This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: cpplib: Separate out translation phases 1 and 2
- From: Devang Patel <dpatel at apple dot com>
- To: Neil Booth <neil at daikokuya dot co dot uk>
- Cc: Devang Patel <dpatel at apple dot com>, gcc-patches at gcc dot gnu dot org
- Date: Fri, 18 Apr 2003 17:41:58 -0700
- Subject: Re: cpplib: Separate out translation phases 1 and 2
On Friday, April 18, 2003, at 05:18 PM, Neil Booth wrote:
Neil Booth wrote:-
Actually, WTF, I think I'll do this, though like I said it's not
really
a fair comparison. I'll report the results tomorrow to "make
bootstrap"
GCC before and after the patch with --enable-languages=c and
--disable-checking; the latter setting will also magnify any slowdown.
After:
real 31m32.023s
user 28m36.130s
sys 1m52.130s
Before:
real 30m55.332s
user 28m18.960s
sys 1m48.990s
Which is 2% slowdown. Since the new code doesn't take advantage of
newly available optimizations, and also has baggage from the old way of
doing things, it can only get better and I don't consider this a
blocker.
In large real life projects, we have seen that CPP is responsible for
larger portion of compiler time compared to its chunk while building
compiler. But that's OK. Now we have some idea about what to expect.
Plus now we know how much speed up we have to achieve with future
optimizations to reach original level.
And when somebody scream on us saying GCC 3.4 is slow than we know where
to start hunt if we keep such numbers for all important reorg/new
features.
It is up to you and/or others to decide if this is a blocker or not.
I was just interested numbers.
Thank you very much,
-Devang