This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc compile-time performance
- From: "David S. Miller" <davem at redhat dot com>
- To: dewar at gnat dot com
- Cc: drow at mvista dot com, gcc at gcc dot gnu dot org, gdr at codesourcery dot com, law at redhat dot com, mark at codesourcery dot com, scott at coyotegulch dot com
- Date: Sat, 18 May 2002 22:20:26 -0700 (PDT)
- Subject: Re: gcc compile-time performance
- References: <20020519051638.D7B18F28CC@nile.gnat.com>
From: dewar@gnat.com (Robert Dewar)
Date: Sun, 19 May 2002 01:16:38 -0400 (EDT)
Surely you must be exaggerating here, several orders of magnitude would
imply at least a factor of 100???
I can bootstrap GCC with all languages enabled (except Ada) and fully
"make check" in about 1 hour or so on a 2 processor Ultrasparc with
about 1 gigabyte of ram running Linux.
Running Solaris, the same bootstrap (and it is an equivalently
targetted bootstrap doing the same amount of multilibbing) takes more
than a day on a 32 processor Solaris machine with several gigabytes of
ram. Most of this time is spent doing the libjava build (which isn't
parallelized) and in "make check" (which shows massive deficiencies in
Solaris's process creation overhead).
So to be more precise, I meant about an order of "20 or so" when I
said several orders of magnitude.