EmbeddedRelated.com
Forums
Memfault Beyond the Launch

ARM IDE & kit similar to CCS for PIC's?

Started by The_Todd August 12, 2007
In article <13c62n1lkr8124a@corp.supernews.com>, Paul Curtis 
<plc@rowley.co.uk> writes
>"Chris Hills" <chris@phaedsys.org> wrote in message >news:<mnjjV9AW5AwGFA1A@phaedsys.demon.co.uk>... >> In article <1186961538.421108.206970@q75g2000hsh.googlegroups.com>, >> larwe <zwsdotcom@gmail.com> writes >> >On Aug 12, 6:17 pm, The_Todd <toddb...@gmail.com> wrote: >> >> My question is: Is there anything similar to the CCS IDE for the ARM >> >> controllers? A reasonably priced ($300) well supported and documented >> >> thats easy to use IDE that supports JTAG etc...? I'm not interested in
>> >Hard to understand why you are not "interested" in gcc when a large >> >percentage of the world's fielded ARM code was built with it - but >> >anyway, look at Rowley's CrossWorks for ARM. A personal license is >> >$149, an education license $299. I am using it with Rowley's own USB >> >JTAG adapter and it works well, I assume it works just as well with >> >other supported adapters. > >> Rowley's compiler IS the GCC compiler. You are paying for their IDE
> >
>As I keep pointint out, only for ARM,
I thought we were discussing your ARM suite.
>not for MSP430, AVR, MAXQ.
Enough with the commercials already ... :-)
> With a >$149 personal license for ARM you get a non-commercial version of the
What is a non-commercial version?
>software *and* a piece of hardware, the CrossConnect Lite, thrown in so you >can do standard debugging. And you're not just "paying for their IDE", you >are paying for the IDE, the library,
You do your own library?
> and all the CPU and board support >packages we have written
Most compilers come with a multitude of BSP's and projects even on the free versions of the other commercial compilers.
> and access to our support helpdesk. If you only >got the IDE, I'd be disappointed.
Point taken but I thought support for GCC was free from the community? Or do people need support from you because you use a different library etc? Regards Chris -- \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ \/\/\/\/\ Chris Hills Staffs England /\/\/\/\/ /\/\/ chris@phaedsys.org www.phaedsys.org \/\/\ \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
On Wed, 15 Aug 2007 17:23:59 +0100, Chris Hills <chris@phaedsys.org>
wrote:

><snip> >Point taken but I thought support for GCC was free from the community?
There are a number of companies that sell configured packages of software that include GCC and GAS at their core. For example: http://www.microcross.com/html/about_us.html "Microcross is the leading international distributor of open-source embedded development tools based on GNU technology." Frankly, I've purchased the tools from Microcross to save time. At the time, my price was about $150 for the set. (That was four years back, I think.) I don't believe any of the configured tools I got were their own product and they were not limited to non-commercial use -- so long as I selected the right library to include. That project never did see the light of day, as we were evaluating a processor and tools at the time and the project didn't get fully funded in the end. So I can't speak from the point of view of a completed project.
>Or do people need support from you because you use a different library >etc?
I think one possible value of a 'different library' from a vendor like this is that you don't have to worry about the licensing. You have paid __them__ to vet this issue for you. Otherwise, you can be sure they would be quickly confronted about the license violation in their package. In using GPL'd tools, one has to be careful and it's nice that someone else is thinking about these issues, too, who is a bit more of an expert about it than most of us application programmers often are. Another is that when competing with other companies, the library can have a material impact on the performance of the compiler output, on the whole, and competing will generally mean some effort will go into providing a more effective library than what is freely available without restriction to GCC. None of this is a guarantee, of course. But it is not an unreasonable expectation. Jon P.S. I have recently been using Rowley's toolset for the ARM (LPC2148, in my case) and he was very generous with me in the process of getting started, so realize that I am speaking from an existing, friendly relationship: Rowley offers a strong, value-added package for the ARM, Paul is very technically informed on myriad details as well as the marketplace, and Rowley is one of the better service-oriented businesses I've worked with. His offer of tools for $149 for non-commercial use, including the JTAG unit he provides, is fine. It's about what you'd expect to pay for a with-demo-board-only special version of JLINK (or ULINK-2) or other modest JTAG device. His JTAG tool includes a feature you may not get with those and may need -- following variable clock rates under JTAG debugging, I think (but my knowledge is not comprehensive here, so take this with a grain of salt.) And you get non-commercial, unlimited use of a pretty easy to use toolset. Looking backwards, if I were doing hobby work with ARM processors, I wouldn't bat an eye about accepting that deal knowing what I now know. The struggle in learning about and the cost in getting a good JTAG unit along with all the rest would be short-circuited and at a price that is probably no different than I'd otherwise spend, anyway. Paul is offering more than just good value, in my opinion, to hobbyists. Jon
In article <13c635o23l2l04c@corp.supernews.com>, Paul Curtis says...
> > "Robert Adsett" <sub2@aeolusdevelopment.com> wrote in message > news:MPG.21297ae966da7ae398979f@free.teranews.com... > > In article <1186966571.106883.155100@k79g2000hse.googlegroups.com>, > > larwe says... > >> On Aug 12, 7:37 pm, Robert Adsett <s...@aeolusdevelopment.com> wrote: > >> > >> > Of course CrossWorks for ARM is GCC plus their IDE and debugger. There > >> > >> It used to be, but ISTR last time I spoke with Paul @ Rowley that this > >> was no longer true - it's now using their own compiler engine. > > > > Their website says GCC. They do supply libraries of course, including a > > C runtime library. > > > > It looks as if ImageCraft's is cheaper, $200-$550 as compared to > > Rowely's $1000 for development versions. > > > > Robert > > $995 is the CrossWorks commercial version. $149 is the CrossWorks hobbyist > version with hardware thrown in. > > Comparing like with like with the top-end commercial tools, you will also > need a debugger with the ImageCraft product, another $159. Ok, and if you > want the ImageCraft compiler, a debugger, and the a USB JTAG adapter that's > an additional $429 which brings the total to$978. This is a fair > comparison. > > Comparing a hobbyist purchase, you can get a CrossConnect USB JTAG adapter > and CrossWorks software from us for $149, but ImageCraft is, oh, $199 + $429 > which is $628 for the same thing. > > So, in the round, I fail to see why ImageCraft works out cheaper, comparing > like for like.
I was comparing based on the licensed for producing code for use in a product. The other items certainly add value, but it's not necessarily what everyone is looking for. Whether they should be is another question. Robert -- Posted via a free Usenet account from http://www.teranews.com
"Jonathan Kirwan" <jkirwan@easystreet.com> wrote in message news:lui6c35vkht9r8uoo5e50iv4qafc7j60g2@4ax.com...
> On Wed, 15 Aug 2007 17:23:59 +0100, Chris Hills <chris@phaedsys.org> > wrote:
>>Or do people need support from you because you use a different library >>etc? > > I think one possible value of a 'different library' from a vendor like > this is that you don't have to worry about the licensing. You have > paid __them__ to vet this issue for you. Otherwise, you can be sure > they would be quickly confronted about the license violation in their > package. In using GPL'd tools, one has to be careful and it's nice > that someone else is thinking about these issues, too, who is a bit > more of an expert about it than most of us application programmers > often are. Another is that when competing with other companies, the > library can have a material impact on the performance of the compiler > output, on the whole, and competing will generally mean some effort > will go into providing a more effective library than what is freely > available without restriction to GCC. None of this is a guarantee, of > course. But it is not an unreasonable expectation.
The licensing is definitely an issue. Even newlib has some restrictions despite not being (L)GPL. However the killer problem is that the freely available libraries are huge and slow(*) and so totally unsuitable for embedded systems. As you say, you can get a lot more out of a compiler by using high quality libraries. AFAIK most companies selling GCC based tools use their own (closed source) libraries. Wilco (*) As an example, long long divide in GCC's libraries takes in total 5KBytes on ARM. My fast version (ie. performance, not codesize optimised) takes 528 bytes...
"Wilco Dijkstra" <Wilco_dot_Dijkstra@ntlworld.com> wrote in message 
news:zSLwi.7447$ka7.1466@newsfe4-gui.ntli.net...
> > "Jonathan Kirwan" <jkirwan@easystreet.com> wrote in message > news:lui6c35vkht9r8uoo5e50iv4qafc7j60g2@4ax.com... >> On Wed, 15 Aug 2007 17:23:59 +0100, Chris Hills <chris@phaedsys.org> >> wrote: > >>>Or do people need support from you because you use a different library >>>etc? >> >> I think one possible value of a 'different library' from a vendor like >> this is that you don't have to worry about the licensing. You have >> paid __them__ to vet this issue for you. Otherwise, you can be sure >> they would be quickly confronted about the license violation in their >> package. In using GPL'd tools, one has to be careful and it's nice >> that someone else is thinking about these issues, too, who is a bit >> more of an expert about it than most of us application programmers >> often are. Another is that when competing with other companies, the >> library can have a material impact on the performance of the compiler >> output, on the whole, and competing will generally mean some effort >> will go into providing a more effective library than what is freely >> available without restriction to GCC. None of this is a guarantee, of >> course. But it is not an unreasonable expectation. > > The licensing is definitely an issue. Even newlib has some restrictions > despite > not being (L)GPL. However the killer problem is that the freely available > libraries > are huge and slow(*) and so totally unsuitable for embedded systems. As > you say, > you can get a lot more out of a compiler by using high quality libraries. > AFAIK > most companies selling GCC based tools use their own (closed source) > libraries.
Licensing is certianly an issue. Although we use GCC as the compiler and binutils as the underpinnings of linking, the libraries, board support packages, flash code, and everything else is certainly not covered by any style of GPL/LGPL license because we wrote the rest in-house with properly licensed software (Visual Studio and Qt). With a Personal (non-commercial) license you can use our tools to write non-commercial software. We make no claims on the source code you write and you are free to distribute the sources to others if you wish. We provide the source code for GCC and binutils and the patches we've made so you can rebuild the tools should you need to. We have already been through the issues involved with the FSF; I don't think we're any different from other companies in some regards. I can't speak for Microcross, I do not know whether they provide their own closed source libraries or not. The issue for commercial customers is that they can purchase (from us) a standard Commercial licence for our tools and know that their end application is not covered by any GPL or LGPL license. If we had used LGPL code in the library we distribute to customers then they would be bound to provide relinkable object code. This is obviously undesirable for many customers and, gee, they may not even be aware of the mess they could land themselves in. -- Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
"Chris Hills" <chris@phaedsys.org> wrote in message 
news:rHxg9lJfiywGFA2K@phaedsys.demon.co.uk...
> In article <13c62n1lkr8124a@corp.supernews.com>, Paul Curtis > <plc@rowley.co.uk> writes >>"Chris Hills" <chris@phaedsys.org> wrote in message >>news:<mnjjV9AW5AwGFA1A@phaedsys.demon.co.uk>... >>> In article <1186961538.421108.206970@q75g2000hsh.googlegroups.com>, >>> larwe <zwsdotcom@gmail.com> writes >>> >On Aug 12, 6:17 pm, The_Todd <toddb...@gmail.com> wrote: >>> >> My question is: Is there anything similar to the CCS IDE for the ARM >>> >> controllers? A reasonably priced ($300) well supported and documented >>> >> thats easy to use IDE that supports JTAG etc...? I'm not interested >>> >> in > >>> >Hard to understand why you are not "interested" in gcc when a large >>> >percentage of the world's fielded ARM code was built with it - but >>> >anyway, look at Rowley's CrossWorks for ARM. A personal license is >>> >$149, an education license $299. I am using it with Rowley's own USB >>> >JTAG adapter and it works well, I assume it works just as well with >>> >other supported adapters. >> >>> Rowley's compiler IS the GCC compiler. You are paying for their IDE > > > > >>As I keep pointint out, only for ARM, > > I thought we were discussing your ARM suite.
You made a statement that the Rowley Compiler is GCC. In fact, it's not--the "Rowley Compiler" is something that I wrote and which generates code for a number of targets, not only the ones that we've turned into products. I made the decision that I did not wish to write an ARM code generator because I had already done so for Rowley Modula-2 and didn't feel the need to do it again. There was also another very good reason to integrate GCC rather than our own toolset into CrossWorks, and that reason has never seen the light of day outside of Rowley Associates and a single silicon company. A more accurate statement is that "The compiler provided with CrossWorks for ARM is derived from the GNU Compiler Collection".
>> With a >>$149 personal license for ARM you get a non-commercial version of the > > What is a non-commercial version?
I won't recite the agreement, but simply put, it states that you will not derive a profit from the use of our tools. This covers hobbyist, educational, and research use of our tools. Now, in the case of CrossWorks for ARM, we are not restricting the use of GCC and binutils, far from it--you can use those freely. However, if you wanted to do that, you would not need to purchase a Personal License--you just download and use GCC and binutls from us or elsewhere. What you can't do is use our IDE and you can't use our libraries without purchasing CrossWorks because those are covered by our own license.
>>software *and* a piece of hardware, the CrossConnect Lite, thrown in so >>you >>can do standard debugging. And you're not just "paying for their IDE", >>you >>are paying for the IDE, the library, > > You do your own library?
Of course. This is clearly stated as one of the benefits of using our tools. In this way commercial customers can be sure that their application will not be covered by a GPL or LGPL license inherited from the inclusion of GPL/LGPL code in the library they link in.
>> and all the CPU and board support >>packages we have written > > Most compilers come with a multitude of BSP's and projects even on the > free versions of the other commercial compilers.
...but I'm pretty much sure they don't offer the range of BSPs that we provide. I regularly survey the development tools landscape. The fact that anybody can write and distribute a BSP for CrossWorks, which you can just plug in, pretty much nails us as fairly unique. We have customers that are developing proprietary ARM chips using CrossWorks and writing their own BSPs.
>> and access to our support helpdesk. If you only >>got the IDE, I'd be disappointed. > > Point taken but I thought support for GCC was free from the community?
Whoever said that we only provide GCC support? Free support is worth every penny, but there is an expectation from a customer that has paid for our tools to receive support from the provider, not seek support from the community. We do not take customer's money and tell them to seek help elsewhere--there is an undertaking that we will support them. To be frank, GCC support is the least of our problems. I would say that 95% of customer problems are related to operation of hardware, NOT software, from the processor right through the JTAG adapter to the OS it runs on. This is the reason that we are now distributing a low-cost JTAG adapter with each CrossWorks license purchase is because that adapter runs our own code and we then have zero excuse for any failings. In fact, the BIGGEST, hairyest problem is the way in which ARMs reset and control of them is gained after reset.
> Or do people need support from you because you use a different library > etc?
Most customers are new to ARM and want something easy to use. We provide that. Experience ranges from complete newbies to hardened engineers. Everybody has their own set of support requirements. -- Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
"Jonathan Kirwan" <jkirwan@easystreet.com> wrote in message 
news:lui6c35vkht9r8uoo5e50iv4qafc7j60g2@4ax.com...
> On Wed, 15 Aug 2007 17:23:59 +0100, Chris Hills <chris@phaedsys.org> > wrote:
< snip>
> Looking backwards, if I were doing hobby work with ARM processors, I > wouldn't bat an eye about accepting that deal knowing what I now know. > The struggle in learning about and the cost in getting a good JTAG > unit along with all the rest would be short-circuited and at a price > that is probably no different than I'd otherwise spend, anyway. Paul > is offering more than just good value, in my opinion, to hobbyists.
The reason to introduce a Personal License for CrossWorks is that my inbox became littered with requests to lower the price of the tools for people who are not interested in producing products but want to play about with microprocessors. Discussions over the Personal License took a long time at CrossWorks Central; in fact, all marketing and pricing takes a long time, this is much harder than writing software. Throwing in a JTAG adapter sweetens the deal. The argument for a Personal License is clear: hobbyists want to play with micros and use a decent hassle-free piece of software when doing so. We had to balance this against the fact that some customers will purchase a personal license even though they are developing commercial code. The pragmatic way to view this is that, in all likelihood, we wouldn't sell a commercial license to these customers anyway, and if we witheld the introduction of a Personal License because of that we'd be spiting the hobbyists that genuinely want to use it. Hence, we took the Eagle model of personal use--sign a piece of paper and we'll let you purchase a Personal License. That makes people think before proceeding and, should we wish to press things later because of license misuse, we have a signed agreement. We take the same view of developers using cracked versions of our software--we're not going to sell to these "customers" anyway, so why get all bent out of shape when they do use it? Why not consider it confirmation of our product's quality that technoheads take the time to crack and use our products? Our focus remains on the class of customer we wish to establish and retain: those that delight in paying for great software and appreciate fantastic support. -- Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
On Aug 16, 3:30 am, "Paul Curtis" <p...@rowley.co.uk> wrote:
> "Jonathan Kirwan" <jkir...@easystreet.com> wrote in message > > news:lui6c35vkht9r8uoo5e50iv4qafc7j60g2@4ax.com... > > > On Wed, 15 Aug 2007 17:23:59 +0100, Chris Hills <ch...@phaedsys.org> > > wrote: > > < snip> > > > Looking backwards, if I were doing hobby work with ARM processors, I > > wouldn't bat an eye about accepting that deal knowing what I now know. > > The struggle in learning about and the cost in getting a good JTAG > > unit along with all the rest would be short-circuited and at a price > > that is probably no different than I'd otherwise spend, anyway. Paul > > is offering more than just good value, in my opinion, to hobbyists. > > The reason to introduce a Personal License for CrossWorks is that my inbox > became littered with requests to lower the price of the tools for people who > are not interested in producing products but want to play about with > microprocessors. Discussions over the Personal License took a long time at > CrossWorks Central; in fact, all marketing and pricing takes a long time, > this is much harder than writing software. Throwing in a JTAG adapter > sweetens the deal. > > The argument for a Personal License is clear: hobbyists want to play with > micros and use a decent hassle-free piece of software when doing so. We had > to balance this against the fact that some customers will purchase a > personal license even though they are developing commercial code. The > pragmatic way to view this is that, in all likelihood, we wouldn't sell a > commercial license to these customers anyway, and if we witheld the > introduction of a Personal License because of that we'd be spiting the > hobbyists that genuinely want to use it. > > Hence, we took the Eagle model of personal use--sign a piece of paper and > we'll let you purchase a Personal License. That makes people think before > proceeding and, should we wish to press things later because of license > misuse, we have a signed agreement. > > We take the same view of developers using cracked versions of our > software--we're not going to sell to these "customers" anyway, so why get > all bent out of shape when they do use it? Why not consider it confirmation > of our product's quality that technoheads take the time to crack and use our > products? Our focus remains on the class of customer we wish to establish > and retain: those that delight in paying for great software and appreciate > fantastic support. > > -- > Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk > > CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
Hello All, This is my first time posting in an usenet group, and I am using google reader, so if something goes wrong, I will appreciated your advice.. (good or bad) Any way, let's go to the meat! So seems that a good deal of these IDEs including Rowley and other is the library, because it could save you a lot of space due they are optimized to deep embedded systems, so providers state really clear this advantage. In the case specify to Rowley, Paul, could you mind to compare your library vs keil's newest library microlib for ARM? I have done some projects with the free tools and the are good but the C runtime lib is big, really big, and sometimes is not acceptable to have this extra space used for this big libs. One good example is when you jump from your vector and stack init section(asm) to __main... from here (without rewriting the C runtime libs ) I expect to do zero down, copy down, stack, heap(optional) and then main(), but what I have found is that the *argv, argc* section, and others sections are there adding a lot of space.
ARod <alejmrm@gmail.com> writes:

> Hello All, > > This is my first time posting in an usenet group, and I am using > google reader, so if something goes > wrong, I will appreciated your advice.. (good or bad) > > Any way, let's go to the meat! > > So seems that a good deal of these IDEs including Rowley and other is > the library, because it could save you > a lot of space due they are optimized to deep embedded systems, so > providers state really clear this advantage. > In the case specify to Rowley, Paul, could you mind to compare your > library vs keil's newest library microlib for ARM? > > I have done some projects with the free tools and the are good but the > C runtime lib is big, really big, and > sometimes is not acceptable to have this extra space used for this big > libs. One good example is when you jump from your > vector and stack init section(asm) to __main... from here (without > rewriting the C runtime libs ) I expect to do zero down, > copy down, stack, heap(optional) and then main(), but what I have > found is that the *argv, argc* section, and others sections > are there adding a lot of space.
gcc works fine as a "bare metal" compiler - I just did a test with a trivial main() and a minimum of startup code. The total size is ~190 bytes. If you use functions such as malloc or printf, you will bloat the code pretty quickly. This is true with all C libraries (but to a lesser extent with the commercial ones). I have found that these functions are all best avoided anyway on "small" systems (e.g. the single chip ARMS). -- John Devereux
On Sun, 12 Aug 2007 22:17:20 -0000, The_Todd <toddberk@gmail.com>
wrote:

>Bear with me as I'm a student mechanical engineer and micro >controllers are my hobby. > >I've become quite familiar with PIC microcontrollers using the CCS >compiler which I am quite fond of. I started out with their education >development kit that was a fantastic resource that came with the great >compiler, hardware, and an excellent work book with simple examples >and explanations. Its here: http://ccsinfo.com/content.php?page=education > >My question is: Is there anything similar to the CCS IDE for the ARM >controllers? A reasonably priced ($300) well supported and documented >thats easy to use IDE that supports JTAG etc...? I'm not interested in >the GNU GCC.
Both Hitex and Raisonance makes commercial IDEs that can integrate with GNU GCC. You do not need to write makefiles etc. Both have demo versions you can download. The basic difference between the demo and full versions is that the debug support is limited with the demo versions. Both sell development kits that has the IDE and gcc bundled with the kit with documentation on getting examples going on the development kits. The Hitex documentation looks better than the Raisonance documentations. Regards Anton Erasmus

Memfault Beyond the Launch