Date   : Sun, 10 Sep 2006 20:03:54 +0100
From   : "Ian Wolstenholme" <BBCMailingList@...>
Subject: Re: Tube File Copying

Examining the times stamped on the files by the MDFS the curious
thing is that the first 8MB chunk took about 50 minutes and the
second one took over an hour and then after that it was about 40
minutes each.  I don't really know how this occurred as there was
no other network activity going on at the time, although judging by 
the orrible clattering noise coming out of the LVROM player part of
the time, maybe there were quite a few retries on that side of
things.

Best wishes,



Ian

----- Original Message -----
From: jgh@... (Jonathan Graham Harston)
To:  bbc-micro@...
Sent:  09 Sep 2006 08:58:43 +0100
Subject: [BBC-Micro] Re: Tube File Copying

>Message-ID: <38B971F35D084FFD925544ADB638189E.MAI@...>
 
"Ian Wolstenholme" <BBCMailingList@...> wrote:
> It's been going for 2 hours now although it's only at the stage of
> loading the second 8MB chunk.  I suppose loading each chunk is
> going to get longer and longer as it has to get to the starting point
> in the file each time before reading any data.
 
It shouldn't do.
 
  PROCgbpb(rd%,chn%,data%,8*1024*1024,0)
 
should take the same amount of time as:
 
  PROCgbpb(rd%,chn%,data%,8*1024*1024,8*1024*1024) [1]
 
The filing system just does:
 
  sector=file's_start_sector+(PTR DIV 256)
 
[1] Using PROCgbpb(action, handle, address, number, pointer)
 
-- 
J.G.Harston - jgh@...                - mdfs.net/User/JGH
Sheffield Boundary Review at http://mdfs.net/User/JGH/Docs/Politics/ParlReview