KEil bought by ARM

You probably haven't used professional compilers then... Yes, in the

8 and 16-bit space some commercial compilers are worse than open source ones. However 10 year old ARM compilers _still_ generate better code than the latest GCC...

GCC is definitely getting better, but is it narrowing the gap? If you consider large C++ applications then the scope for improvement is enormous - I guess a factor 2 of improvement is achievable.

Applications grow about as fast as memory and performance do, so any improvements are immediately swallowed.

Interestingly in the 32-bit RISC space GCC is furthest behind. If you look at EEMBC or Spec for example you don't see (m)any uses of GCC.

Wilco

Reply to
Wilco Dijkstra
Loading thread data ...

I have not used a "professional" (commercial) ARM compiler, this is true.

Maybe - my programs tend to be "medium sized" C. I would suggest that we can expect further improvements in gcc with respect to "large C++" programs. This is because efforts by the gcc maintainers in this area would benefit all targets enormously, compared to working on micro-optimising a single backend. This is especially likely if the factor if 2 you mention is correct.

Well I guess that depends on who does the benchmarking!

I did come accross this

In summary, it looks like Keil ran their own compiler with full optimisation, and gcc with none! When both were run with sensible settings for embedded work, gcc came out ahead in both speed and memory footprint.

--

John Devereux
Reply to
John Devereux

It's not closing. Gcc is still about a decade behind. I get that figure independently from several compiler vendors and tool makers. In places Gcc is apparently an appalling mess.

It is certainly not a sensible choice for quite a few architectures.

Not for embedded systems or the 8 bit market would have disappeared LONG ago.

What is 50 cents sent difference on the MCU? well multiply it by 50K per hear and you get some idea! then there is the additional cost of the memory, the more complex PCB... the cost is on a size* pads* holes type equation. it all ads up and 1 dollar per board is 50,000 dollars per year.

The other problem is EMC... many want to run the MCU SLOWER not faster.

YEs for some but there is a LONG way to go (if at all) before it will over take the 8 bit market.

Yes Gcc can not hope to compete in the 8 and most 16 bit markets.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
/\/\/ chris@phaedsys.org      www.phaedsys.org \/\/\
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Reply to
Chris Hills

And of course these same compiler vendors and tool makers are completely objective and have no axe to grind? Yeah, pull the other one.

Ian

Reply to
Ian Bell

Hmm, sounds like these all have a vested interest in saying that...

Agreed, the PIC springs to mind. In fact all the older 8 bit parts (those with very few registers).

Yes, for sufficient production quantities, even the smallest margin becomes significant. Of course you can make the same arguments for C vs assembler! We sell The other problem is EMC... many want to run the MCU SLOWER not faster.

Actually avr-gcc is pretty good.

--

John Devereux
Reply to
John Devereux

None of the conversations were commercial or public. In the case of some of the tool vendors they had no axe to grind as their tools work with GCC where required.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
/\/\/ chris@phaedsys.org      www.phaedsys.org \/\/\
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Reply to
Chris Hills

[...]

I hope you realize you have destroyed your credibility by claiming GCC is about a decade behind. Or were you talking about places that GCC never was and never will be? Might as well be infinity behind.

And your relevant point would be?

Reply to
Bryan Hackney

Rather than killing off Keil 8051 (and other) compilers, ARM could well gradually raise the prices of them. That way they squeeze more profit from the existing 8051 market, while encouraging new designs to move to other chips (i.e., ARM).

Reply to
David Brown

The oldest old compiler we sent as support to a customer this year was

16 years old.

Walter Banks Byte Craft Limited

larwe wrote:

Reply to
Walter Banks

Why would you choose to archive source for an open source compiler instead of executables?

The difference with the commercial compiler is archiving the executables is your only option. Most significantly you can't distribute copies of the archived commercial complier so any other party who may need to maintain the software in the future will have to buy their own copy of the commercial compiler to archive now or hope with crossed fingers the compiler vendor will still be in business and able to supply (at a cost) obsolete versions of their tools.

Reply to
nospam

I think a followup to this is warranted. After I posted this a representative from Altium contacted me for details. Apparently the back version I asked for was available but the sales person I had contact with wasn't aware of it.

Too late to be of use to me anymore but anyone getting the same response from their sales representative might try pushing a little harder.

I did appreciate getting an apology though.

Robert

Reply to
R Adsett

When I worked for Racal Comms a few years ago, the software engineers I worked with on a military system had so many problems with the very expensive compiler they were supposed to use that they switched to gcc. They didn't tell management who were horrified when they found out; they let them continue using gcc, however.

Leon

Reply to
Leon

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.