Re: Optimization Question

[Ted Dennison <dennison@telepath.com>]

In article <94hm5q$nmc$1@nnrp1.deja.com>,
  Robert Dewar <robert_dewar@my-deja.com> wrote:
> In article <94hjbp$ks6$1@nnrp1.deja.com>,
>   Ted Dennison <dennison@telepath.com> wrote:
>
> > My first iteration took the dumb approach and used Direct_IO
> > instantiated on bytes.
>
> This of course is even *WORSE* than using sequential_io, since
> there is extra positioning overhead. I can't imagine why you
> would choose Direct_IO for what is obviously a sequential
> problem. Just shows that if there is a way to abuse things
> someone will take advantage of it :-) :-)

:-)

I used Direct_IO because for my purpose (splitting the file into chunks
that will fit on the floppy), I needed to know how big the file actually
is. Direct_IO has a length-of-file routine (Size), while Sequential_IO
does not.

I suppose I could have used Direct_IO just for that purpose and
Sequential_IO for the actual IO operations. But that seemed a little
silly when Direct_IO also had all the operations I needed.

So are you saying that the versions of Read and Write in Direct_IO that
don't have the "Positive_Count" typed parameters still take longer than
the routines with the exact same parameter profile in Sequential_IO? Not
that it really matters that much. We've already ascertained that its
going to be slow either way.

--
T.E.D.

http://www.telepath.com/~dennison/Ted/TED.html


Sent via Deja.com
http://www.deja.com/