2106 | ADC | ARM7 | Atmel | Bootloader | CAN | CrossStudio | CrossWorks | DDS | ECos | Ethernet | ETM | FIFO | FLASH | FPGA | GCC | GDB | GNU | GNUARM | GPIO | I2C | IAP | IAR | JTAG | Kickstart | LCD | Linux | LPC | LPC-E2294 | LPC2000 | LPC2100 | LPC2104 | Lpc2106 | Lpc210x | LPC2114 | LPC2119 | LPC2124 | LPC2129 | Lpc2138 | LPC213x | LPC21xx | LPC2210 | LPC2212 | LPC2214 | LPC2292 | LPC2294 | LPC2xxx | LPC3128 | MCB2100 | Olimex | Philips | PWM | Rowley | RTC | RTOS | SPI | SSP | UART | UART0 | UART1 | ULINK | USB | Watchdog | Wiggler

XMOS are taking orders for a great-looking piece of kit:

https://products.xmos.com/xc1

This is worth pre-ordering just for the ultimate satisfaction of owning one
and feeling the throbbing power in your hands. Old transputer guys aren't
ready to be put out to grass just yet...

I'm sure you'd love to own one, Leon...

Rgds,

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

----- Original Message -----
From: "Paul Curtis"
To:
Sent: Tuesday, September 30, 2008 12:06 AM
Subject: [lpc2000] OT: To all those geeks that like to play with stuff
> XMOS are taking orders for a great-looking piece of kit:
>
> https://products.xmos.com/xc1
>
> This is worth pre-ordering just for the ultimate satisfaction of owning
> one
> and feeling the throbbing power in your hands. Old transputer guys aren't
> ready to be put out to grass just yet...
>
> I'm sure you'd love to own one, Leon...

I've been in touch with them.

I'm waiting for the chips to become available (they are sampling them). They
offered me one of the kits (ex-demo) at a reduced price.

I've started creating a PCB part for the chip, it isn't easy.

Leon
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Paul,

> XMOS are taking orders for a great-looking piece of kit:
>=20
> https://products.xmos.com/xc1

Thanks for bringing this to my attention

> This is worth pre-ordering just for the ultimate satisfaction of owning o=
ne
> and feeling the throbbing power in your hands. Old transputer guys aren'=
t
> ready to be put out to grass just yet...

Hey, I=C2=B4ve got an old ATW (Atari Transputer Workstation), but never got=
=20
it going. So maybe I catch the train this time :-)
--=20
42Bastian

Note: SPAM-only account, direct mail to bs42@...
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

XMOS is interesting but based on a quick review of the architecture=20
it is a fairly standard multicore SoC with a non-standard ISA. The=20
chip on the board in the kit (XC-1) has only 4 cores and tops at 1600=20
MIPS or 1.6 BOPS according to the XMOS.=20

For comparison IntellaSys makes a 40 core chip with up to 25000=20
billion operations per second or 25 BOPS. The only twist is that=20
these cores are Forth machines which is not surprising when see who=20
is behind them .

But back to ARMs.

Looks like multicore SoCs are beginning to invade the microcontroller=20
space so it will be interesting to see how things develop. Personally=20
I would love to see an ARM mutlicore SoC with 32 or more cores plus a=20
hierarchical interconnect and some SRAM/Flash. But for now we have to=20
contend with the dual core ARM9s/ARM7s found in most of the MP3s and=20
iPods. Incidentally NXPs own LPC288X seems to be just "half" of the=20
chipset used for the MP3 players, or maybe it is just a reject from a=20
production with one dead core and/or Flash (just a pure speculation)?

Pawel

>
> Paul,
>=20
> > XMOS are taking orders for a great-looking piece of kit:
> >=20
> > https://products.xmos.com/xc1
>=20
> Thanks for bringing this to my attention
>=20
> > This is worth pre-ordering just for the ultimate satisfaction of=20
owning one
> > and feeling the throbbing power in your hands. Old transputer=20
guys aren't
> > ready to be put out to grass just yet...
>=20
> Hey, I=C2=B4ve got an old ATW (Atari Transputer Workstation), but never=20
got=20
> it going. So maybe I catch the train this time :-)
>=20
>=20
> --=20
> 42Bastian
>=20
> Note: SPAM-only account, direct mail to bs42@
>

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

> XMOS is interesting but based on a quick review of the architecture
> it is a fairly standard multicore SoC with a non-standard ISA. The
> chip on the board in the kit (XC-1) has only 4 cores and tops at 1600
> MIPS or 1.6 BOPS according to the XMOS.

At 400MHz, yes.

> For comparison IntellaSys makes a 40 core chip with up to 25000
> billion operations per second or 25 BOPS. The only twist is that
> these cores are Forth machines which is not surprising when see who
> is behind them.

But there are stark differences between the two:

(a) You can run C easily on the XMOS product, not on the IntellaSys product.
(b) The XMOS product has real honest-to-goodness development tools
available.
(c) The XMOS product has 256K RAM and can run 32 threads.

For me the IntellaSys' product is niche, just like Tilera's TILE64 and
Picochip's PCxxx chips. They are chips designed with a specific market
segment in mind. The XMOS product looks like it has some legs on it, but
whether it goes mainstream is another matter. If you don't like XC you can
just use plain C anyway.

At least XMOS and IntellaSys are producing cheap development kits... :-)

> Looks like multicore SoCs are beginning to invade the microcontroller
> space so it will be interesting to see how things develop. Personally
> I would love to see an ARM mutlicore SoC with 32 or more cores plus a
> hierarchical interconnect and some SRAM/Flash. But for now we have to
> contend with the dual core ARM9s/ARM7s found in most of the MP3s and
> iPods. Incidentally NXPs own LPC288X seems to be just "half" of the
> chipset used for the MP3 players, or maybe it is just a reject from a
> production with one dead core and/or Flash (just a pure speculation)?

You can put multiple Cortex-M1 cores on a FPGA. I'm not sure what you'd do
with them.

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

--- In l...@yahoogroups.com, "elektrknight" wrote:
>
> Hi,
>
> XMOS is interesting but based on a quick review of the architecture
> it is a fairly standard multicore SoC with a non-standard ISA. The
> chip on the board in the kit (XC-1) has only 4 cores and tops at 1600
> MIPS or 1.6 BOPS according to the XMOS.
>
> For comparison IntellaSys makes a 40 core chip with up to 25000
> billion operations per second or 25 BOPS. The only twist is that
> these cores are Forth machines which is not surprising when see who
> is behind them .

They are aimed at completely different applications and markets. One
use for XMOS devices is to replace FPGAs in many applications, which
is a massive market if they can crack it.

Leon

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

At 01:05 AM 9/30/2008, Paul Curtis wrote:

> > For comparison IntellaSys makes a 40 core chip with up to 25000
> > billion operations per second or 25 BOPS. The only twist is that
> > these cores are Forth machines which is not surprising when see who
> > is behind them.
>
>But there are stark differences between the two:
>
>(a) You can run C easily on the XMOS product, not on the IntellaSys product.
>(b) The XMOS product has real honest-to-goodness development tools
>available.
>(c) The XMOS product has 256K RAM and can run 32 threads.
>
>For me the IntellaSys' product is niche, just like Tilera's TILE64 and
>Picochip's PCxxx chips. They are chips designed with a specific market
>segment in mind. The XMOS product looks like it has some legs on it, but
>whether it goes mainstream is another matter. If you don't like XC you can
>just use plain C anyway.

Admit it, the real reason is just that you can't let go of your
Transputer fixation :-P So you are going dust off your transputer compiler?!

// richard (This email is for mailing lists. To reach me directly,
please use richard at imagecraft.com)
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi Richard,

> >For me the IntellaSys' product is niche, just like Tilera's TILE64 and
> >Picochip's PCxxx chips. They are chips designed with a specific market
> >segment in mind. The XMOS product looks like it has some legs on it, but
> >whether it goes mainstream is another matter. If you don't like XC you
can
> >just use plain C anyway.
>
> Admit it, the real reason is just that you can't let go of your
> Transputer fixation :-P So you are going dust off your transputer
compiler?!

That time has long passed and I'm not about to resurrect anything. :-)

But, of course, XMOS does indeed have heritage:

http://en.wikipedia.org/wiki/David_May_(computer_scientist)

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Paul Curtis schrieb:
> (a) You can run C easily on the XMOS product, not on the IntellaSys produ=
ct.
> (b) The XMOS product has real honest-to-goodness development tools
> available.
> (c) The XMOS product has 256K RAM and can run 32 threads.

I=C2=B4d say, a) and c) are the real advantage.

> You can put multiple Cortex-M1 cores on a FPGA. I'm not sure what you'd =
do
> with them.=20

The difference between homemade multicore and XMOS is IMHO the=20
interconnects.

--=20
42Bastian

Note: SPAM-only account, direct mail to bs42@...
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

----- Original Message -----
From: "42Bastian"
To:
Sent: Tuesday, September 30, 2008 9:40 AM
Subject: Re: [lpc2000] Re: OT: To all those geeks that like to play with
stuff
Paul Curtis schrieb:
> (a) You can run C easily on the XMOS product, not on the IntellaSys
> product.
> (b) The XMOS product has real honest-to-goodness development tools
> available.
> (c) The XMOS product has 256K RAM and can run 32 threads.

I´d say, a) and c) are the real advantage.

> You can put multiple Cortex-M1 cores on a FPGA. I'm not sure what you'd
> do
> with them.

The difference between homemade multicore and XMOS is IMHO the
interconnects.

----------------------

I've just put my name down for one of the $99 XC-1 kits. That has a
single-core device.

Leon
--
Leon Heller
Amateur radio call-sign G1HSM
Yaesu FT-817ND transceiver
Suzuki SV1000S motorcycle
l...@btinternet.com
http://www.geocities.com/leon_heller
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Leon schrieb:

> I've just put my name down for one of the $99 XC-1 kits. That has a
> single-core device.

From the flyer:
"The credit card sized XC-1 board is based on the four-core
XS1-G4 programmable device, and provides basic I/O ..."

So I guess even this small board will be fun hacking :-)

--
42Bastian

Note: SPAM-only account, direct mail to bs42@...

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi Leon,

> I've just put my name down for one of the $99 XC-1 kits. That has a
> single-core device.

Actually the flyer says it's a 4-core device, otherwise it would have been a ho-hum board. The 4 makes the difference IMO, that's what you're purchasing an XMOS chip for.

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

----- Original Message -----
From: "Leon"
To:
Sent: Tuesday, September 30, 2008 9:44 AM
Subject: Re: [lpc2000] Re: OT: To all those geeks that like to play with
stuff
> ----- Original Message -----
> From: "42Bastian"
> To:
> Sent: Tuesday, September 30, 2008 9:40 AM
> Subject: Re: [lpc2000] Re: OT: To all those geeks that like to play with
> stuff
> Paul Curtis schrieb:
>> (a) You can run C easily on the XMOS product, not on the IntellaSys
>> product.
>> (b) The XMOS product has real honest-to-goodness development tools
>> available.
>> (c) The XMOS product has 256K RAM and can run 32 threads.
>
> I´d say, a) and c) are the real advantage.
>
>> You can put multiple Cortex-M1 cores on a FPGA. I'm not sure what you'd
>> do
>> with them.
>
> The difference between homemade multicore and XMOS is IMHO the
> interconnects.
>
> ----------------------
>
> I've just put my name down for one of the $99 XC-1 kits. That has a
> single-core device.

It's actually got the 4-core device!

Leon
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

----- Original Message -----
From: "Paul Curtis"
To:
Sent: Tuesday, September 30, 2008 9:51 AM
Subject: RE: [lpc2000] Re: OT: To all those geeks that like to play with
stuff
> Hi Leon,
>
>> I've just put my name down for one of the $99 XC-1 kits. That has a
>> single-core device.
>
> Actually the flyer says it's a 4-core device, otherwise it would have been
> a ho-hum board. The 4 makes the difference IMO, that's what you're
> purchasing an XMOS chip for.

Yes, I just noticed that. It's a good price for 1600 MIPS! I think they've
put the chip in a reduced pin package, to keep the cost down.

I designed a transputer systems with 16 T4 or T8 modules, each with 1 Mbyte
of RAM. At 20 MIPS each that made 320 MIPS, and cost a lot of money.

Leon
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

> > This is worth pre-ordering just for the ultimate satisfaction of owning=
one
> > and feeling the throbbing power in your hands. Old transputer guys are=
n't
> > ready to be put out to grass just yet...
>=20
> Hey, I=C2=B4ve got an old ATW (Atari Transputer Workstation), but never g=
ot
> it going. So maybe I catch the train this time :-)

The ATW was over-priced, IMO. I had one for a few months, but never really=
liked it that much. I did, however, love my Atari 1040ST. For transputer=
s, I much preferred plug-in ISA cards.

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk=20
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

----- Original Message -----
From: "Paul Curtis"
To:
Sent: Tuesday, September 30, 2008 10:16 AM
Subject: RE: [lpc2000] OT: To all those geeks that like to play with stuff
Hi,

> > This is worth pre-ordering just for the ultimate satisfaction of owning
> > one
> > and feeling the throbbing power in your hands. Old transputer guys
> > aren't
> > ready to be put out to grass just yet...
>
> Hey, I´ve got an old ATW (Atari Transputer Workstation), but never got
> it going. So maybe I catch the train this time :-)

The ATW was over-priced, IMO. I had one for a few months, but never really
liked it that much. I did, however, love my Atari 1040ST. For transputers,
I much preferred plug-in ISA cards.
I designed a double-height ISA card for a company with lots of my modules on
it (I can't remember how many). Unfortunately, they didn't sell any of them.
It was intended to run some exotic CAD software and the people developing it
coundn't get it to work.

My contact at XMOS is away, I was going to ask him when they start shipping
the XC-1. I've registered for the software and membership of their forum,
anyway.

Leon
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

>> use for XMOS devices is to replace FPGAs in many applications, which
>> is a massive market if they can crack it.

NXP's next MCUs will have similar capabilities, self contained programmable logic. The real issue is the increasing use of FPGAs with soft core CPU code. I am using those as well right now in several of my FPGA designs, and eliminating many onboard MCUs. The MCU makers are worried about soft core CPUs within FPGAs eating up a big share of the market. So the MCU makers want to add prog logic to their MCUs to counter the attack.

In the coming years there will be a lot of multisexual chips that have both prog logic and CPU functionality. It may well become standard. In the future we may well view any chip that does not have both as archaic.

Chris.

[Non-text portions of this message have been removed]
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

> >> use for XMOS devices is to replace FPGAs in many applications, which
> >> is a massive market if they can crack it.
>
> NXP's next MCUs will have similar capabilities, self contained
programmable
> logic.

Really, who cares? If they could just get their damned devices into the
world without so many defects, so we didn't have to deal with the multitude
of workarounds and customer problems, it would be much better for all.

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Paul Curtis schrieb:

> If they could just get their damned devices into the
> world without so many defects, so we didn't have to deal with the multitude
> of workarounds and customer problems, it would be much better for all.

Maybe they offer CPU-Update libraries for the FPGA part :-)

--
42Bastian

Note: SPAM-only account, direct mail to bs42@...

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

>>Really, who cares? If they could just get their damned devices into the
>>world without so many defects, so we didn't have to deal with the multitude
>>of workarounds and customer problems, it would be much better for all.

Gee Paul, you sound surprisingly pissed. I'm just finishing a large design with the LPC2468 and I have to say there were very few problems with the CPU. This is my 4th NXP LPC design. Compared to the old days with PICs, 8051's, and Z80s, I am absolutely in love with the LPCs. There were many bugs with the PICs. I don't have to worry about reading all of the "gray bar" special notes like I did with the PICs. I think NXP has been pretty open about what the defects are in early spins, and they get the defects fixed pretty fast in the next chip revs. ATMEL has far more problems, and ST has problems as well. Any chips this complex will always have bugs in the early releases. I think Philips/NXP has done much better than many of the others.

Chris.

[Non-text portions of this message have been removed]
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

I've just spoken to my XMOS contact. He was about to catch a plane to the
USA so it was quite brief. They have already had some design wins, including
one in Japan for a completely unexpected application. They will be releasing
chips in small pin-count packages, including QFP or QFN, which will make
them more hobbyist-friendly.

I'd like to investigate their use in software-defined amateur radio systems.
They typically use both FPGAs and DSPs, and a single XMOS chip whould
replace both devices.

Leon
--
Leon Heller
Amateur radio call-sign G1HSM
Yaesu FT-817ND transceiver
Suzuki SV1000S motorcycle
l...@btinternet.com
http://www.geocities.com/leon_heller
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

> >>Really, who cares? If they could just get their damned devices into the
> >>world without so many defects, so we didn't have to deal with the
multitude
> >>of workarounds and customer problems, it would be much better for all.
>
> Gee Paul, you sound surprisingly pissed.

No, just frustrated. Having found a problem in LPC2148s that NXP knew about
but didn't bother to document upset me greatly. The number of
code-execution-related problems remains high.

How about answering this question. Assume you're running your code in RAM
at 72MHz with the MAM fully enabled on the LPC2378 (which avoids nasty
MAM/Flash bugs) and then call an ISP routine, what's the likelihood it
failing to program correctly? Does it make a difference which revision 2378
you're using? Would this also apply to a 2103? A 2148? You have to
contend with flash, PLL, MAM problems, and silicon revision problems. Given
we have to program all these parts, not just a specific revision of a single
device, how do you go about making sure that your flash routines *do* work
correctly?

Then you need to write the board support packages for all these and set the
PLL up for correct execution. And who know which device revision is on the
board...

--
Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

--- In l...@yahoogroups.com, "HM2" wrote:
>
> >>Really, who cares? If they could just get their damned devices
into the
> >>world without so many defects, so we didn't have to deal with the
multitude
> >>of workarounds and customer problems, it would be much better for all.
>
> Gee Paul, you sound surprisingly pissed. I'm just finishing a large
design with the LPC2468 and I have to say there were very few problems
with the CPU. This is my 4th NXP LPC design. Compared to the old
days with PICs, 8051's, and Z80s, I am absolutely in love with the
LPCs. There were many bugs with the PICs. I don't have to worry
about reading all of the "gray bar" special notes like I did with the
PICs. I think NXP has been pretty open about what the defects are in
early spins, and they get the defects fixed pretty fast in the next
chip revs. ATMEL has far more problems, and ST has problems as well.
Any chips this complex will always have bugs in the early releases.
I think Philips/NXP has done much better than many of the others.
>
> Chris.

We received the communication from NXP that they solved the MAM bug on
LPC2138 just one month ago while we started using the chip 2 years
ago: it is not what we call a 'prompt user support'.
Reading more recent posts I see that things are not changed (LCD
controller flicker problem, for example).
Regarding Z80, I never seen a problem with that processor, we still
have in field products sold more than 10 years ago.

Returning to transputer stuff: I used it an experimental Inmos board
at university, in 1990, programmed in Occam (a multitasking C like
language)
I didn't know that there were other devices (like Atari) that employed
that processor.

Carlo
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

----- Original Message -----
From: "c.barbaro"
To:
Sent: Tuesday, September 30, 2008 1:36 PM
Subject: [lpc2000] Re: OT: To all those geeks that like to play with stuff
> --- In l...@yahoogroups.com, "HM2" wrote:
>>
>> >>Really, who cares? If they could just get their damned devices
> into the
>> >>world without so many defects, so we didn't have to deal with the
> multitude
>> >>of workarounds and customer problems, it would be much better for all.
>>
>> Gee Paul, you sound surprisingly pissed. I'm just finishing a large
> design with the LPC2468 and I have to say there were very few problems
> with the CPU. This is my 4th NXP LPC design. Compared to the old
> days with PICs, 8051's, and Z80s, I am absolutely in love with the
> LPCs. There were many bugs with the PICs. I don't have to worry
> about reading all of the "gray bar" special notes like I did with the
> PICs. I think NXP has been pretty open about what the defects are in
> early spins, and they get the defects fixed pretty fast in the next
> chip revs. ATMEL has far more problems, and ST has problems as well.
> Any chips this complex will always have bugs in the early releases.
> I think Philips/NXP has done much better than many of the others.
>>
>> Chris.
>>
>> We received the communication from NXP that they solved the MAM bug on
> LPC2138 just one month ago while we started using the chip 2 years
> ago: it is not what we call a 'prompt user support'.
> Reading more recent posts I see that things are not changed (LCD
> controller flicker problem, for example).
> Regarding Z80, I never seen a problem with that processor, we still
> have in field products sold more than 10 years ago.
>
> Returning to transputer stuff: I used it an experimental Inmos board
> at university, in 1990, programmed in Occam (a multitasking C like
> language)
> I didn't know that there were other devices (like Atari) that employed
> that processor.

A friend of mine interfaced one of my modules to a Sinclair QL!

Leon
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

>
> I've just spoken to my XMOS contact. He was about to catch a plane
to the
> USA so it was quite brief. They have already had some design wins,
including
> one in Japan for a completely unexpected application. They will be
releasing
> chips in small pin-count packages, including QFP or QFN, which will
make
> them more hobbyist-friendly.

Hobbyist as a target? Not at all! A chip in 36 pin QFN with 4 400MHz
cores has some real world appeal as an embedded compute engine.
I have couple of good uses for it provided it does not cost ARM and
leg so to speak and is fairly low power.

>
> I'd like to investigate their use in software-defined amateur radio
systems.
> They typically use both FPGAs and DSPs, and a single XMOS chip
whould
> replace both devices.
>

That makes perfect sens. With this kind of core speed you could
simulate logic on some of the cores and run your control on
another one.

Or you could emulate another core like ARM with all the
peripherals, fixing any "hardware" problems would be a non issue.

Just looking at the other hardware kit XS1-G4 with QVGA display
it has a lot more appeal except for the price. Seems like a very
good emulation platform for all those game consoles out there.
> Leon
> --
> Leon Heller
> Amateur radio call-sign G1HSM
> Yaesu FT-817ND transceiver
> Suzuki SV1000S motorcycle
> leon355@...
> http://www.geocities.com/leon_heller
>

Pawel
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi,

>
> >> use for XMOS devices is to replace FPGAs in many applications,
which
> >> is a massive market if they can crack it.
>
> NXP's next MCUs will have similar capabilities, self contained
>programmable logic. The real issue is the increasing use of FPGAs
I was told about year ago by NXP engineer that they will be releasing
multicore chips, well I am still waiting.

They might have a prototype but considering NXPs current situation
it might be a while before we see these chips.
>with soft core CPU code. I am using those as well right now in
>several of my FPGA designs, and eliminating many onboard MCUs. The
>MCU makers are worried about soft core CPUs within FPGAs eating up a
>big share of the market. So the MCU makers want to add prog logic
>to their MCUs to counter the attack.
>

It goes both ways, FPGA vendors are adding hard cpu cores to their
logic chips while MCU vendors are adding programmable logic to their
SoCs. It will be interesting to see who wins this integration race.
> In the coming years there will be a lot of multisexual chips that
> have both prog logic and CPU functionality. It may well become
> standard. In the future we may well view any chip that does not
> have both as archaic.
>
> Chris.
>
>

Pawel
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

Hi Paul, Leon, et al..

Thanks for the information on the XMOS chip, I've registered to get an
XC-1 Dev Kit.

In my dim and distant past I reverse engineered the Siemens PLC I/O
bus and then controlled a rack of PLC I/O modules using custom
designed Transputer to PLC backplane interface.

I'm looking forward to plating with this board a great deal. :o)

---

PJE

------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )

>> How about answering this question. Assume you're running your code in RAM
>> at 72MHz with the MAM fully enabled on the LPC2378 (which avoids nasty
>> MAM/Flash bugs) and then call an ISP routine, what's the likelihood it
>> failing to program correctly? Does it make a difference which revision 2378
>> you're using? Would this also apply to a 2103? A 2148? You have to
>> contend with flash, PLL, MAM problems, and silicon revision problems. Given
>> we have to program all these parts, not just a specific revision of a single
>> device, how do you go about making sure that your flash routines *do* work
>> correctly?

I understand. I use IAP in all my designs and they all work. The one problem I ran into was with the 2104 which didn't work right with 24MHz xtal until I did some special things. My current LPC2468 design seems to work fine at full 72MHz Fcpu no matter how I set the MAM. The only wierdness I ran into with that part was in the EMC which requires certain reg settings they don't document. I think chips this complex are always going to have some issues, especially in early revs. But comparing what Philips does with other vendors, I think they have done quite a bit better than many others. Reading some of the posts on other MCU forums makes me glad I use LPC parts.

Chris.

[Non-text portions of this message have been removed]
------------------------------------



(You need to be a member of lpc2000 -- send a blank email to lpc2000-subscribe@yahoogroups.com )