This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: gcc compile-time performance

Marc Espie wrote:
> People work with old architectures for the challenge, and because they
> have fun with it. And some other people are very happy to be able to DO
> something, like run Unix, on their 1980 amiga/atari/vax/m88k/whatever.

> The big problem is that, once this stops being fun, people stop playing
> with old hardware.

Use of "old" hardware goes far beyond hobbyists. Many important, large
groups who use "antiquated" hardware. In my experience, this includes third
world countries, school districts, small businesses, colleges, and anyone on
a budget -- the precise markets where "free" software feels it can make
inroads on proprietary systems.

High-reliability software often runs on "old" hardware; the Hubble Space
Telescope runs a recently-installed 80486-based computer, while the space
shuttle uses 8086s!

A common complaint about the "Wintel" platform is how slow, bloated software
"upgrades" require ever-faster hardware -- a force hardware upgrade cycle.
GNU and Linux seem to be falling into the false trap that "bigger is better"
and that people can just "upgrade" if performance gets too slow.

The reality is that people do serious, professional work for a large segment
of the computing populace on "old" systems. And gcc should recognize that,
both in terms of generated code performance and compilation time. I think
they have the former nailed; now it's time to work on the latter.

> I think that the main problem comes from the simple fact that it is rather
> unsexy to work on speed increases in gcc, compared to adding new-fangled
> optimizations that gain 0.05% on a SPEC test, and which are not
> even finished overall.

Well, articles like mine () probably engender this attitude, and I, for one,
care deeply about "new-fangled" optimizations. However, optimizations don't
do me a damned bit of good if I can't get the product finished in time for
my customers. I don't see this as an either-or proposition; why can't we
have fast compile times for development and powerful optimizations for
release builds?


Scott Robert Ladd
Coyote Gulch Productions,
No ads -- just very free (and somewhat unusual) code.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]