Hallo Ingo,


> EXOS is a kernal-replacement-only speeder,

It was my favourit speedloader before JD.


> so that address refers to the original 1541 rom. I think it
> also works with a Jiffy-1541, but I'm not 100% certain.

If that is part of the the code that is uploaded to the drive, I'm quite 
sure it won't work because $F82B is the second byte of an instruction 
starting at $F82A. So if Exos indeed relies on this routine, I have a 
problem as JD doesn't have it. The only solution would be to copy the whole 
routine from the original 1541 DOS. Lucky for me (EP)ROM space is not a 
problem :)

Now I regret I discarded my own message. I know I mentioned I could gain an 
extra three seconds. I can use the $F3B1 - Read Header routine to read the 
sector number And then start reading the data starting with the next 
sector. In this way I will loose a little bit less then 1/3 of a second but 
I'll be using routines I'm familiar with.



Hallo Patryk,


>> No, you weren't ignored, but I cannot use the Warp routines because
>> I cannot use the GCRed data.
>
> I think I would double-check this statement.

The Warp routine sends the data directly from the 6522 to the PC. This 
means it is still GCRed and thus I cannot use it. More interesting would be 
the Turbo routine if it wasn't slower as what I already got.


> I have those rewritten in my copier. Can give you access to my
> SVN if you like.

Please do. In my last email I was quite enthousiastic about the way Exos 
decoded the header. But only after sending it I realised that it wouldn't 
speed up things; as you mentioned before, the GCR decoding is still the 
routine where most can be gained.

I hope you also have the routines to write things to disk again?

Some info: I just managed to program my 29C256 again after all the trouble 
of yesterday. This time I used the Find Sector routine for all tracks and 
things worked out fine: still 35 seconds. 
Next thing to try the idea I mentioned above. I'm busy with it but the 
copying stops in the middle of nowhere W/O any reason (at least to me).


--
    ___
   / __|__
  / /  |_/     Groetjes, Ruud Baltissen
  \ \__|_\
   \___|       http://Ruud.C64.org







       Message was sent through the cbm-hackers mailing list