Does ARMs support bytes?

Jim Granville wrote:

> 
> * Thumb-2 is not binary compatible with Thumb[16] or ARM[32]

Hi Jim,

Correct. You need updated development tools. That's why it is important 
that development tools vendors are involved during development of the V7 
architecture.

> * Thumb-2 removes the need to juggle/optimise your app for Size.Speed 
> across the 16 and 32 bit opcode alternatives.

That's is true for Cortex-M Series.  For Cortex-A and Cortext-R series 
processors, it will support ARM, Thumb and Thumb-2 instruction sets. So 
some users might still want to separate part of their code as ARM 
instruction set and other parts as Thumb/Thumb-2 instruction set.

(http://www.arm.com/products/CPUs/families/CortexFamily.html)

> * Thumb-2 introduces (among other things) variable length opcodes, and 
>  atomic Bit access.
> * Thumb-2 also adds opcodes for 16 bit constant load,
> short skips [IF THEN], bit fields, bit reversal, TestZ&Branch,
> TableBranch, improved interrupt handling (HW Stacking).....
> 

I am not sure how much I can tell you (in case I upset the marketing 
guys ;-) So I better not to detail these areas.

>  They state all this gives you a 5% smaller and 2-3% faster code than 
> Thumb. [presumably those values are averages, maybe even 
> 'marketing-filtered' averages :) ? ]
> 

Time will tell ;)  And don't forget that the core size is smaller (lower 
power consumption). For more advantages of Cortex-M3 core please see
http://www.arm.com/products/CPUs/ARM_Cortex-M3.html

>  Variable length opcodes, atomic Bit access, TestZ&Branch will be very 
> familiar to 80C51 users. Thus yes, Cortex will be a better 
> MicroController core than earlier ARMs.
> 

This one I am sure :)
But don't forget it is called Cortex-M3.  (See next answer)

> Comment:
>  There is likely to be some confusion in the field, between ARMv7, and 
> ARM7.
>   Reminds me of the Philips XA51, where Philips tried to coat-tail on 
> the large 80C51 market, with a device that was NOT binary compatible.
> 
>  Time will tell how Cortex flies.
> 
> <paste>

It will certainly takes time for the market to adapt to the new naming 
system.  Instead of calling it ARM12, the development has now divered to 
three areas and come with a new naming system:
Applications   - Cortex-A
Real time      - Cortex-R
Microcontroller- Cortex-M
See  http://www.arm.com/products/CPUs/families/CortexFamily.html
In this way it should help to reduce confusion between V7 architecture 
and ARM7 core.

<Personal Opinion>
I guess people will get soon lazy and just call Cortex-M3 as "M3"
(This reminded me of the M3 motorway in UK  8-)
</Personal Opinion>

> 
> 
>  Hmmm, well one of those papers is dated June 2003, and the Opcodes are 
> defined very early in a processor design. Releasing opcode info is also 
> just a simple text file, so consumes very little actual publishing effort.
>  Of course the silicon follows some time later, but designers can decide 
> what to place on their radar much earlier in a life cycle.
> 
> -jg
> 

The difficult part is to define the corner cases behaviours.  These 
needed to be clearly documented.

Hope this information helps.  If there is any further questions, please 
post it on comp.sys.arm.  There are more ARM people reading that 
newsgroup and they can help answer questions.

Joseph

This e-mail message is intended for the addressee(s) only
and may contain information that is the property of, and/or
subject to a confidentiality agreement between the intended
recipient(s), their organisation and/or the ARM Group of
Companies. If you are not an intended recipient of this
e-mail message, you should not read, copy, forward or
otherwise distribute or further disclose the information
in it; misuse of the contents of this e-mail message may
violate various laws in your state, country or jurisdiction.
If you have received this e-mail message in error, please
contact the originator of this e-mail message via e-mail
and delete all copies of this message from your computer
or network, thank you.

Joseph wrote:

> ...
> Correct. You need updated development tools. That's why it is 
> important that development tools vendors are involved during 
> development of the V7 architecture.

I guess we are too small for ARM Inc. to notice for now. I asked several 
times for NDA and info but to no avail. Oh well. I am betting that in a 
year's time, a good number of non-GNU users for ARM7 MCU will be using 
our compilers (fully functional for 45 days, 10K limited afterward w/o 
asking us for a license or anything). May be we can get some info on the 
Thumb-2 then :-)

-- 
// richard
http://www.imagecraft.com

Richard M. wrote:
> Joseph wrote:
> 
>> ...
>> Correct. You need updated development tools. That's why it is 
>> important that development tools vendors are involved during 
>> development of the V7 architecture.
> 
> 
> I guess we are too small for ARM Inc. to notice for now. I asked several 
> times for NDA and info but to no avail. Oh well. I am betting that in a 
> year's time, a good number of non-GNU users for ARM7 MCU will be using 
> our compilers (fully functional for 45 days, 10K limited afterward w/o 
> asking us for a license or anything). May be we can get some info on the 
> Thumb-2 then :-)
> 

Hi Richard,

I really shouldn't comment on this non-technical issue ;)
<My own opinion>
Please check if any of your current customers will be using Cortex.
If there is any, try push the ARM sale guys saying that you need to 
support your customer.  Also it should be much easier now as the V7 
architecture is officially released.
</My own opinion>

Joseph

This e-mail message is intended for the addressee(s) only
and may contain information that is the property of, and/or
subject to a confidentiality agreement between the intended
recipient(s), their organisation and/or the ARM Group of
Companies. If you are not an intended recipient of this
e-mail message, you should not read, copy, forward or
otherwise distribute or further disclose the information
in it; misuse of the contents of this e-mail message may
violate various laws in your state, country or jurisdiction.
If you have received this e-mail message in error, please
contact the originator of this e-mail message via e-mail
and delete all copies of this message from your computer
or network, thank you.

On Wed, 16 Mar 2005 13:59:54 +0000, Joseph
<joseph.yiu@somewhere-in-arm.com> wrote:

><My own opinion>
>Please check if any of your current customers will be using Cortex.
>If there is any, try push the ARM sale guys saying that you need to 
>support your customer.  Also it should be much easier now as the V7 
>architecture is officially released.
></My own opinion>
>
>Joseph

Joseph, it seems to me that it should be enough to hear that Richard
is willing to apply his own personal time into this business area.  He
already exists, is making his own way in the world, has a demonstrated
level of competence, and is asking for very little by way of help in
order to invest a serious amount of his own time.  Would you imagine
any rational person would do that, without any prospects of evidence
or belief that the time will pay off??

It seems that Richard may have more confidence in the future of this
processor than ARM's own people do.

If I were Richard, I might be a little offended at the suggestion that
I might have such bad business sense and hadn't already found that
there was customer interest in the new CPU core before committing
myself to the idea of fielding a new target for my compiler tool.

I know you aren't marketing or sales here.  And I'm only yelling at
you because you are here to hear it.  You cannot do anything about
this.  But it does shock me to the core, Joseph, that ARM acts this
way towards a legitimate compiler vendor.  (It's the kind of thing
that makes me wonder if they have a personal axe to grind in the
compiler arena or hidden associations they are trying to protect.)

I see no excuse, none, for Richard to have experienced what he has
from ARM, enough to say, "I asked several times for NDA and info but
to no avail."

Jon

Jonathan Kirwan wrote:
> 
> 
> Joseph, it seems to me that it should be enough to hear that Richard
> is willing to apply his own personal time into this business area.  He
> already exists, is making his own way in the world, has a demonstrated
> level of competence, and is asking for very little by way of help in
> order to invest a serious amount of his own time.  Would you imagine
> any rational person would do that, without any prospects of evidence
> or belief that the time will pay off??
> 
> It seems that Richard may have more confidence in the future of this
> processor than ARM's own people do.
> 
> If I were Richard, I might be a little offended at the suggestion that
> I might have such bad business sense and hadn't already found that
> there was customer interest in the new CPU core before committing
> myself to the idea of fielding a new target for my compiler tool.
> 
> I know you aren't marketing or sales here.  And I'm only yelling at
> you because you are here to hear it.  You cannot do anything about
> this.  But it does shock me to the core, Joseph, that ARM acts this
> way towards a legitimate compiler vendor.  (It's the kind of thing
> that makes me wonder if they have a personal axe to grind in the
> compiler arena or hidden associations they are trying to protect.)
> 
> I see no excuse, none, for Richard to have experienced what he has
> from ARM, enough to say, "I asked several times for NDA and info but
> to no avail."
> 
> Jon

Hi Jonathan, Richard,

I apologise if I offend anyone here. Hope Richard is not offended by my 
previous comments.

And thanks for your comment Jonathan.

There are large number of issues about releasing information in this 
business which are undisclosed to public. Especially when the 
information could be linked to intellectual properties or patents of ARM 
or 3rd parties.  So potentially the NDA you have to sign is not only 
one, but a number of NDA from different companies. Or if patent issue is 
found, we have to stop signing new NDA immediately until problem is solved.

As you understand there is nothing I can do (or shouldn't even comment 
in public) about how marketing people do their work. But in general we 
are trying our best to provide information the best we can. I am not in 
the position of handling enquiries of this nature, but you might try 
write to Press Contacts staffs (http://www.arm.com/news/contacts.html) 
to see if they have idea about how you can get hold of the required 
information. I am quite certain that ARM will release instruction set 
details to public, but possibly not on ARM website (as you might noticed 
the current ARM architecture reference manual is published by ADDISON 
WESLEY, and it not available on ARM web site).

If you don't mind I hope to finish this news thread here.
If anyone got questions regarding the Cortex or V7 architecture, it is 
better to post it on comp.sys.arm as there are more ARM people reading 
that newsgroup so you could get better answer from there.

Hope this helps.

Joseph

This e-mail message is intended for the addressee(s) only
and may contain information that is the property of, and/or
subject to a confidentiality agreement between the intended
recipient(s), their organisation and/or the ARM Group of
Companies. If you are not an intended recipient of this
e-mail message, you should not read, copy, forward or
otherwise distribute or further disclose the information
in it; misuse of the contents of this e-mail message may
violate various laws in your state, country or jurisdiction.
If you have received this e-mail message in error, please
contact the originator of this e-mail message via e-mail
and delete all copies of this message from your computer
or network, thank you.

Jonathan Kirwan wrote:
> On Wed, 16 Mar 2005 13:59:54 +0000, Joseph
> <joseph.yiu@somewhere-in-arm.com> wrote:
> 
> 
>><My own opinion>
>>Please check if any of your current customers will be using Cortex.
>>If there is any, try push the ARM sale guys saying that you need to 
>>support your customer.  Also it should be much easier now as the V7 
>>architecture is officially released.
>></My own opinion>
>>
>>Joseph
> 
> 
> Joseph, it seems to me that it should be enough to hear that Richard
> is willing to apply his own personal time into this business area.  He
> already exists, is making his own way in the world, has a demonstrated
> level of competence, and is asking for very little by way of help in
> order to invest a serious amount of his own time.  Would you imagine
> any rational person would do that, without any prospects of evidence
> or belief that the time will pay off??
> 
> It seems that Richard may have more confidence in the future of this
> processor than ARM's own people do.
> 
> If I were Richard, I might be a little offended at the suggestion that
> I might have such bad business sense and hadn't already found that
> there was customer interest in the new CPU core before committing
> myself to the idea of fielding a new target for my compiler tool.
> 
> I know you aren't marketing or sales here.  And I'm only yelling at
> you because you are here to hear it.  You cannot do anything about
> this.  But it does shock me to the core, Joseph, that ARM acts this
> way towards a legitimate compiler vendor.  (It's the kind of thing
> that makes me wonder if they have a personal axe to grind in the
> compiler arena or hidden associations they are trying to protect.)
> 
> I see no excuse, none, for Richard to have experienced what he has
> from ARM, enough to say, "I asked several times for NDA and info but
> to no avail."

  Perhaps Richard should ask again, and report back ?
  He could also ask whom he could talk with to get engineering samples
of Coretex silicon ( because ARM will know who has this 'nearly out' )- 
some of the docs were from mid 2003, so that indicates
ES silicon is around about now....
  -jg

Joseph wrote:

 > Jonathan Kirwan wrote:
 >
 >>
 >>
 >> Joseph, it seems to me that it should be enough to hear that Richard
 >> is willing to apply his own personal time into this business area.  He
 >> already exists, is making his own way in the world, has a demonstrated
 >> level of competence, and is asking for very little by way of help in
 >> order to invest a serious amount of his own time.  Would you imagine
 >> any rational person would do that, without any prospects of evidence
 >> or belief that the time will pay off??
 >>
 >> It seems that Richard may have more confidence in the future of this
 >> processor than ARM's own people do.
 >>
 >> If I were Richard, I might be a little offended at the suggestion that
 >> I might have such bad business sense and hadn't already found that
 >> there was customer interest in the new CPU core before committing
 >> myself to the idea of fielding a new target for my compiler tool.
 >>
 >> I know you aren't marketing or sales here.  And I'm only yelling at
 >> you because you are here to hear it.  You cannot do anything about
 >> this.  But it does shock me to the core, Joseph, that ARM acts this
 >> way towards a legitimate compiler vendor.  (It's the kind of thing
 >> that makes me wonder if they have a personal axe to grind in the
 >> compiler arena or hidden associations they are trying to protect.)
 >>
 >> I see no excuse, none, for Richard to have experienced what he has
 >> from ARM, enough to say, "I asked several times for NDA and info but
 >> to no avail."
 >>
 >> Jon
 >
 >
 >
 > Hi Jonathan, Richard,
 >
 > I apologise if I offend anyone here. Hope Richard is not offended by 
my previous comments.


  Joseph, I am sure Jon was not offended by yourself, but by the low 
technical merit in ARMs policies..

 > And thanks for your comment Jonathan.
 >
 > There are large number of issues about releasing information in this 
business which are undisclosed to public. Especially when the 
information could be linked to intellectual properties or patents of ARM 
or 3rd parties.  So potentially the NDA you have to sign is not only 
one, but a number of NDA from different companies. Or if patent issue is 
found, we have to stop signing new NDA immediately until problem is solved.
 >
 > As you understand there is nothing I can do (or shouldn't even 
comment in public) about how marketing people do their work. But in 
general we are trying our best to provide information the best we can. I 
am not in the position of handling enquiries of this nature, but you 
might try write to Press Contacts staffs 
(http://www.arm.com/news/contacts.html) to see if they have idea about 
how you can get hold of the required information. I am quite certain 
that ARM will release instruction set details to public, but possibly 
not on ARM website (as you might noticed the current ARM architecture 
reference manual is published by ADDISON WESLEY, and it not available on 
ARM web site).


.. and I can see no technical reason for this chestnut either!
Could there possibly be some obtuse legal reason for this ?

 > If you don't mind I hope to finish this news thread here.
 > If anyone got questions regarding the Cortex or V7 architecture, it 
is better to post it on comp.sys.arm as there are more ARM people 
reading that newsgroup so you could get better answer from there.


a) I thought this was not publicly released info ?
b) Nice to see you anticipate Cortex to have zero relevence to embedded 
apps & designers ?

-jg

On Wed, 16 Mar 2005 18:08:05 +0000, Joseph
<joseph.yiu@somewhere-in-arm.com> wrote:

>Jonathan Kirwan wrote:
>> 
>> 
>> Joseph, it seems to me that it should be enough to hear that Richard
>> is willing to apply his own personal time into this business area.  He
>> already exists, is making his own way in the world, has a demonstrated
>> level of competence, and is asking for very little by way of help in
>> order to invest a serious amount of his own time.  Would you imagine
>> any rational person would do that, without any prospects of evidence
>> or belief that the time will pay off??
>> 
>> It seems that Richard may have more confidence in the future of this
>> processor than ARM's own people do.
>> 
>> If I were Richard, I might be a little offended at the suggestion that
>> I might have such bad business sense and hadn't already found that
>> there was customer interest in the new CPU core before committing
>> myself to the idea of fielding a new target for my compiler tool.
>> 
>> I know you aren't marketing or sales here.  And I'm only yelling at
>> you because you are here to hear it.  You cannot do anything about
>> this.  But it does shock me to the core, Joseph, that ARM acts this
>> way towards a legitimate compiler vendor.  (It's the kind of thing
>> that makes me wonder if they have a personal axe to grind in the
>> compiler arena or hidden associations they are trying to protect.)
>> 
>> I see no excuse, none, for Richard to have experienced what he has
>> from ARM, enough to say, "I asked several times for NDA and info but
>> to no avail."
>> 
>> Jon
>
>Hi Jonathan, Richard,
>
>I apologise if I offend anyone here. Hope Richard is not offended by my 
>previous comments.
>
>And thanks for your comment Jonathan.

I've no horse in this race, Joseph.  So I just tried putting myself in
Richard's place and wondering about how I'd feel then.  When I took
that simple step, I noticed I had some very strong feelings and so I
wrote them out.  (Since Richard has a business here, it might be
somewhat more difficult for him to write publicly about such a
reaction, if he had one.)

I know you had NO intention at all of harm and were only trying to
think about constructive ideas to suggest.  So I don't mean to
suggest, in any way at all, that there was anything about your
comments other than trying to be as helpful as you could.

>There are large number of issues about releasing information in this 
>business which are undisclosed to public. Especially when the 
>information could be linked to intellectual properties or patents of ARM 
>or 3rd parties.  So potentially the NDA you have to sign is not only 
>one, but a number of NDA from different companies. Or if patent issue is 
>found, we have to stop signing new NDA immediately until problem is solved.

I know.  But entering into such hidden relationships carries with it a
certain risk, too, in terms of outside perceptions.  At a minimum.
And I don't much like the sound of the words you had to write above to
remain accurate and honest with us.

It would be better if all relationships ARM has are publicly available
and readable.  I don't mean *before* they are entered into, as that
may be a critical negotiation period where the need for secrecy makes
sense.  But after the fact, yes.  Everything above board and in the
clear.  That would help a great deal when others must make and risk
their own situations in aligning themselves with some initiative they
want to promote.

>As you understand there is nothing I can do (or shouldn't even comment 
>in public) about how marketing people do their work.

Totally understood!  I empathize that you had no inclination that I
might write what I did in response to an otherwise "helpful" comment
to Richard and now things rapidly spinning in a difficult direction.

>But in general we 
>are trying our best to provide information the best we can.

That's a given, of course.  Who wouldn't?  The question arises whether
or not the _right_ decisions are being made, in practice.  And there,
people can disagree.

>I am not in 
>the position of handling enquiries of this nature, but you might try 
>write to Press Contacts staffs (http://www.arm.com/news/contacts.html) 
>to see if they have idea about how you can get hold of the required 
>information.

Another possible tip for Richard, I suppose.  I've a hunch it's not
going to open doors for Richard, though, that he hasn't already
knocked on.

>I am quite certain that ARM will release instruction set 
>details to public, but possibly not on ARM website (as you might noticed 
>the current ARM architecture reference manual is published by ADDISON 
>WESLEY, and it not available on ARM web site).

It seems to me that this is great for people (like me) who would
consider the idea of using a product after it is released.  But of
course, this will be of little help to people like Richard who need to
be out "priming the pump" a little, before the rest of us start
looking for a compiler tool.  If he doesn't have it ready when I'm
looking, and I buy someone else's tool and invest my time in that,
then he has forever lost out on the opportunity.  Bad news.

>If you don't mind I hope to finish this news thread here.

I'm sorry I felt the inclination to write further.  But I did.
However, I fully understand why it makes little sense for you to
pursue this line here.  I might exactly the same thing in your shoes
and I don't expect a reply -- nor should you provide one.

>If anyone got questions regarding the Cortex or V7 architecture, it is 
>better to post it on comp.sys.arm as there are more ARM people reading 
>that newsgroup so you could get better answer from there.

And not here?  Aren't those people in comp.sys.arm doing embedded
work?  Have they no presence here?

>Hope this helps.
>
>Joseph

My own 'resonances' began only because I wanted you to see and feel
how your last suggestion might have come across, if you put yourself
into the seat of someone who is putting their money where their mouth
is at.  I've been self-employed most of my adult life and have had to
worry about these kinds of things and there never is a day that goes
by that you aren't continually aware of the details of that business.
If you forget it, you find yourself looking for employment fast.

There is NO question at all in my mind that Richard wouldn't have even
bothered asking in the first place, if there wasn't already very good
reason for ARM to provide him with the information.  It's not like
Richard is some no-account, wanna-be with no track record to look at.

If I were him, I might take it kind of personal.  But perhaps that's
just me.

Jon

> As you understand there is nothing I can do (or shouldn't even comment
> in public) about how marketing people do their work. But in general we
> are trying our best to provide information the best we can. I am not
> in the position of handling enquiries of this nature,

Yes you are, you could go up some floors to the Press contacts
department and inform them that Richard needs the information.
You could try to find out who the local ARM representative is
and make a personal phone call and convince them that it is in ARMs,
best interest that they release the information to Richard.
If you have no clout in ARM, then you might find some reasonable
fellow with some clout, convince him/her and let them do the job.

> Joseph
>

-- 
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB

[snip all for brevity :-) ]

Thanks Jonanthan and Ulf for your support and thanks Joseph for the 
info. First of all, as a general statement, we small 3rd party vendors 
get all sorts of responses from silicon manufacturers. For example, we 
have good experience with Atmel and their product mixes are good for our 
target audience so we try to support them as much as possible. I can't 
say that for everyone else. Some vendors practically treat us like a 
disease :-)

As for Cortex/Thumb-2 information in particular, I inquired the ARM 3rd 
party vendor relation a few months ago, after I found out about Thumb-2 
in the ARM conference, but no reply on how I can even get such an NDA or 
NDAs. Some ARM marketing guy also responded to my request form I filled 
out in the conference and I asked for the NDA but never gotten a reply 
from him either.

As for why we would want such information. Obviously if no silicon 
manufacturers jump on the bandwagon, then no, we won't be interested 
:-). Fundamentally though, it takes at least 6 months to get a new 
compiler out, and the more advance information we have, the better we 
can plan. It's a constantly changing market place and as a business 
person, supporting a new target is a big risk for us. I wish I can say I 
am betting 100% but I have not. I understand ARM Inc. wishes to protect 
its IP in terms of patent applications and such, but I can't say I am 
not frustrated either.

The ARM compiler is a big gamble for us. Notice that the existing 
vendors: Greenhills, ARM, IAR, Code Warrior have been in the ARM market 
for a while, Keil and us are the only vendors writing our own compilers 
recently. We take a risk of not jumping on the bandwagon on wrapping an 
IDE on top of GNU for a variety of reasons, and obviously we think we 
will be successful. We will see how it pan out  in the next year. I see 
a new group of customers going for the ARM7 MCU, so that's our initial 
audience. We will then move on to support other ARM variants. As 
business grows, we will continue to improve our code performance and 
look into the high end applications as well.Anyone can google 
groups.google.com and see that we started out by selling an HC11 
compiler for $39!!! While we are not a "big" embedded compiler 
powerhouse, we have craved out a nice niche with good customer 
reputation.. So we are doing pretty OK

-- 
// richard
http://www.imagecraft.com