[Coco] 1995 vintage 68B09, a good year?

[jdaggett at gate.net]

On 28 Apr 2005 at 12:25, Kevin Diggs wrote:

Date sent:      	Thu, 28 Apr 2005 12:25:59 -0700
From:           	Kevin Diggs <kevdig at hypersurf.com>
To:             	CoCoList for Color Computer Enthusiasts 
<coco at maltedmedia.com>
Subject:        	[Coco] 1995 vintage 68B09, a good year?
Send reply to:  	CoCoList for Color Computer Enthusiasts 
<coco at maltedmedia.com>
	<mailto:coco-
request at maltedmedia.com?subject=unsubscribe>
	<mailto:coco-
request at maltedmedia.com?subject=subscribe>

> JDag,
> 
>  What do you think this could be pushed to?
> 
>  I thought in earlier discussions, many moons ago, you alluded to the
> fact that these later parts would be done in smaller geometries for
> economic reasons (yeilds?). Wouldn't this also make them faster?
>
*******
Kevin The 1995 date code is post cancelllation date code and should be the latest 
mask set. This will have all the corrections and changes from the original design in 
1979. As for clock speed it is very much possible that the processor will run with 
buss speeds up to 3MHz.  Take care that after 2 MHz the processor does get warm. 
You will  need to do some heat sinking to keep the part cooler. 


>  Mindless ramble:  Ya know the TFR and EXG instructions? Ya think the
> "size mismatched" values could be used as a 16 register cache (I know
> - the 68HCxx embedded processors use this for extension. This ain't
> one of those. It is a 6809)? For practical use these instructions
> would need to be much faster, like single cycle, and fast context dump
> and restore instructions would be needed.
> 
*******

The EXG instruction will exchange the contents of R1 and R2. So what ever is in R1 
will end up in R2. What ever is in R2 will end up in R1. The registers must be of the 
same bit width. That is 8 bit registers can only be EXG with 8 bit registers. 

The TFR instruction is essentially a copy instruction . That is what is in R1 will be 
copied over to R2. Again this can only be done with registers of the same bit width. 


james

>      kevin
> 
> P.S.:  Mac bus speed now 45. With 9x multiplier processor is at 405.
> As soon as I figure out what to stick in the L2CR I'm gonna try to
> build Mozilla. Think that is a good test of system stability? Other
> suggestions for Linux stress tests appreciated.
> 
> -- 
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco