almost full HD slows things down?

[DZM]

Could anyone verify this claim?

When you start using a new HD, the files are allocated with sectors close to the outside of the HD, and so the linear speed of the head is faster there. As the HD is being occupied, the sectors move to the inner circles, and the linear speed drops because of the smaller radius of the inner circle. Someone told me that the sustained speed of HD when transferring large files could drop by half when more than 2/3" occupied. Please verify.

 

[angrybusdriver]

Maybe. It may use up to the maximum reported seek time, but that's not really enough to warrant a slow down.

Usually when hard drives are pretty full, Windows has been installed for a long time, which increases the chance of there being spyware, virii, too many installed programs/huge registry/etc. installed on the computer.

Just defrag once in awhile and there should be no slowdown noticed.

 

[STL]

[DZM]
> Could anyone verify this claim?

Yes, hard drives start from the outside in. Sustained Transfer Rate is higher at the beginning of the drive (the outside of the disk) than at the end of the drive (the inside of the disk).

> Someone told me that the sustained speed of HD when transferring
> large files could drop by half when more than 2/3" occupied.

See the WD2500JD Transfer Rate graph at StorageReview - Sustained Transfer Rate for that drive drops from 57.9 MB/sec to 36.5 MB/sec.

It is a gradual process.

[angrybusdriver]
> Maybe. It may use up to the maximum reported seek time

What you are talking about is not relevant to the original poster's question.

> spyware, virii, too many installed programs/huge registry/etc.

Not relevant.

> Just defrag once in awhile and there should be no slowdown noticed.

Defragmentation reduces the need for unnecessary seeks. Optimization (e.g. what Norton Speed Disk does) can move frequently-accessed files to the front of the drive so they are closer together and can be read/written with higher STR. But nothing can hide the decreased STR at the end of a full drive.

 

[DZM]

Originally posted by angrybusdriver
Maybe. It may use up to the maximum reported seek time, but that's not really enough to warrant a slow down.

Usually when hard drives are pretty full, Windows has been installed for a long time, which increases the chance of there being spyware, virii, too many installed programs/huge registry/etc. installed on the computer.

Just defrag once in awhile and there should be no slowdown noticed.

Then, am I right on this ---

Even if a NB 2.5" HD has the same RPM (e.g. 7.2K) and cache as a 3.5" HD, it could not have the same performance due to its smaller caliber?

 

[Ice Czar]

if the areal density is the same youd be correct
(number of bits per square inch)

http://www.hardforum.com/showthread.php?s=&threadid=699246
http://www.hardforum.com/showthread.php?s=&threadid=741512

but the sustained transfer rate isnt as important as many people make out, there isnt that many aps that benifit from it (the files being smaller, or fragmented) and so the access times and fragmentation level often have more to do with the percieved performance

notable exception video editing

 

[Snugglebear]

Originally posted by DZM
Then, am I right on this ---

Even if a NB 2.5" HD has the same RPM (e.g. 7.2K) and cache as a 3.5" HD, it could not have the same performance due to its smaller caliber?

The disk with the 2.5" platters will have a lower density but allow for faster access times due to the restricted size. Most companies no longer do this, but rather destroke the drive and only use, say, the first inch on that 3.5" drive. It's cheaper to make all 3.5" platters and only use the fastest portion on smaller (in total storage terms) drives. Also note that most high-speed SCSI units use smaller platters to allow them both faster access times and higher rotational latencies. If you use SCSI like some of the rest of us, you'll notice that the areal densities on our drives are lower, the platters smaller, yet the overall performance is much higher. Higher density, after all, makes random accesses harder and harder because the chance of missing the spot during a given rotation goes up.