Linked by Thom Holwerda on Sat 15th Sep 2007 20:14 UTC, submitted by deanna
BSD and Darwin derivatives Anders Magnusson's BSD-licensed pcc compiler has been imported into NetBSD's pkgsrc and OpenBSD's src tree. Anders wrote to NetBSD's tech-toolchain list: "It is not yet bug-free, but it can compile the i386 userspace. The big benefit of it is that it is fast, 5-10 times faster than gcc, while still producing reasonable code. The only optimization added so far is a multiple-register-class graph-coloring register allocator, which may be one of the best register allocators today. Conversion to SSA format is also implemented, but not yet the phi function. Not too difficult though, after that strength reduction is high on the list."
Thread beginning with comment 271695
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[3]: Oh, great!
by Marcellus on Sun 16th Sep 2007 10:06 UTC in reply to "RE[2]: Oh, great!"
Marcellus
Member since:
2005-08-26

IMHO, it's a better, cleaner design than GCC, but then again, it has the luxury of being built from the ground-up, and doesn't have decades of legacy to support.


I'm not sure what you mean here. If you mean that GCC have to support old source, isn't it the case that old sources that compiled with 2.95 can't be compiled with a current version? Not the only case of GCC no longer supporting legacy sources.

If you mean GCC supporting features, there's plenty of stuff that has been removed from support over the years as well.


Aside from that, GCC is slow period. Even if you disable optimizations, it's painfully slow.

If trying to use the same main sources for all possible targets makes it slower for all, they really should split it up into target classes instead.

Reply Parent Bookmark Score: 1

RE[4]: Oh, great!
by luzr on Sun 16th Sep 2007 10:53 in reply to "RE[3]: Oh, great!"
luzr Member since:
2005-11-20

> Aside from that, GCC is slow period. Even if you disable optimizations, it's painfully slow.

But it gets better.

IME GCC4.2 is almost 50% faster than GCC3.4

Reply Parent Bookmark Score: 2

RE[5]: Oh, great!
by dylansmrjones on Sun 16th Sep 2007 13:09 in reply to "RE[4]: Oh, great!"
dylansmrjones Member since:
2005-10-02

Slow compilation is the price you pay for fast binaries.

GCC usually beats the crap out of Visual C++ in terms of producing the fastest binaries, as your earlier link to the Ultimate++ benchmark showed.

However, that doesn't mean PCC isn't interesting. It is very much so.

Just don't forget Marcellus is FUD'ing* and trolling as usual.

* http://www4.osnews.com/permalink?271690

Reply Parent Bookmark Score: 3

RE[4]: Oh, great!
by rayiner on Sun 16th Sep 2007 18:25 in reply to "RE[3]: Oh, great!"
rayiner Member since:
2005-07-06

My point was that the major reason LLVM is faster than GCC is because it was built from the ground-up using the latest compiler techniques. GCC, in comparison, has incrementally evolved from a 20 year old codebase. The GCC folks don't have the luxury of going off somewhere and spending five years totally rebuilding things. They have way too many existing customers and existing platforms to support. LLVM doesn't have those kinds of constraints.

If trying to use the same main sources for all possible targets makes it slower for all, they really should split it up into target classes instead.

This wouldn't work in any way that didn't require huge code duplication.

Edited 2007-09-16 18:31

Reply Parent Bookmark Score: 3