? Ive only heard one name. Western Digital Raptor. Are there any other reliable 10k HDDs out there?
 
You'd think the other manufacturers would have provided some competition by now, but WD is continually kicking ass as usual. They have been ever since they started the JB Caviars IMO :)
 
movax said:
You'd think the other manufacturers would have provided some competition by now, but WD is continually kicking ass as usual. They have been ever since they started the JB Caviars IMO :)
Click to expand...
Competition against their own 10K SCSI lines?
Not likely ;)
 
Vertigo Acid said:
Competition against their own 10K SCSI lines?
Not likely ;)
Click to expand...

Bingo.
The 10k Raptor was (and I guess still is to some degree) supposed to be an enterprise product. WDC didn't have a SCSI line when they premired the raptor, and the rest is history.

Edit: Spelling
 
Yeah, forgot that WD doesn't have a SCSI line ><. I <3 Raptors.
 
yeah too bad no one has any extra 74gigs or 36gigers to sell to me :( PST if you have one! My hdd failed and I cant use my main comp. Will pay 90ish shipped for raptor
 
Post a "WTB: Raptor" Thread in FS/FT, and you'll have one faster than you say...HardOCP? :p
 
You're not looking hard enough. this thread has 36GB Raptors for $80 each.
*EDIT* Oops, you're looking for a 74GB drive...try here.
 
Incorrect. On the desktop, buffer size/strategy is #1, followed by seek performance, then density (intensifies locality) then transfer performance.

Clicky

Seek performance is #1 in game loading, as games tend to be very poorly coded and be unnecessarily seek heavy. However, buffer size/strategy is still a big factor.

Also, increasing spindle speed only reduces rotational latency. In this case, that means 1ms less average rotational latency, as a 10K spindle has RL of 3.0ms, while 15K has 2.0ms. WD probably could get that 1ms just by putting a 15K grade actuator on the Raptor and reducing seek time, without taking the environmental hit of a 15K spindle.
 
WD probably could get that 1ms just by putting a 15K grade actuator on the Raptor and reducing seek time, without taking the environmental hit of a 15K spindle.
Click to expand...
Err, seek latency is actuator latency plus rotational latency. No matter how fast your actuator, you still have to wait for the disk sector to move underneath the head.
 
But if you get the actuator there 1ms faster, you reduce seek time, as 1/6 of the sectors requested have below average latency. If you knock seek time down from 5ms to 4ms, that's 1/6 of the seeks that just pass underneath the read/write heads, and is read immediately, instead of just missing and having to wait 5ms plus.
 
But in this case, the actuators would be the same speed. You were saying WD could meet the performance of a 15K disk simply by adding that disks actuator.

A 10K disk with a "15K-grade actuator" cannot be as fast as a 15K disk with that same actuator.
 
No, but a 10K disk with a 15K grade actuator would achieve about the same seek performance as a 15K drive with a 10K grade actuator.
 
Ok, if your point was that upgrading the actuator would give more bang than the spindle speed, then I agree.
 
This has been a key reason why consumer drives have stayed at 7200RPM. It has been more cost effective for the makers to simply make the drive's bigger (therefore easier to market) and as a side effect, intensify locality (seeks over a smaller area of the disk) increasing application level performance, without having to add louder and more power intensive, but faster, actuators or spindle motors.

Also, if you reduce seek time 1.2ms, you achieve the same benefit as going from 7200 to 10K.
 
Also, if you reduce seek time 1.2ms, you achieve the same benefit as going from 7200 to 10K.
Click to expand...
The same benefits on seek times...but not bandwidth.

I wonder if drive makers have ever looked at adding a second actuator to each platter, mounted 180 degrees opposite. That would halve rotational latency as well as doubling transfer rate-- it should allow a 10K drive to perform like a 20K. Expensive...but at some point its got to be cheaper than upping spindle speeds still further.
 
I have heard it argued both ways. This article explains how, after a few months or years of PC use your drive becomes fragmented, even after running defrag programs and whatnot. In doing so, seek times become more and more important as the drive has to seek through more data to find what it is looking for. We all know how hard drives get slower as more and more stuff is put on them. 10K's experience this less, while 15k's hardly experience this at all.
 
masher said:
The same benefits on seek times...but not bandwidth.

I wonder if drive makers have ever looked at adding a second actuator to each platter, mounted 180 degrees opposite. That would halve rotational latency as well as doubling transfer rate-- it should allow a 10K drive to perform like a 20K. Expensive...but at some point its got to be cheaper than upping spindle speeds still further.
Click to expand...
If you run a raid 1 array on a good controller, it'll effectively do this (on reads, anyways...).

 
unhappy_mage said:
If you run a raid 1 array on a good controller, it'll effectively do this (on reads, anyways...).
Click to expand...
Right...too bad it kills the write performance though :)
 
You must log in or register to reply here.
Back
Top