In news:h17gtj$ekk$1@news.eternal-september.org,
~misfit~ typed on Tue, 16 Jun 2009 19:17:04 +1200:
> Hi Bill.

Hi Shaun!

> I find that, using PerfectDisk set to 'Smart Defrag', it makes a big
> difference. Smart Defrag puts all the files needed for booting at the
> fastest part of the drive, then categorises the remaining files into
> three groups, depending on when the were last altered. 'Rarely
> Modified' are put next, "Occasionally Modified' follows with
> 'Freqently Modified' left at the end of the data on the drive so
> that, when it is modified, it's not squeezed into little gaps left
> where, for instance, you may have deleted a file that had been on the
> drive since the begining.
>
> The times for the different groups are user-adjustable but I find
> that the defaults work well. The bus isn't the bottleneck. For
> instance this R51 has an ATA100 IDE interface that is theoretically
> capable of moving 100MB/second. The HDD, even though it's been
> upgraded from the original 4,200rpm to a 5,400rpm, would be
> hard-pressed to hit 40MB/second on a sustained read and far less that
> that when seeking all over the disk for scattered (fragmented) files.
> Therefore, IMO, the bus isn't the bottleneck...

Well if the I/O isn't a bottleneck, then why do they put buffers on hard
drives for? As there is no need for buffers. Remember the I/O chips has
to handle other things too. So the hard drive has to time share with
other devices. Thus makes the I/O the bottleneck in most cases IMHO.

I'll try defragging my Gateway laptop again, but I swear I never see any
speed improvements. So I don't know what the big deal is all about. On
old MFM drives, they were so slow, it *did* make a huge difference
there. But I never seen any improvement on IDE drives yet.

--
Bill
Windows 2000 SP4
Asus EEE PC 701G4 ~ 2GB RAM ~ 16GB-SDHC