[MacDV] Re: how and why to benchmark

Dmytro Koval'ov dmytro.kovalov at nikkocitigroup.com
Wed Jan 7 17:08:14 PST 2004


>>>>> On Wed, 7 Jan 2004, "James" == James Asherman wrote:


  James> Let me say I have already performed my own biased tests.  I use
  James> Norton speed disk.  Takes 5 seconds to defrag and OTIMIZE the
  James> video storage disk because it is empty.  It gets done w/ the
  James> Media setting.  When I do my boot disk it take somewhat longer

What you've written here is completely wrong. Time used for disk 
optimizing has nothing to do with access time or I/O. For sure 
any defragmenter will finish ''defragmenting'' empty disk faster that disk 
with files on it. But because it's EMPTY, not because it's 
defragmented.

  James> virtual disk space and the updates on the other and little
  James> holes where small pieces of OS used to be but are now deleted
  James> that this is a factual if not accurate representation .(the
  James> same goes for Apps and other files) I believe that when I
  James> defrag and Optimize, those holes are in fact filled, reducing

James, I think you still didn't get the idea of what Peter is talking 
about. 

Because of all that reasons that Peter mentioned, you (or Speed disk -
which is actually a DOS product ported to Mac -- completely different OS
and completely different disk access methods) can not tell for sure that
two consecutive rectangles in the Speed Disk window, represent sectors
that are actually located in the consecutive (or even adjacent) areas on
disk. And it is quite probable that while filling those gaps Speed Disk
is actually making things worse because it is moving pieces of the same
file into different areas of disk, thus making access time longer.
Especially with Panther, which on itself optimizes disk access. And be
sure that Panther knows better HOW to optimize that Norton.

  James> the number of possible bad or wrong choices that the disc
  James> controller might make in the future. I also believe that when
  James> it shows me and tells me the numbers of fragmented files and
  James> shows the contiguous space available that, that too, may not be
  James> actual but the fact that the space is being reported to speed

Controller DOES NOT report to speed disk whether space contiguous or 
not. Speed disk decides this and draws the picture of what it THINKS is 
representing the drive.

  James> I know even though one could run for years without
  James> doing it, my business dictates that in addition to constant
  James> updates and downloads and my personal fun, that it is prudent
  James> to do this procedure on my boot disc in order to have the
  James> greatest flexibility and capacity I can.

Existence of the empty sectors on disk or fragmentation/
de-fragmentation has nothing to do with capacity and/or flexibility. 

--Dmytro





More information about the MacDV mailing list