Are velociraptors the fastest

It depends.

Many SAS and SCSI drives are theoretically much faster, but their firmware tends to be optimised for server usage patterns, so there might be little or no advantage in a desktop PC.
 
Other things being equal, the faster-spinning a drive, the lower the seek times. So I would expect a 15k SAS/SCSI drive to beat out a velociraptor, regardless of firmware, but different firmware optimisations could make a bigger difference than I am allowing for..
 
Other things being equal, the faster-spinning a drive, the lower the seek times. So I would expect a 15k SAS/SCSI drive to beat out a velociraptor, regardless of firmware, but different firmware optimisations could make a bigger difference than I am allowing for..


Does anyone here actually have one of those Seagate 15k RPM drives though?
 
Just bought a 500GB Seagate for use as a boot drive, going to apply this to it. Sure I'm only going to get around 150GB of the 500GB storage, but that 150GB is faster than the 150GB Velociraptor and much much cheaper

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

saw aruffell's post on the mod in the other thread, i picked up the same 1.5tb seagate the other day but dont have the heart to reduce it 5 times the amount for the extra speed :o

would be cool to see a user review once you done it though :)
 
wow that link is really interesting..
SO how does it actually speed up the drive? does it jsut use the faster part o teh drive? or is there something else to it?

Have you tried doing different comparisons. Down to like 50GB? 2 in raid and you could potential get vertex speeds but many many times the capacity.

Can you do this with any drives or just that one?
 
I wasn't thinking of SAS/SCSI drives, my mistake. A lot of them would indeed be faster than a Velociraptor, however IIRC they're quite expensive for what storage space you get, as well as needing an adapter card.

The Velociraptor is without a doubt the fastest of the SATA/IDE drives, bar SSD.
 
wow that link is really interesting..
SO how does it actually speed up the drive? does it jsut use the faster part o teh drive? or is there something else to it?

Have you tried doing different comparisons. Down to like 50GB? 2 in raid and you could potential get vertex speeds but many many times the capacity.

Can you do this with any drives or just that one?

I originally read about short stroking in an article using WD caviar drives with some software WD provide. That article I posted uses the freeware SeaTools from Seagate.

Indeed, you utilise the the outer part of the platter (the fastest rotating part) which also shortens access times as there is less area to be covered. According to that article, the Veliciraptors still have 3.3ms over the Seagate's access times even when using this technique, but hey, for the price/performance I really couldn't care.

I am seriously considering buying two 1GB drives, short stroking them to 300GB each and then setting them up in RAID 0!

I will have the drive I ordered by Tuesday and let you know how I get on. Currently running everything off my RAID 0 250GB drives, I wanted a separate boot drive as this should improve general system performance by keeping Windows separate.
 
Last edited:
Any chance you can do like 5 different Size tests? so we can see a comparison and if there is a sweet spot. All thought that would take a while, so understand if you can't.
 
Any chance you can do like 5 different Size tests? so we can see a comparison and if there is a sweet spot. All thought that would take a while, so understand if you can't.

The max transfer speed will be the same regardless of the restricted capacity. It's as you move closer to the centre of the platter the read/write speeds drop. By limiting the size, you maintain higher speeds and therefore a greater average overall. Check the graph from HD Tune below to see what I mean.

2ljkgab.png


In theory, the smaller you stroke the drive, the better the overall performance, but the performance cannot exceed the maximum read/write speed. You just gain quicker access times with less area to cover.

Ill shove the drive in before I install windows onto and do a few tests at various sizes though ;)
 
I dont get this stroking business. Surely if you just partition your drive correctly and have Windows sitting on the first partition, at the fastest part of the disk, everything you do on that partition will be fast. Its only in benchmarks (averages) that this mod would actually make a difference surely.
 
Does anyone here actually have one of those Seagate 15k RPM drives though?

Someone was playing around with half a dozen in raid0 a few months ago on here. on my phone atm though so can't be bothered go dig up the thread, but I'm sure it'll turn up with a search.

I dont get this stroking business. Surely if you just partition your drive correctly and have Windows sitting on the first partition, at the fastest part of the disk, everything you do on that partition will be fast. Its only in benchmarks (averages) that this mod would actually make a difference surely.

I would think so.. but you'd have to leave the rest of the drive unpartitioned/unused.
 
Back
Top Bottom