Date   : Fri, 05 Feb 2010 12:10:48 +0000
From   : francis@... (Francis Devereux)
Subject: 'BASIC 512K'

On 5 Feb 2010, at 11:33, John Kortink wrote:
>
> On Fri, 5 Feb 2010 01:47:04 -0000, Mike Pepper <profpep@...>
> wrote:
>
>>
>> ----- Original Message -----
>> From: "John Kortink" <kortink@...>
>> Sent: Friday, February 05, 2010 12:26 AM
>>>
>>> On Fri, 5 Feb 2010 00:03:52 -0000, "David Hunt" <dm.hunt@... 
>>> >
>>> wrote:
>>>
>>>>> John Kortink wrote:
>>>>>> ReCo6502 can now run a 65816-patched HiBASIC 4r32
>>>>>> that provides 448 KB of variable storage space,
>>>>>> with no significant performance penalty (due to
>>>>>> the 24-bit addressing capabilities of the 65816).
>>>>>
>>>>> Time to get obtain some appropaitely shaped tuits and finally get
>>>>> around to writing that 65816 emulator I've been meaning to.
>>>>
>>>> How's about a 32016 emulator? It'd be v.useful to have running on  
>>>> the ARM
>>>> co-pro.
>>>
>>> It'll probably end up being in BASIC ... ;-)
>>>
>>>
>> And will probably run about as fast as the original 32016!
>
> Not by a long shot.
>
> An ARM7 copro could probably manage 1 to 3 times the
> speed of a 2MHz 6502, if coded in slick assembly. A
> 6 MHz 32016 will most likely end up below the 100%
> line. Coded in BASIC it will be slower yet by a huge
> amount (probably between 5 to 15 times).

ISTR there was a BBC emulator for the Archimedes that emulated a model  
B at full speed on an ARM3 (or maybe even an ARM2).  So if my memory  
is accurate (a big if ;-) then wouldn't an ARM7 do better than 3 times  
a 2MHz 6502? Especially considering that it wouldn't need to emulate  
the graphics hardware etc. if it was only emulating the 'parasite'  
side of the Tube.

Francis