This is related to my "D'oh post" regarding the Cortex M4 floating point size. According to ARM, the Cortex R4 supports double-precision floating point, and it looks like it's maybe more amenable to embedding than the Cortex Ax cores. I really want to find a chip that I can just lick the back of and stick onto a board, rather than something that'll become the CPU in a built-up computer with external memory and all that razz-matazz. TI has a Cortex R4 part, but I've used the predecessor part and the only tools available for it were from Keil -- I dislike single-sourcing, and I'd much prefer using Gnu tools if there are such for the R4. So, has anyone used the R4? Do you happen to know if the Gnu tools support it well? Can you recommend any off the shelf embedded processors that use it? (TI included -- just because the SM470R1B1M was hard to work with doesn't mean that TI hasn't solved those problems with their Cortex version). Thanks much. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Cortex R4 Tools and Processors
Started by ●August 4, 2013
Reply by ●August 4, 20132013-08-04
On Sun, 04 Aug 2013 11:53:58 -0500, Tim Wescott <tim@seemywebsite.really> wrote:>This is related to my "D'oh post" regarding the Cortex M4 floating point >size. > >According to ARM, the Cortex R4 supports double-precision floating point, >and it looks like it's maybe more amenable to embedding than the Cortex >Ax cores. I really want to find a chip that I can just lick the back of >and stick onto a board, rather than something that'll become the CPU in a >built-up computer with external memory and all that razz-matazz. > >TI has a Cortex R4 part, but I've used the predecessor part and the only >tools available for it were from Keil -- I dislike single-sourcing, and >I'd much prefer using Gnu tools if there are such for the R4. > >So, has anyone used the R4? Do you happen to know if the Gnu tools >support it well? Can you recommend any off the shelf embedded processors >that use it? (TI included -- just because the SM470R1B1M was hard to >work with doesn't mean that TI hasn't solved those problems with their >Cortex version).Both Rowley http://www.rowley.co.uk/arm/Support_Packages.htm and Imagecraft https://www.imagecraft.com/ seem to support TI TMS570 series processors. I've used both compilers but not that series (or other R4) processors so I'm not absolutely sure that's what you're looking for. Rowley is GNU-based, Imagecraft is independent. I've used both for commercial embedded products. There's probably a pure GNU solution out there, but I'm in a situation where (L)GPL issues arise.
Reply by ●August 5, 20132013-08-05
On 04/08/2013 17:53, Tim Wescott wrote:> TI has a Cortex R4 part, but I've used the predecessor part and the only > tools available for it were from Keil -- I dislike single-sourcing, and > I'd much prefer using Gnu tools if there are such for the R4.TI also supply their own compiler and Eclipse based IDE in their Code Composer Studio (CCS) download. Regards, Richard. + http://www.FreeRTOS.org Designed for microcontrollers. More than 103000 downloads in 2012. + http://www.FreeRTOS.org/plus Trace, safety certification, FAT FS, TCP/IP, training, and more...
Reply by ●August 5, 20132013-08-05
Tim Wescott <tim@seemywebsite.really> wrote:> Do you happen to know if the Gnu tools support it well?The documentation at least claims it does. OpenOCD should also support it as well. -a
Reply by ●August 6, 20132013-08-06
Tim Wescott wrote:> This is related to my "D'oh post" regarding the Cortex M4 floating point > size. > > According to ARM, the Cortex R4 supports double-precision floating point, > and it looks like it's maybe more amenable to embedding than the Cortex > Ax cores. I really want to find a chip that I can just lick the back of > and stick onto a board, rather than something that'll become the CPU in a > built-up computer with external memory and all that razz-matazz. > > TI has a Cortex R4 part, but I've used the predecessor part and the only > tools available for it were from Keil -- I dislike single-sourcing, and > I'd much prefer using Gnu tools if there are such for the R4. > > So, has anyone used the R4? Do you happen to know if the Gnu tools > support it well? Can you recommend any off the shelf embedded processors > that use it? (TI included -- just because the SM470R1B1M was hard to > work with doesn't mean that TI hasn't solved those problems with their > Cortex version). > > Thanks much. >We just worked on a TMS570 (R4) port for uCOS-II using IAR EWARM 6.5. Worked great. Scott
Reply by ●August 7, 20132013-08-07
On 04/08/13 19:51, Rich Webb wrote:> On Sun, 04 Aug 2013 11:53:58 -0500, Tim Wescott > <tim@seemywebsite.really> wrote: > >> This is related to my "D'oh post" regarding the Cortex M4 floating point >> size. >> >> According to ARM, the Cortex R4 supports double-precision floating point, >> and it looks like it's maybe more amenable to embedding than the Cortex >> Ax cores. I really want to find a chip that I can just lick the back of >> and stick onto a board, rather than something that'll become the CPU in a >> built-up computer with external memory and all that razz-matazz. >> >> TI has a Cortex R4 part, but I've used the predecessor part and the only >> tools available for it were from Keil -- I dislike single-sourcing, and >> I'd much prefer using Gnu tools if there are such for the R4. >> >> So, has anyone used the R4? Do you happen to know if the Gnu tools >> support it well? Can you recommend any off the shelf embedded processors >> that use it? (TI included -- just because the SM470R1B1M was hard to >> work with doesn't mean that TI hasn't solved those problems with their >> Cortex version). > > Both Rowley http://www.rowley.co.uk/arm/Support_Packages.htm and > Imagecraft https://www.imagecraft.com/ seem to support TI TMS570 > series processors. I've used both compilers but not that series (or > other R4) processors so I'm not absolutely sure that's what you're > looking for. > > Rowley is GNU-based, Imagecraft is independent. I've used both for > commercial embedded products. There's probably a pure GNU solution out > there, but I'm in a situation where (L)GPL issues arise. >The GPL nature of gcc does not affect you unless you are trying to run the compiler on the microcontroller. Any sane toolchain will have libraries that can be linked with your code without any license restrictions - but if you use additional third-party libraries or code, you obviously need to check the licence (whether commercial or open source).
Reply by ●August 7, 20132013-08-07
On 07/08/2013 15:34, David Brown wrote:> On 04/08/13 19:51, Rich Webb wrote: >> On Sun, 04 Aug 2013 11:53:58 -0500, Tim Wescott >> <tim@seemywebsite.really> wrote: >> >>> This is related to my "D'oh post" regarding the Cortex M4 floating point >>> size. >>> >>> According to ARM, the Cortex R4 supports double-precision floating point, >>> and it looks like it's maybe more amenable to embedding than the Cortex >>> Ax cores. I really want to find a chip that I can just lick the back of >>> and stick onto a board, rather than something that'll become the CPU in a >>> built-up computer with external memory and all that razz-matazz. >>> >>> TI has a Cortex R4 part, but I've used the predecessor part and the only >>> tools available for it were from Keil -- I dislike single-sourcing, and >>> I'd much prefer using Gnu tools if there are such for the R4. >>> >>> So, has anyone used the R4? Do you happen to know if the Gnu tools >>> support it well? Can you recommend any off the shelf embedded processors >>> that use it? (TI included -- just because the SM470R1B1M was hard to >>> work with doesn't mean that TI hasn't solved those problems with their >>> Cortex version). >> >> Both Rowley http://www.rowley.co.uk/arm/Support_Packages.htm and >> Imagecraft https://www.imagecraft.com/ seem to support TI TMS570 >> series processors. I've used both compilers but not that series (or >> other R4) processors so I'm not absolutely sure that's what you're >> looking for. >> >> Rowley is GNU-based, Imagecraft is independent. I've used both for >> commercial embedded products. There's probably a pure GNU solution out >> there, but I'm in a situation where (L)GPL issues arise. >> > > The GPL nature of gcc does not affect you unless you are trying to run > the compiler on the microcontroller. Any sane toolchain will have > libraries that can be linked with your code without any license > restrictions - but if you use additional third-party libraries or code, > you obviously need to check the licence (whether commercial or open source). >Besides which, as far as I know Rowley provide their own (not open source) optimised libraries when you purchase their tools. Only the compiler is GPL. Regards, Richard. + http://www.FreeRTOS.org Designed for microcontrollers. More than 103000 downloads in 2012. + http://www.FreeRTOS.org/plus Trace, safety certification, FAT FS, TCP/IP, training, and more...
Reply by ●August 7, 20132013-08-07
On Wed, 07 Aug 2013 16:34:28 +0200, David Brown <david@westcontrol.removethisbit.com> wrote:>On 04/08/13 19:51, Rich Webb wrote: >> On Sun, 04 Aug 2013 11:53:58 -0500, Tim Wescott >> <tim@seemywebsite.really> wrote: >> >>> This is related to my "D'oh post" regarding the Cortex M4 floating point >>> size. >>> >>> According to ARM, the Cortex R4 supports double-precision floating point, >>> and it looks like it's maybe more amenable to embedding than the Cortex >>> Ax cores. I really want to find a chip that I can just lick the back of >>> and stick onto a board, rather than something that'll become the CPU in a >>> built-up computer with external memory and all that razz-matazz. >>> >>> TI has a Cortex R4 part, but I've used the predecessor part and the only >>> tools available for it were from Keil -- I dislike single-sourcing, and >>> I'd much prefer using Gnu tools if there are such for the R4. >>> >>> So, has anyone used the R4? Do you happen to know if the Gnu tools >>> support it well? Can you recommend any off the shelf embedded processors >>> that use it? (TI included -- just because the SM470R1B1M was hard to >>> work with doesn't mean that TI hasn't solved those problems with their >>> Cortex version). >> >> Both Rowley http://www.rowley.co.uk/arm/Support_Packages.htm and >> Imagecraft https://www.imagecraft.com/ seem to support TI TMS570 >> series processors. I've used both compilers but not that series (or >> other R4) processors so I'm not absolutely sure that's what you're >> looking for. >> >> Rowley is GNU-based, Imagecraft is independent. I've used both for >> commercial embedded products. There's probably a pure GNU solution out >> there, but I'm in a situation where (L)GPL issues arise. >> > >The GPL nature of gcc does not affect you unless you are trying to run >the compiler on the microcontroller. Any sane toolchain will have >libraries that can be linked with your code without any license >restrictions - but if you use additional third-party libraries or code, >you obviously need to check the licence (whether commercial or open source).Correct. Rowley, for example, does use a gcc compiler and, of course, the generated code itself isn't GPL-burdened. They also (being a sane toolchain ;-) provide their own non-LGPL libraries. In the "check the license" vein, however, using C++ RTTI or exceptions does link in GPL-licensed code (not just LGPL). I've been puzzled about why that's the case but since I stay away from C++ in embedded work I haven't worried about (or investigated) that too much.
Reply by ●August 7, 20132013-08-07
Rich Webb <webb.ra@example.net> wrote:> In the "check the license" vein, however, using C++ RTTI or exceptions > does link in GPL-licensed code (not just LGPL). I've been puzzled > about why that's the case but since I stay away from C++ in embedded > work I haven't worried about (or investigated) that too much.The GCC C++ library license contains an exception allowing its use in proprietary software: <http://gcc.gnu.org/onlinedocs/gcc-4.8.1/libstdc++/manual/manual/license.html> -a
Reply by ●August 7, 20132013-08-07
On Wed, 7 Aug 2013 18:29:56 +0000 (UTC), Anders.Montonen@kapsi.spam.stop.fi.invalid wrote:>Rich Webb <webb.ra@example.net> wrote: > >> In the "check the license" vein, however, using C++ RTTI or exceptions >> does link in GPL-licensed code (not just LGPL). I've been puzzled >> about why that's the case but since I stay away from C++ in embedded >> work I haven't worried about (or investigated) that too much. > >The GCC C++ library license contains an exception allowing its use in >proprietary software: ><http://gcc.gnu.org/onlinedocs/gcc-4.8.1/libstdc++/manual/manual/license.html>Interesting. Thanks!