[Coco] What would a CoCo successor have to have as a minimum?
Aaron Wolfe
aawolfe at gmail.com
Sat Nov 20 08:47:43 EST 2010
On Sat, Nov 20, 2010 at 8:32 AM, Mark McDougall <msmcdoug at iinet.net.au> wrote:
> On 20/11/2010 11:39 AM, Aaron Wolfe wrote:
>
>> I'm not sure where the idea that FPGA is slower than a PC came from,
>> but it's simply not true.
>
> Unfortunately, it simply *is* true Aaron.
>
> IIRC the absolute best we can squeeze out of Coco3FPGA is... what... 25MHz?
>
> How fast do you think you can emulate a 6809 on a 3GHz machine? The mediocre
Well, unless we are saying VCC is not optimized for speed (others here
have said that it is, I do not know myself), then based on
guesstimates here I'd say 3.0 might give the rough equivalent of
40Mhz? if we assume "about 8 seconds" is what 25Mhz takes, and my
2.66Ghz machine with Vcc takes 5.5... VCC is doing ~36.36Mhz on my
2.66 Ghz machine.
There are plenty of factors unaccounted for here, but I think its
unlikely that any coco4 project will be writing their own emulator
from scratch, so using the fastest emulator that we already have seems
the best way to measure. On the other hand, a 25Mhz FPGA Coco is
considerable *faster* than my modern PC running MESS, so if the
project will be using MESS as the base since it is open source, they
may not be able to match current FPGA speed on any current general
purpose computer without a lot of tweaking.
Anyway, my point is not to "speed test" things. A few comments on the
list seemed to say that if a project supported both emulaltors and
FPGA, the FPGA's performance would limit the capabilities of this
proposed coco 4. This seems unlikely. The processor alone in my gen.
purpose PC costs twice what a DE1 board costs. All things considered,
FPGA would become a limiting factor only when compared to very
expensive computers, and even then we are talking about performance in
the same "class", not an order of magnitude difference.
-Aaron
> performance you're seeing is because the emulators in question aren't
> optimised for speed - because they don't need to be. In most cases they're
> optimised for cross-platform compatibility, and for ease of programming.
>
> I've seen a 1?2MHz Amiga copro emulate a 2MHz Z80 system at full speed.
> Imagine what you can do with a processor clocked 250 times faster...
>
> Regards,
>
> --
> | Mark McDougall | "Electrical Engineers do it
> | <http://members.iinet.net.au/~msmcdoug> | with less resistance!"
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco
>
More information about the Coco
mailing list