Just some numbers that I got with a small testsuite. This
is the setiathome client running on an AMD Athlon-500 (K7),
FreeBSD 4.0-current (as of 1999-09-18). Of course, I've
used the same work unit for all tests. I also changed the
code to stop after a certain amount of data has processed,
beca
On Fri, Sep 24, 1999 at 08:43:37AM -0700, Darryl Okahata wrote:
> "Zach N. Heilig" <[EMAIL PROTECTED]> wrote:
>
> > The application for the tests is mpg123.
> > test mp3 playing time: 373 seconds.
> [ ... ]
> > 1) No Optimization
> >225.08 real 224.30 user 0.23 sys
>
Darryl Okahata wrote in list.freebsd-current:
> "Zach N. Heilig" <[EMAIL PROTECTED]> wrote:
> > The application for the tests is mpg123.
> > test mp3 playing time: 373 seconds.
> [ ... ]
> > 1) No Optimization
> >225.08 real 224.30 user 0.23 sys
> [ ... ]
< said:
> The point is that it _does_ hurt. Anything above -O3 is _likely_ to
> have bugs.
And more to the point: the FreeBSD Project will not support those who
compile their kernel or world with anything other than the default
optimization settings and compilers.
-GAWollman
--
Garrett A. Woll
"Zach N. Heilig" <[EMAIL PROTECTED]> wrote:
> The application for the tests is mpg123.
> test mp3 playing time: 373 seconds.
[ ... ]
> 1) No Optimization
>225.08 real 224.30 user 0.23 sys
[ ... ]
> 2) -O3 -mcpu=i486 -march=i486 -fomit-frame-pointer -fschedule-i
On Thu, Sep 23, 1999 at 10:57:07AM -0400, Luke wrote:
> I have -03 -mpentium -fomit-frame-pointer and the only inexpicable
> problem I have is I cant compile groff with more than -O for some reason.
> Other than that make world and make release have worked many times.
> I wonder does any
On Fri, Sep 24, 1999 at 12:08:10PM +0900, Daniel C. Sobral wrote:
> The point is that it _does_ hurt. Anything above -O3 is _likely_ to
> have bugs.
I am forced to agree about the fact that optimisation is traditionally
buggy :)
I tried to optimize my system with -O3 and -pipe. So I build myself
"Daniel C. Sobral" <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] wrote:
> >
> > But specifying something too high (-O99) doesn't hurt - I'm using -O6 for
> > gcc 2.95.1 (which, by the way, compiles almost everything in 3.3-RELEASE
> > and 4.0-CURRENT, the only thing still troubling me with it
Oliver Fromme <[EMAIL PROTECTED]> writes:
> The gcc optimizer is traditionally buggy. I wouldn't trust a
> system compiled with anything more than -O (especially on
> production servers). The higher optimization levels don't
> provide much of a speed improvement anyway, sometimes they make
> t
[EMAIL PROTECTED] wrote:
>
> But specifying something too high (-O99) doesn't hurt - I'm using -O6 for
> gcc 2.95.1 (which, by the way, compiles almost everything in 3.3-RELEASE
> and 4.0-CURRENT, the only thing still troubling me with it is the kernel).
The point is that it _does_ hurt. Anythin
[EMAIL PROTECTED] wrote in list.freebsd-current:
> On Thu, 23 Sep 1999, The Hermit Hacker wrote:
> > just curious, but what is the max setting that can be used to compile the
> > kernel?
>
> Works rather nicely here with
>
> -O6 -mpentiumpro -march=pentiumpro -pipe -s -fexpensive-optimiza
In the last episode (Sep 23), [EMAIL PROTECTED] said:
> On Thu, 23 Sep 1999, Matthew Dillon wrote:
> > I tend not to like the higher optimization levels because they
> > cause the compiler to attempt to turn static functions into inlines
> > and, in my opinion, it doesn't do a very good job of sel
On Thu, 23 Sep 1999, Matthew Dillon wrote:
>I tend not to like the higher optimization levels because they cause
>the compiler to attempt to turn static functions into inlines and,
>in my opinion, it doesn't do a very good job of selecting which functions
>to convert. The result
On Thu, 23 Sep 1999, Sheldon Hearn wrote:
> On Thu, 23 Sep 1999 08:23:09 -0300, The Hermit Hacker wrote:
>
> > well, I just did -O3 -mpentium, and it both compiled cleanly, and appears
> > to be running okay, so is -O the max that makes a difference, or...?
>
> Or...
>
> Try build world with t
:>
:> -O6 is too much; -O3 is the highest level tested for by egcs.
:
:But specifying something too high (-O99) doesn't hurt - I'm using -O6 for
:gcc 2.95.1 (which, by the way, compiles almost everything in 3.3-RELEASE
:and 4.0-CURRENT, the only thing still troubling me with it is the kernel).
:
On Thu, 23 Sep 1999, Dan Nelson wrote:
> > Works rather nicely here with
> >
> > -O6 -mpentiumpro -march=pentiumpro -pipe -s -fexpensive-optimizations
> > -ffast-math
>
> -O6 is too much; -O3 is the highest level tested for by egcs.
But specifying something too high (-O99) doesn't hurt - I'm u
:On Thu, 23 Sep 1999 08:00:50 -0300, The Hermit Hacker wrote:
:
:> just curious, but what is the max setting that can be used to compile the
:> kernel?
:
:-O
:
:Ciao,
:Sheldon.
I've found -Os to produce quite excellent code. It's an option specific
to the new EGCS compiler (known as 'GCC
In the last episode (Sep 23), [EMAIL PROTECTED] said:
> On Thu, 23 Sep 1999, The Hermit Hacker wrote:
> > just curious, but what is the max setting that can be used to
> > compile the kernel?
>
> Works rather nicely here with
>
> -O6 -mpentiumpro -march=pentiumpro -pipe -s -fexpensive-optimizati
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
>> > well, I just did -O3 -mpentium, and it both compiled cleanly, and
>> > appears
>>
>> Try build world with that kernel running. :-)
>
> I should have mentioned that I'm speaking from fairly stale experience.
> I got bitten by this last year and
On Thu, 23 Sep 1999, Sheldon Hearn wrote:
>
>
> On Thu, 23 Sep 1999 08:23:09 -0300, The Hermit Hacker wrote:
>
> > well, I just did -O3 -mpentium, and it both compiled cleanly, and appears
> > to be running okay, so is -O the max that makes a difference, or...?
>
> Or...
>
> Try build world
On Thu, 23 Sep 1999 13:24:50 +0200, Sheldon Hearn wrote:
> > well, I just did -O3 -mpentium, and it both compiled cleanly, and appears
>
> Try build world with that kernel running. :-)
I should have mentioned that I'm speaking from fairly stale experience.
I got bitten by this last year and h
On 23 Sep, Sheldon Hearn wrote:
>> well, I just did -O3 -mpentium, and it both compiled cleanly, and appears
>> to be running okay, so is -O the max that makes a difference, or...?
>
> Or...
>
> Try build world with that kernel running. :-)
Just to let you know: -Os works for me (since egcs hi
On Thu, 23 Sep 1999, The Hermit Hacker wrote:
> just curious, but what is the max setting that can be used to compile the
> kernel?
Works rather nicely here with
-O6 -mpentiumpro -march=pentiumpro -pipe -s -fexpensive-optimizations
-ffast-math
LLaP
bero
To Unsubscribe: send mail to [EMAIL
On Thu, 23 Sep 1999 08:23:09 -0300, The Hermit Hacker wrote:
> well, I just did -O3 -mpentium, and it both compiled cleanly, and appears
> to be running okay, so is -O the max that makes a difference, or...?
Or...
Try build world with that kernel running. :-)
Ciao,
Sheldon.
To Unsubscribe:
On Thu, 23 Sep 1999, Sheldon Hearn wrote:
>
>
> On Thu, 23 Sep 1999 08:00:50 -0300, The Hermit Hacker wrote:
>
> > just curious, but what is the max setting that can be used to compile the
> > kernel?
>
> -O
well, I just did -O3 -mpentium, and it both compiled cleanly, and appears
to be runn
On Thu, 23 Sep 1999 08:00:50 -0300, The Hermit Hacker wrote:
> just curious, but what is the max setting that can be used to compile the
> kernel?
-O
Ciao,
Sheldon.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
just curious, but what is the max setting that can be used to compile the
kernel?
looking through the man page, I can do something like:
-O3 -march=pentium
and, if I'm reading the man page right:
-mcpu=cpu type
Assume the defaults for the machine type CPU TYPE
27 matches
Mail list logo
