Would you rather have a faster sustained transfer rate or lower access seek time?

Soldato
Joined
27 Jan 2005
Posts
3,655
Location
Expat in HK
A generic 74GB Raptor with an average 65MB sustained transfer rate should have quicker access times than all SATA/PATA with <10,000RPM. What I was wondering though is does anyone think the overall performance of a SATA/PATA can make up for the lack of quicker access times by having a greater sustained transfer rate?

If so, then how much greater than 65MB do you think warrants such a compensation.

They are two completely different aspects of a drives ability, but i'm looking for personal views on the matter.
 
Last edited:
Personally I'm more interested in transfer rates. I don't reboot that often so startup times aren't that important rather I'm more interested in things like thumbnailing large numbers of photos (25Mb TIFFs) which need long sustained reads.
 
I prefer transfer rate, a good defragged hdd wont seek that much and most games there days have everything in one huge .dat file (the games that i play anyhow), doesnt need much seeking, so i prefer read times...
 
This is the current deliberation as I'm reformatting. It's in RAID 0 now, which gives slightly decreased seek times (13 compared to 9ms for example) but much greater sustained/burst speeds (think 316 MB/s burst and sustained about 100). Leaning towards raid atm.
 
joeyjojo said:
This is the current deliberation as I'm reformatting. It's in RAID 0 now, which gives slightly decreased seek times (13 compared to 9ms for example) but much greater sustained/burst speeds (think 316 MB/s burst and sustained about 100). Leaning towards raid atm.
Another thing to consider is stripe size. Couldn't a 16k stripe make Windows snappier to make up for the lowered seek/access times? Too many variables and a lack of evidence to support anything clearcut for me at the moment.
 
Kaiju said:
Another thing to consider is stripe size. Couldn't a 16k stripe make Windows snappier to make up for the lowered seek/access times? Too many variables and a lack of evidence to support anything clearcut for me at the moment.
Might backfire and reduce sustained speeds :p
 
Only way really to find out, is to install your OS with different combinations of stripe size and benchmark your commonly-used programs or use a synthetic benchmark (*spit*). I did this when I setup my RAID-0 for games (I have far too much spare time) because its obviously essential that you can get the cool toys before everyone else when the map changes in Battlefield 2. :cool:
 
matja said:
Only way really to find out, is to install your OS with different combinations of stripe size and benchmark your commonly-used programs or use a synthetic benchmark (*spit*). I did this when I setup my RAID-0 for games (I have far too much spare time) because its obviously essential that you can get the cool toys before everyone else when the map changes in Battlefield 2. :cool:
What stripe did you find best with BF2?
 
Back
Top Bottom