[sdiy] Small MCU MIPS, DMIPS?

[karl dalen]

>Scott Gravenhorst <music.maker at gte.net>:
 
> As mentioned before, RAM is expensive real estate-wise.

Yes i know that's why im disappointed on the marketing
departments they want it to be expensive to protect their
business. (now im repeating 12 year old sdiy statement).

What does it matter for a diy or small scale company
if a chip without 1MB dram cost 2USD and one with cost
4-5USD? The cost involved to get those external 1MB, 
design, solder, space debug ect could be higher in the
end in small runs of 1000units or even less.

Do they even manufacture 1Mbyte Dram's anymore?
Isn't it all SDDR 1Gbyte? I consider this a valid question.
 
>JJS wrote
>But for what price?
Slightly above regular 8bitters of same spec
but without 1Mb dram. Nothing spectacular.

>And what else did it have?
The regular stuff, nothing spectacular.
>What was its purpose?
General purpose MCU.

>And who was buying it? - which reflects the reality that some big >consumers of devices can skew the perception that a certain design is >commonplace.

Sound to me your a bit suppriced that someone actually had made a
silly MCU with 1MB DRAM, well the japps has always been on the
edge of thinking anew.
 
Besides how many MCU's haven't just lived a 1-2 year life and
diapered because of bad promotion not a bad design? I bet several
hundreds. It's actually a matter of decenet promotion and toolsets
much more then fancy MIPS and amount of peripherals in hell, a lot
many devices have had so many PEP's and insanely complex to use so
the the device self died because of that..

Why not a quad core dsPIC 40Mhz and less PEP's, i buy that any day
the time spent on understanding and debug PEP's could be put into
bit bang emulate the simplest, SPI for instance, whatever!

There needs to be a paradigm shift soon to get away from all
those silly 8,16bit 20-40Mhz MCUs. (hopefully im provocative
enough here) :-) 
 
> I bolted from analog a while ago for just this
> reason.  

You may have not if the entire CEM line would have been
still alive and well, including new devices who combined
table reading and VCO VCF. 

>My choice was the FPGA for it's sheer logic gate count
>which can accomodate very complex systems. 

Same thing here, i'm disappointed there are no DIP16 FPGA
a million apps dont need 2000balls in the pants. How many
balls have you ever used in your apps? Less then 50?

>high in my opinion.  Yeah, I don't get to solder
>solder solder, but I think I've done some things 
>I couldn't possibly have done without the FPGA at
>my disposal.  DIY for me has become writing my
>ideas in Verilog. Maybe not so much fun for everyone,
>but then  neither is mountain climbing.

yes, but you have also spent years to get where you are now
so the mountain was there already it's slopes where perhaps
not so steep but long..which declares the point of do we 
really want to solder (analog) write (digital) rather then
make tunes and sounds?

I bet the older you get the more you might value your time
for either one?! At least i do. No im not saying one shouldent
learn new, of course on should but on on your own premises.

Why spend years on something that should be way easier, there
are still the prestige thing in engineering that dictates if
you can do something that's easy make it complex so it sells
better.;-) hehe...

KD