Fast mechanical HDD with clever Partitioning.

Roland0 · 24 Nov 2009 at 16:28

Hello peeps,

I remember seeing a post that explained that you could do something along the lines of partition a disk. So that only the very edge of the drive was used, in order to get very fast read/write speeds, and reduced access times obviously at the cost of space. However it wasn't just partitioning, as it needed special software from the HDD vendor. However I can't find or remeber anything aout it or how its done

I would be getting an SSD but since the decent ones are the best part of £200 and the single platter 500GB drives are less that £40 its seems like im bets of waiting for SSDs to come down in price.

Cheers

ChileanLlama · 24 Nov 2009 at 16:32

You're looking for "short stroking". I've never done it myself, so can't offer you more than that

AbsenceJam · 24 Nov 2009 at 16:33

Just partition, they're created from the outer edge so first partition is fastest. You can use most manufacturer disks to limit the size of the drive but why bother when partitioning is exactly the same. And if you really don't want to use all the space then don't partition the unsued space.

elpedro · 24 Nov 2009 at 16:34

Its called Short Stroking, I've done it with my 250GB Barracuda using SeaTools from Seagate.

It does indeed reduce access time and you get the maximum read/write speeds with the outside of the platter being the fastest moving point.

http://www.tomshardware.com/reviews/short-stroking-hdd,2157.html#xtor=RSS-182

PiKe · 24 Nov 2009 at 16:39

AbsenceJam said:
Just partition, they're created from the outer edge so first partition is fastest. You can use most manufacturer disks to limit the size of the drive but why bother when partitioning is exactly the same. And if you really don't want to use all the space then don't partition the unsued space.

This doesn't produce the same result.

miniyazz · 24 Nov 2009 at 16:48

PiKe said:
This doesn't produce the same result.

Do please explain how you would go about achieving the result then?

elpedro · 24 Nov 2009 at 16:49

PiKe said:
This doesn't produce the same result.

Indeed, the rest of the disk space, although unpartitioned, still 'exists'. Short stroking makes the disk appear to be what ever capacity you set it as, the unused space cannot be detected nor used.

AbsenceJam · 24 Nov 2009 at 17:04

PiKe said:
This doesn't produce the same result.

elpedro said:
Indeed, the rest of the disk space, although unpartitioned, still 'exists'. Short stroking makes the disk appear to be what ever capacity you set it as, the unused space cannot be detected nor used.

No it produces the EXACT same results whether you partition the rest of the disk or not. Whether your computer/OS can see part of the drive or not doesn't make a difference to how it operates.

miniyazz · 24 Nov 2009 at 17:07

elpedro said:
Indeed, the rest of the disk space, although unpartitioned, still 'exists'. Short stroking makes the disk appear to be what ever capacity you set it as, the unused space cannot be detected nor used.

What, you mean like partitioning, or something? :rolleyes:

AbsenceJam said:
No it produces the EXACT same results whether you partition the rest of the disk or not. Whether your computer/OS can see part of the drive or not doesn't make a difference to how it operates.

This

Roland0 · 24 Nov 2009 at 20:25

Hello again,

Do we have any objective evidence that short stroking and partitioning are the same or different in performance? i can't really see how partitioning can be all that different, unless somehow the controller not knowing that the majority of the disc isnt usable speeds it up.

elpedro · 24 Nov 2009 at 20:59

AbsenceJam said:
No it produces the EXACT same results whether you partition the rest of the disk or not. Whether your computer/OS can see part of the drive or not doesn't make a difference to how it operates.

Read/write will be the same, be slight difference in access times as the actuator will still move the full width of the platter if the unpartitioned area is still seen, whereas if it's short stroked, it doesn't know that part exists and remains soley on the out part of the platter.

Technically there will be a difference, I highly doubt anyone would ever be able to notice it. But yes, the read/write speed will be the same.

icysniper · 24 Nov 2009 at 21:07

this may help and uses a 1.5TB Seagate

http://www.techwarelabs.com/seagate_1-5tb-mod/all/1/

AbsenceJam · 24 Nov 2009 at 21:14

elpedro said:
Read/write will be the same, be slight difference in access times as the actuator will still move the full width of the platter if the unpartitioned area is still seen, whereas if it's short stroked, it doesn't know that part exists and remains soley on the out part of the platter.

Technically there will be a difference, I highly doubt anyone would ever be able to notice it. But yes, the read/write speed will be the same.

Changing the max address does nothing except change how much capacity can be addressed, it does nothing to the mechanics or anything like that. The actuator works exactly the same as ever, only moving to where it needs to. Why do you think it would move across the full width to access somthing on the outside? It wouldn't....

AbsenceJam · 24 Nov 2009 at 21:20

Stoned_SpaZZZ said:
Hello again,

Do we have any objective evidence that short stroking and partitioning are the same or different in performance? i can't really see how partitioning can be all that different, unless somehow the controller not knowing that the majority of the disc isnt usable speeds it up.

Using ATTO which will benchmark partitions rather than whole disks I confirmed this earlier after the wrong replies to your post but forgot to keep the caps. Limited one disk to 17GB (just tapped in a random no.) with the Samsung utility and benched it, then removed the address limitation and did the same with a 17GB partition. Same results. There's no app that tests access time for a single partition/drive letter afaik so couldn't do that. Anyone can do the same themselves with ease if they want to see proof.