Re: gcc 3.1

faq
flatforty
contribute
subscribe
configure
search
rdf

main
parent
thread

Re: gcc 3.1
by emmanuel on Thursday 11/Jul/2002, @12:34

unfortunately, it seems g++ 3.1 isn't quite that nice, at least for kde:
http://lists.kde.org/?l=kde-core-devel&m=102166265006814&w=2

:O/

emmanuel

Related Links

·   Articles on KDE Official News
·   Also by emmanuel
·   Contact author

The Fine Print: The following comments are owned by whomever posted them.
( Reply )

Re: gcc 3.1
by Anonymous on Friday 12/Jul/2002, @01:04

I read it
That's a shame :-(
What has happened with all what was said?
I am really surprised of that results... What is happening?
And why should we use gcc 3.1 if it's WORSE BOTH in time compiling and starting apps?

I guess KDE will be faster starting apps only in a couple of years due to faster PCs. Or should we hope on something I missed from the post? ->http://lists.kde.org/?l=kde-core-devel&m=102166265006814&w=2

[ Reply To This | View ]

Re: gcc 3.1
by Ryan Cumming on Friday 12/Jul/2002, @01:36

The results are a bit dubious, as the -g3 parameter does very different things under GCC 2.95 and GCC 3.1. I suspect that relocating the DWARF symbols that 3.1 produces with -g3 accounts for the extra relocation time. Oh, and GCC 3.1 produces faster code, more friendly warnings and errors, and is more standards compliant. Some people care about more than speed.

[ Reply To This | View ]

Re: gcc 3.1
by dwyip on Friday 12/Jul/2002, @01:36

Why? Because GCC 2.95 sucked as far as C++ support was concerned. Any C++ compiler suite that

(1) didn't implement the STL correctly (look at the header files)
(2) won't let you do templated friend classes (least never worked for me)
(3) doesn't understand namespaces (2.95 faked understanding)
(4) allows file descriptors to work with iostreams (bad behavior; the C++ Standard outlaws this)
(5) a lot more that I'm missing

just, well, is bleh. GCC 3.1 fixes these.

As far as GCC 3.1 goes for speed, eh, oh well. If you turn on the optimizing features of the thing, I've found that the performance is great -- both in synthetic benchmarks and in everyday use. I've used GCC 3.1 to compile everything on this system, and although C++ compile time is slower (who really cares about that?), performance is more than acceptable.

[ Reply To This | View ]

Re: gcc 3.1
by Anonymous on Friday 12/Jul/2002, @01:50

I guess you are right, but it really shocked me the benchmark results showed above.
I didn't know about the 5 nice fixes you said. I only knew that speed was going to be better because this was the only thing explained in user forums like this one.
Good to see that at least some problems have been solved.
Thanks for your work developers!!! :-)

Could anyone give a reasonable schedule to see when (if ever) we'll see loading apps speed increase with better compilers? After all, is it really a problem, or just waiting for faster PCs is the way to go? Should developers concentrate their efforts on writting a nice desktop waiting for those faster machines in future?
Developers replies with their feelings on the topic will be greatly appreciated.

[ Reply To This | View ]

Re: gcc 3.1
by azhyd on Friday 12/Jul/2002, @04:43

> and although C++ compile time is slower (who really cares about that?)

well developpers on not so new (but no so old) cpus :(

[ Reply To This | View ]

Re: gcc 3.1
by dwyip on Friday 12/Jul/2002, @10:00

> well developpers on not so new (but no so old) cpus :(

Eh, I don't really mind. Gives me an excuse to say "I'm working" when all I'm doing is waiting for a compile to finish :D

[ Reply To This | View ]

Re: gcc 3.1
by Perfomance must improve. on Sunday 14/Jul/2002, @11:14

Not everybody in this world has the money to buy a new PC. Look at most of thecountries in the third-world. Linux (and ofcourse KDE and other open-source)has very much chances, because thos people don't have much money to spend and Linux is free. But if people have to spend a lot of money to run programs with acceteble performance they will proberly stick to a system like Windows 95, 98 or ME. Simply because these sytems run very well on low-resource machines.

[ Reply To This | View ]

Re: gcc 3.1
by aleXXX on Monday 15/Jul/2002, @04:16

This is certainly true.
But it ain't simple to make it run faster.
Another option would be to have one quite fast server and a bunch of slow (e.g. old pentiums) boxes acting as X terminals.

Bye
Alex

[ Reply To This | View ]

Re: gcc 3.1
by dan on Monday 15/Jul/2002, @19:16

> But it ain't simple to make it run faster.

well.. i hope gcc 3.1 will be further improved (a lot!) during the next months.

another idea: what about using the intel compiler suite instead of gcc-3.1? anybody tried this one together with kde 3.0/3.1? i've read the intel compiler made a lot of progress regarding "gcc compatibility" recently.. and it's supposed to be faster and create about 10 to 30 percent faster binaries than gcc 2.95.3.

[ Reply To This | View ]

The Fine Print: The previous comments are owned by whomever posted them.
( Reply )


		"Feature freeze: You're allowed to add new bugs, but no new features." -- Richard J. Moore
KDE®, "K Desktop Environment", "KDE Dot News", "got the dot?" and the KDE Logo® are trademarks or registered trademarks of KDE e.V. in the European Union, the United States and other countries. All other trademarks and copyrights on this page are owned by their respective owners. Comments are owned by the poster. The rest: Copyright © 2000-2008 KDE e.V. for The KDE Project. For further information or comments on this site, please contact the Webmaster.