Hack your barracuda to be quicker than a velociraptor

[Elixer]

File this hack under Hmmmmmmmm.

In short, you hack the 1.5TB HD into a 300GB HD, and then run benchmarks on it, and then say it is faster than a velociraptor, in everything but access times.

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

 

[rcpratt]

Oh dear.

 

[imported_Scoop]

That is pretty cool. The point is to save money to have the performance of the Velociraptor. You don't have 1.5TB Velociraptor drives do you?

 

[TemjinGold]

How's that any better than just partitioning 300/1200? All you are doing is keeping the 300 partition and pissing away the rest.

 

[brblx]

1tb+ drives already read/write at near-velociraptor rates. the access time is what matters.

silly 'mod.'

 

[rcpratt]

Originally posted by: brblx
1tb+ drives already read/write at near-velociraptor rates. the access time is what matters.

silly 'mod.'

ding ding ding

 

[Fox5]

Does this do anything? They could have at least shown before and after results. I don't recall raptors putting out very impressive read/write speeds anyway, it was all about access times.

 

[Ben90]

Because all drives are structured like disks, the data located towards the end of the drive (towards the edges), is more separated than the data located near the center, hence it takes less time to locate, read and write information to and from the drive. Since the drive is 1500 GB in size, 300 GB ends up being 20%. By setting the size to 20 % of the max, the drive uses the inside 20% of the disc thus bringing superior performance.

THATS IT!! we just need to "long-stroke" our hard drives to beat the VR.

 

[F1shF4t]

Using same methodology I can "prove" that my laptop 500gig 5400rpm Seagate is faster than the gen 1 74gig raptor I had :Q

Doen't make it any more less of a joke. Sequantial read/write speeds are not everything.

 

[TC91]

Originally posted by: Ben90

THATS IT!! we just need to "long-stroke" our hard drives to beat the VR.

It's actually called short stroking. What you are doing is just using the fastest part of the drive, and since benches usually "average out" the entire drive including the slowest parts of the drive (inner part of the platters) which lowers the averaged out benchmark result.

Short stroking won't make your drive go any faster, it just keeps you from using the slower parts of the hard drive.

 

[taltamir]

Originally posted by: TemjinGold
How's that any better than just partitioning 300/1200? All you are doing is keeping the 300 partition and pissing away the rest.

because then you are still using the entire drive, and your heads still go to the "data" area often.

Although, as long as you are not using the 1200GB partition, disabled indexing, defrag, etc.. then you are right.

 

[Ben90]

Originally posted by: TC91

Originally posted by: Ben90

THATS IT!! we just need to "long-stroke" our hard drives to beat the VR.

It's actually called short stroking. What you are doing is just using the fastest part of the drive, and since benches usually "average out" the entire drive including the slowest parts of the drive (inner part of the platters) which lowers the averaged out benchmark result.

Short stroking won't make your drive go any faster, it just keeps you from using the slower parts of the hard drive.

yea, im aware what he was doing in the article was short stroking. He clearly doesnt know how a hard drive works though.

I was quoting him "the drive uses the inside 20% of the disc thus bringing superior performance."

 

[Emulex]

wow thats rad; and its still slower than a decent ssd

 

[Rubycon]

Very misleading! Short stroking does nothing to increase real world performance. Just makes benchmarks look good.

I tried this with 10 Fujitsu MBA147RC 15K SAS disks in RAID0. Made a 73GB logical drive (discarding the remaining 9 1/2 disks worth of space!). This was the result:

Results