Re: horrible disk thorughput on itanium

[Andrew Morton <akpm at zip dot com dot au>]

Ulrich Drepper wrote:
> 
> Andreas Jaeger <aj@suse.de> writes:
> 
> > This should be doable.  We could easily implement libio/putc.c as
> > follows:
> 
> That's stupid.  Either you use putc_unlocked() as any sane person
> should in non-threaded code which then expands inline or you use
> __fsetlocking(FSETLOCKING_BYCALLER).  Just because people cannot
> program correctly doesn't require the implementation to fix their
> problems.

I can take a perfectly good program written in 1995, run it
under current glibc and see it run like a crippled duck
purely because of changes in the C library.  Are you saying that
all existing code needs to be edited, recompiled, retested and
redistributed, just so it can perform as well as it used to?

I don't, for example, see any of these new-fangled hacks in the
bison-1.30 source.   




In answer to Geoff Keating's question, this program:

#include <stdio.h>

main()
{
        int i;

        for (i = 0; i < 100*1000*1000; i++)
                putc(0, stdout);
        exit(0);
}

when directed to /dev/null takes 7.5 seconds.  But only 1.3 seconds
with putc_unlocked.