Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!zephyr.ens.tek.com!tekig5!wayneck
From: wayneck@tekig5.PEN.TEK.COM (Wayne C Knapp)
Newsgroups: comp.sys.amiga.hardware
Subject: Re: (A) Amy 500 -> 2000? (B) 68040 vs. gfx coprocessor
Message-ID: <5575@tekig5.PEN.TEK.COM>
Date: 9 Feb 90 21:24:00 GMT
References: <633@xdos.UUCP> <9527@cbmvax.commodore.com>
Organization: Tektronix Inc., Beaverton, Or.
Lines: 58

In article <9527@cbmvax.commodore.com>, mks@cbmvax.commodore.com (Michael Sinz - CATS) writes:
> >Obviously a 25Mhz 68040 with matching memory would be faster at graphics
> >than the current blitter. On the other hand, it would still be slower
> >(at BLIT operations) than a 25Mhz blitter, because of the optimal
> >blitting instructions.
> 
> But, the 68040 PLUS another chip will always be faster than JUST
> the 68040 (in a multitasking system)  That is because you could
> have the 68040 do a complex 3D transform with shading and then
> start the blitter with that object and continue to calculate on
> the next object while the blit happens. 

How would the blitter help?  If the 68040 is shading the object then it must
be drawing the object.  So the CPU would simply just put the object in the
bitmap as it is drawing.  There is nothing for the blitter to do here.

> Note that I am not saying 
> that I want the current blitter to stay the same.  Maybe going 
> with a TI 34010 or 34020 would be the best way, but in any case, 
> making the main processor do the work is wrong.  Yes, it might be 
> faster than the blitter (34010) in a single-task-do-this-now 
> situation, but when more than one thing is going on, that is not
> the best way.  (Divide and conquer...)
> 

Now this is what I like to see! People at commodore doing some real
thinking!!

The current Amiga blitter is very limited.  For the most part it is 
only used a few logical operations between two banks of memory or even
less than that.  Part of the problem is that it takes a fair amount of
work to write code to use the blitter's functions not supported by
intuition (like doing a 3 source cuttie cut blit) and the chip is faily
limited in scope.  That is the blitter has to be setup and controlled
by the CPU, it can't be programmed to do anything on it's own, it can't
even access its own registers, in any reasonable fashion.  So if you 
already have a 68020 or 030 or 040 the blitter is of less and less use 
since the processor can easily replace the blitter's functions and in 
many cases such as blitting text to the screen a fast processor can get
the job done in the time it takes to set up the blitter in the first place.

So if you are really going to have a coprocessor then it should be something
like the TI 340x0 chips.  These chips are programable meaning they are 
useful for many appilications.  The blitter is so limited that instead of
being called a coprocessor it should be just called a co-ALU.  After all
it mainly gives addional logic functions between memory.  

So here is a question.  Saying that a very enhance Amiga were built out
of a 68040 and 34020 could the 34020 be programmed to allow all the old 
Amiga programs to run without changes?  That is programs that didn't access
the old Amiga hardware directly.  Besides the much lower chip count, faster
and greatly enhanced features, I can think of dozens of uses that I could
put a 340x0 to use that I could never even begin to use the current hardware
on.  Things like decompression of anim files, image processing with say 3x3
matrices, FFTs or maybe faster cosine transforms, ... then with the power
of the 68040 with it this would be one impressive graphics machine.

                                                  Wayne Knapp