Searching \ for '[EE] Programming different systems' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/microchip/devprogs.htm?key=programming
Search entire site for: 'Programming different systems'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Programming different systems'
2006\01\18@054906 by Alan B. Pearce

face picon face
Tag changed again (common guys, its long past being PIC)

>I quite enjoy this lack of space in the PICs; it harks
>back to a time 20-odd years ago when it was a challenge
>to write the application you wanted to on the micro of
>your choice.

I once worked with a guy work was a whizz with unix of various flavours. He
reckoned the nicest was the ones that ran on 80286 based machines as the
segmented architecture forced the programmers to be careful with their code,
so the result was the code tended to be fast and lean. Once the 386
architecture with large flat memory space hit the streets he reckoned that
the coding got sloppier and bloated, with resultant performance hits.

As to the x86 architecture becoming unlicensed, I would have to ask - who
would want it? Why use a segmented architecture when there are so many flat
ROM space micros around that can handle as much memory - ARM, various
extended Z80 (although this is segmented to some degree), the various
Motorola family chips, and so on.

2006\01\18@205647 by Chen Xiao Fan

face
flavicon
face

> -----Original Message-----
> From: spam_OUTpiclist-bouncesTakeThisOuTspammit.edu
> [.....piclist-bouncesKILLspamspam@spam@mit.edu] On Behalf Of Alan B. Pearce
> Sent: Wednesday, January 18, 2006 6:48 PM
> To: Microcontroller discussion list - Public.
> Subject: Re: [EE] Programming different systems
>
> As to the x86 architecture becoming unlicensed, I would have
> to ask - who would want it? Why use a segmented architecture
> when there  are so many flat ROM space micros around that can
> handle as much memory - ARM, various extended Z80 (although
> this is segmented to some degree), the various Motorola family
> chips, and so on.

I do not know what aspect of the x86 architecture becomes
unlicensed but I guess that many people will be interested.

AMD has just licensed x86 architecture to China (actually
only Geode based embedded x86) and it is touted as a
very success PR for AMD in China. Before this China has
only the Godsend (Long Xin) roughly based on MIPS which
roughly approach Pentium III at 500Mhz.

There are still many 80186 class MCU based application in
the market. I guess the shear volume of software and
x86 experts in the market will make them very competitive
with ARM. Now the embed x86 is not as popular because
Intel/AMD/Via controlled it. If x86 is really opened up
(I doubt this is really the truth), more players will
come. For example, if Taiwanese MCU players like Holtek,
Sunplus and Winbond can be in this, they will make it
a success. Even though Holtek/Sunplus/Winbond are not
famous in the MCU market at all, they are actually within
the top 10 in the 8-bit market. They are not famous because
their major market is in Asia only (especially the greater
China area) where cost is of the utmost importance.

Z80? No I do not think they will be main stream again even
though I learned it in the university along with MCS51.
MCS51 will live for very long. So does Z80 but Z80 will
only be a niche.

Regards,
Xiaofan

2006\01\19@161915 by Peter

picon face

On Thu, 19 Jan 2006, Chen Xiao Fan wrote:

> famous in the MCU market at all, they are actually within
> the top 10 in the 8-bit market. They are not famous because
> their major market is in Asia only (especially the greater
> China area) where cost is of the utmost importance.

Speaking of cost, I think that the x86 frenezy will cool down very fast
when the real price war will start and silicon area will begin to
matter. The ARM6 core has something like 35,000 transistors. That's
about a fiftieth of a Pentium afaik (Pentium at 60MHz already had 3.1
million transistors, even 386dx had 275,000), and it runs on 1 Watt at
233 MHz (that is part of the reason why).

Links:

http://en.wikipedia.org/wiki/ARM_architecture
http://en.wikipedia.org/wiki/List_of_Intel_microprocessors

The Z80 was a 'niche' but it appeared in a lot of products that were
mass manufactured. The Nintentdo game boy and Sharp calculators come to
mind here. It is interesting to note that the GBA uses an ARM cpu
instead of the Z80 used by its predecessor.

According to Wikipedia ARM holds 75% of the embedded (32 bit) market
now.

Note that Intel's XScale is ARM. 1GHz clock ARMs are available from
Intel.

Peter

2006\01\19@165501 by Padu

face picon face
From: "Peter" <plpspamKILLspamactcom.co.il>
<snip>
> Note that Intel's XScale is ARM. 1GHz clock ARMs are available from
> Intel.
>
> Peter

There are a few dev boards and software development solutions for the
Philips platform. What about for the intel platform?

I think what I'm trying to ask is: I feel confident in using Philips for
hobby and small scale stuff... and I have this impression that intel is just
for big scale commercial applications... am I right?

Cheers

Padu

2006\01\20@141457 by Peter

picon face

On Thu, 19 Jan 2006, Padu wrote:

> From: "Peter" <.....plpKILLspamspam.....actcom.co.il>
> <snip>
>> Note that Intel's XScale is ARM. 1GHz clock ARMs are available from
>> Intel.
>>
>> Peter
>
> There are a few dev boards and software development solutions for the Philips
> platform. What about for the intel platform?

See here for some nice 'development boards complete' with case and
display:

http://tuxmobil.org/pda_linux_xscale.html

(almost) POSIX compliant operating system and development environment
supplied for free. Just add applications. There is much more on the
internet. Search Google.

> I think what I'm trying to ask is: I feel confident in using Philips for
> hobby and small scale stuff... and I have this impression that intel is just
> for big scale commercial applications... am I right?

Imho the Philips parts are a breakthrough because they are available
directly from mail order vendors. Afaik it is the first time this
happens (ARM chips in ones by mail order and mostly <$10 each).

Peter

2006\01\26@012313 by Chen Xiao Fan

face
flavicon
face


> -----Original Message-----
> From: EraseMEpiclist-bouncesspam_OUTspamTakeThisOuTmit.edu
> [piclist-bouncesspamspam_OUTmit.edu]On Behalf  Of Peter
>
> Imho the Philips parts are a breakthrough because they are available
> directly from mail order vendors. Afaik it is the first time this
> happens (ARM chips in ones by mail order and mostly <$10 each).

I think they are very ambitious and they want to conquer sub-US$3
market as well. It is interesting to know quite these ARM MCU
players (eg: Philips and Atmel and Keil ) are also 8051 players.

Regards,
Xiaofan

2006\01\26@013105 by Chen Xiao Fan

face
flavicon
face

> -----Original Message-----
> From: @spam@piclist-bouncesKILLspamspammit.edu
> [KILLspampiclist-bouncesKILLspamspammit.edu] On Behalf Of Peter
>
> Speaking of cost, I think that the x86 frenezy will cool down
> very fast
> when the real price war will start and silicon area will begin to
> matter. The ARM6 core has something like 35,000 transistors. That's
> about a fiftieth of a Pentium afaik (Pentium at 60MHz already had 3.1
> million transistors, even 386dx had 275,000), and it runs on
> 1 Watt at 233 MHz (that is part of the reason why).

I think this is not so true since the silicon area is not
the most important factor in the slightly higher end market.
I think the development cost is one of the most important
factor since the chip development cost is prohibitive high
now in the higher end. That is one reason FPGA/CPLD are doing
quite well.

Even if the silicon area is the major concern, however, in the
embedded MCU market, the core is only a small fraction of the
whole silicon area. Often the memory sub-system
and peripherals are taking more spaces.

Regards,
Xiaofan

2006\01\26@031515 by Shawn Tan

flavicon
face
On Thursday 26 January 2006 06:31, Chen Xiao Fan wrote:
> Even if the silicon area is the major concern, however, in the
> embedded MCU market, the core is only a small fraction of the
> whole silicon area. Often the memory sub-system
> and peripherals are taking more spaces.

Yea, much of the silicon area is definitely consumed by RAM.. If you look at
the chip micrographs, you'd see this.. This is because they use SRAM.. But
even with DRAM, you'd need a sufficiently large transistor to store the
charge.. There are attempts at reducing this by using novel ram designs.. one
that i've heard of recently is using MEMS ram.. but many of these things are
still at an early stage..

cheers..

--
with metta,
Shawn Tan.

It is easier to change the specification to fit the design than vice versa.

--
with metta,
Shawn Tan

2006\01\26@035257 by Chen Xiao Fan

face
flavicon
face

> -----Original Message-----
> From: RemoveMEpiclist-bouncesTakeThisOuTspammit.edu
> [spamBeGonepiclist-bouncesspamBeGonespammit.edu] On Behalf Of Shawn Tan

> Yea, much of the silicon area is definitely consumed by RAM..
> If you look at the chip micrographs, you'd see this.. This
> is because they  use SRAM.. But  even with DRAM, you'd need a
> sufficiently large transistor to  store the
> charge.. There are attempts at reducing this by using novel
> ram designs.. one that i've heard of recently is using MEMS ram..
> but many of these things are  still at an early stage..
>

MEMS RAM? Interesting. I thought MEMS is not small at all and
people are using those obsolete semiconductor lines to produce
MEMS. I think MEMS is a very interesting subject like those
mirrors used in telecom equipment and those nano-sensors ...

Regards,
Xiaofan

2006\01\26@051711 by Peter Todd

picon face
On Thu, Jan 26, 2006 at 02:31:03PM +0800, Chen Xiao Fan wrote:
> Even if the silicon area is the major concern, however, in the
> embedded MCU market, the core is only a small fraction of the
> whole silicon area. Often the memory sub-system
> and peripherals are taking more spaces.

Heck, I remember seeing a die shot of a PIC chip, forget exactly which
one but I think it was one of the smallest series, 12F or 10F, where
about one quarter of the die was taken up by the charge pump for the
eeprom. One quarter, just for a capacitor!

--
TakeThisOuTpeteEraseMEspamspam_OUTpetertodd.ca http://www.petertodd.ca

2006\01\26@083128 by Shawn Tan

flavicon
face
On Thursday 26 January 2006 10:31, Peter Todd wrote:
> Heck, I remember seeing a die shot of a PIC chip, forget exactly which
> one but I think it was one of the smallest series, 12F or 10F, where
> about one quarter of the die was taken up by the charge pump for the
> eeprom. One quarter, just for a capacitor!

yea.. capacitors are notably difficult to construct in modern digital CMOS
processes.. MIM caps are only 1fF/um2.. the reason being that large
capacitors are usually used in analogue designs and so, digital processes
don't bother much with them..

cheers..

--
With Metta,
Shawn Tan.

"Engineering is the Art of making what you Want from what you can Get."
- Dave Tweed

2006\01\26@083143 by Shawn Tan

flavicon
face
On Thursday 26 January 2006 08:52, Chen Xiao Fan wrote:
> MEMS RAM? Interesting. I thought MEMS is not small at all and
> people are using those obsolete semiconductor lines to produce
> MEMS. I think MEMS is a very interesting subject like those
> mirrors used in telecom equipment and those nano-sensors ...

oh silly me.. it was CNT ram... i don't know why i thought of MEMS.. must've
been because it was mechanical memory.. i think they bent the tube one way or
the other to store a bit.. anyway, was in a recent IEE..

cheers..

--
With Metta,
Shawn Tan.

"Engineering is the Art of making what you Want from what you can Get."
- Dave Tweed

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...