GCC quality (was Re: AMD-64 optimizations, some (new) code)

Ashod Nakashian saghmos at xter.net
Wed Sep 28 04:38:10 CEST 2005


Torbjorn Granlund wrote:
> Ashod Nakashian <saghmos at xter.net> writes:
> 
> I think the entire software community has a very serious attitude
> problem leading to quality problems with almost all software.  GCC,
> the Linux kernel, Windows, etc, are examples.
> 

Is that always an attitude problem. I thought it was more of a 
complexity management problem. You can't be suggesting that most 
everyone has the same attitude issues in the community; try saying that 
to the devoted long-term GNU or Microsoft developers... they won't be 
calm in their responses. Some are dedicated and take pride in their 
work, and they don't do it for the wrong reasons. It wouldn't be fair, 
for example, to suggest that bugs in GMP are because of this community's 
attitude; you would probably be insulted by the such a hint.

>   By the way, the same problems you speak of are very serious issues in 
>   Intel's C++ compilers as well. [...]
> 
> I believe Intel's compiler is written with SPEC in mind; it might
> be unsuitable for any other programs.  :-)

Right. And the irony of things: a commercial product to promote their 
processors!

>   I thought it would be interesting to put a $400+ professional and 
>   commercial compiler in prespective with an open-source, FREE compiler.
> 
> Yes, and HP's compilers (targeting HPPA and IA64) are also very
> buggy.  But let's not argue that a certian piece of Free Software
> is good enough, just because there are commercial horror examples!
> Is GNU/Linux's buggyness OK just because Windows is worse?  (And is
> it really still worse?)
> 

Not quite. Again, it's a complexity management issue at the heart of 
things. One can't ignore the fact that a compiler per se IS one of the 
more complex software projects out there. Forget the fact that 
languages, back-end hardware differences, portability, backwards 
compatibility and optimizations add more than just spices to the mix.

So while HP's compiler issues are as much unacceptable as GNU's, and 
Windows bugs are not an excuse for GNU/Linux's (and not a single bug in 
Linux was more problematic than downloading a patch, don't get me 
started on Windows, where I downgraded due to serious bugs) I still 
think it's not fair not to address the real problem; complexity. You 
really make it sound like GCC is the most unreliable compiler you can 
run into. That's not true. Plus, take any other compiler any day and try 
a project such as GMP on them and you'd sooner rather than later 
discover that GMP is probably the most complex/hard test-case that 
compiler ever had... and it will probably fail it too.

-Ash


More information about the gmp-discuss mailing list