[avr-gcc-list] Re: avr superoptimizer
John Regehr wrote: That's why I'm (desperately) hoping that the LTO project in GCC will eventually help in this area. http://gcc.gnu.org/wiki/LinkTimeOptimization Alternatively, an LLVM backend for AVR would give this capability (approximately) for free. approximately free being just the time taken to write an AVR backend of LLVM, as well as to figure out how to handle non-standard issues such as interrupt handling and flash data access, and then to port avr-libc. I fully agree that an LLVM backend for the AVR has the potential to be a wonderful thing, and maybe that if one were starting from scratch today, it would be easier to write an AVR backend for LLVM than for gcc, but I think unfortunately it would be quite far from approximately free. Perhaps we could petition Atmel to sponsor a couple of students to make a starting point over as a summer project? ___ AVR-GCC-list mailing list AVR-GCC-list@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-gcc-list
RE: [avr-gcc-list] Re: avr superoptimizer
-Original Message- From: avr-gcc-list-bounces+eric.weddington=atmel@nongnu.org [mailto:avr-gcc-list-bounces+eric.weddington=atmel@nongnu. org] On Behalf Of David Brown Sent: Monday, April 20, 2009 2:17 PM To: avr-gcc-list@nongnu.org Subject: [avr-gcc-list] Re: avr superoptimizer Perhaps we could petition Atmel to sponsor a couple of students to make a starting point over as a summer project? I don't think we're quite at a point to do that yet. But I'm not ruling it out completely. ___ AVR-GCC-list mailing list AVR-GCC-list@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-gcc-list
RE: [avr-gcc-list] Re: avr superoptimizer
-Original Message- From: avr-gcc-list-bounces+eric.weddington=atmel@nongnu.org [mailto:avr-gcc-list-bounces+eric.weddington=atmel@nongnu. org] On Behalf Of David Brown Sent: Monday, April 20, 2009 2:36 PM To: avr-gcc-list@nongnu.org Subject: [avr-gcc-list] Re: avr superoptimizer I really don't know the status of LLVM and how well it would work for a small device like the AVR - but it would be *very* interesting to find out! There are a lot of small microcontrollers with limited or no open source compilers, most of which are poorly suited to gcc ports. Maybe they are equally unsuitable for LLVM ports - but maybe not. The only way to know for sure is for someone to work on such a port and see how it goes. That sounds like a great student project to me! Oh, you're certainly preaching to the choir. I understand the potential benefits of such a thing. It's the sponsor part that may not fly right now. ___ AVR-GCC-list mailing list AVR-GCC-list@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-gcc-list