This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Long bootstrap/check times on sparc
- From: "David S. Miller" <davem at redhat dot com>
- To: lucier at math dot purdue dot edu
- Cc: gcc at gcc dot gnu dot org
- Date: Wed, 08 May 2002 18:17:39 -0700 (PDT)
- Subject: Re: Long bootstrap/check times on sparc
- References: <200205081646.g48GkAg24582@banach.math.purdue.edu>
From: Brad Lucier <lucier@math.purdue.edu>
Date: Wed, 8 May 2002 11:46:10 -0500 (EST)
Solaris is not so bad when you can build the libraries with make -j 4, say,
and overlap some of their outrageous file access and system times.
Perhaps someone can figure out how to fix this problem with gnu make.
In my experience 'make -k -jN check' is the real dog. I can fully
bootstrap twice or so on a dual-cpu Ultra running Linux in the time a
monster 32 cpu Solaris box takes to make check.
It's all process creation + file system overhead.
Doing the build on tmpfs (rofl) helps the build out but doesn't help
the make check times much at all.
I am happy someone brought up the libjava parallel building issue, I
had noticed that as well and it would help substantially to fix that
on every platform :-)