RE: Some Code to Look At

["Albertson, Chris" <CAlbertson@primeadvantage.com>]

I'm seeing times in the "hand full of seconds" range for
writes of 16MB images to disk.  This is in line with the
hardware specs the disk manufactures publish.  Modern
drives have an I/O rate of about 6 megabytes per second at
the head to platter.  They can accept data _much_ faster
across the IDE cable but the buffer fills quickly as the
the 6 MB/sec rate is determined by the bit density and the
rotation rate of the platter.  So we should expect a 16MB
image to be written in well under 5 seconds.  If it takes
longer, something other then the disk drive is at fault.

I can get these speeds on a P200 system with a three year old
IDE drive.  Nothing fancy at all.



> -----Original Message-----
> From: Andrew Bennett [mailto:andrew.bennett@ns.sympatico.ca]
> Sent: Sunday, January 28, 2001 12:27 PM
> To: Tom Droege; tass@listserv.wwa.com
> Subject: Re: Some Code to Look At
> 
> 
> On Sat, 27 Jan 2001 11:46:27 -0600, Tom Droege 
> <tdroege@veriomail.com> wrote:
> 
> >Below is the code that writes the contents of the Mark IV 
> Memory card to a 
> >disk file.
> > ...
> [ referring to writing MK IV data ]
> 
> I dug out my old 486/33MHz and got the following results:
> Free Pascal or Borland Pascal 7.0:
> Transription of (most of) the BASIC code 109 sec
> Buffering to 16384 Byte buffers 73 sec
> and for BP7.0:
> Turning off all checking 49 sec (Free Pascal 53 sec)
> Inner loop in Assembler 45 sec
> There is not much more to be had by Assembler coding!
> 
> Here is the Pascal, for what it is worth.
> Andrew Bennett, Avondale Vineyard
>