This is the mail archive of the gcc@gcc.gnu.org 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


 In message <20020519.045228.03127844.davem@redhat.com>, "David S. Miller" 
writes:
 >    From: dewar@gnat.com (Robert Dewar)
 >    Date: Sun, 19 May 2002 07:49:24 -0400 (EDT)
 > 
 >    What we are comparing here (Solaris vs Redhat Linux) is doing a full boot
 > strap
 >    of C and Ada, and this seems to take reasonably comparable times on the
 >    two machines (the Sun is slower by a factor of 2-3, but it is also a slow
 > er
 >    machine by this factor).
 >    
 > What about "make check", do you see the same matching ratio?
 > 
 > I really doubt you will.... the Linux box should nuke the Solaris
 > one in terms of how long it takes to run "make check" even just
 > for the C language.
 > 
 > Linux also nukes Solaris in terms of how long it takes to build
 > libjava, but that is mostly a bug in the java compiler which Mark is
 > fixing (during compilation the java compiler calls stat() on the same
 > file a couple thounsand times, something like this).
 > 
 > But, again, "make check" is the real killer under Solaris.
Also true for HP-UX ("make check" being a killer time-wise).

jeff


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