David Daney wrote:
Joe Buck wrote:
Maybe there could be a "semi-primary" or "experimental primary" status;
a feature could be treated as primary, but with the understanding that
the requirement will be waived if it causes excessive delay. The
"experimental" label could be dropped after a few su
Hi,
SH2A, SH4 and SH4A FPU targets support mode switching i.e. switching
between single precision and double precision. Double-precision FPU
is also available for the above mentioned targets.
The option "-mfmovd" is enabled by default for SH2A which generates
"fmov.d" instruction by default. Ho
On the GCC site (http://gcc.gnu.org/contribute.html#testing) under
"Which Tests to Perform" it is stated:
"If your change is to code that is not in a front end, or is to the C
front end You must bootstrap all languages, not just C."
Does that mean Ada, Objective-C++, and treelang, too? Or ju
On 2/19/08 2:27 PM, Fran Baena wrote:
Hi everybody,
i am studing how gcc carries out Alias Representation and some questions appear.
For instance, given this code portion:
if ( ... )
p = &a;
else
if ( ... )
p = &b;
else
p = &c;
a = 5;
b = 3;
d = *p4;
My questions are
Use of long double math builtins on powerpc-darwin is the cause of the
remaining gfortran issues on this target (at least, we can't
investigate much until this is fixed). An example case is simple: the
following gives wrong results.
$ cat s.c
int main (void)
{
long double x;
double y;
x = -
I very grateful for your help and wisdom
Testcase and MD Patch attached
unsigned long f (unsigned char *P)
{
unsigned long C;
C = ((unsigned long)P[1] << 24)
| ((unsigned long)P[2] << 16)
| ((unsigned long)P[3] << 8)
| ((unsigned long)P[4] << 0);
return C;
}
Index: avr.md
Putting fwprop after combine is no problem - but is too early - none of
the patterns would be split at that time - preventing byte level
propagations.
Yeah, I meant "after split" actually.
Anyway, the problem is that if the RHS becomes a constant, fwprop does
not propagate the LHS anymore.
Paolo
I placed extra fwprop before local-alloc as this was just before NOP
got created and after splitting.
Putting fwprop after combine is no problem - but is too early - none of
the patterns would be split at that time - preventing byte level
propagations.
As register usage as well as c
If I understand correctly:
Prop. of "0" causes simplfy-rtx to create NOP from OR Rx,0
This NOP (deletion?) creates another set of potential uses - as now
the prior RHS def now passes straight thru to a new set of uses - but
we miss those new uses. (which in the testcase are often 0)
I will
On Wed, 20 Feb 2008, "Weddington, Eric" <[EMAIL PROTECTED]> wrote:
> > Maybe there could be a "semi-primary" or "experimental
> > primary" status;
> > a feature could be treated as primary, but with the understanding that
> > the requirement will be waived if it causes excessive delay. The
> > "ex
Hi. I'm the owner of Russian annual programming contest site
http://zcontest.ru
The contestants are usually use GCC compiler as their main platform
for development. So I decided to put GCC archives on my site.
Fell free to post link on it at this page:
http://gcc.gnu.org/mirrors.html
My site loca
2008/2/21, Diego Novillo <[EMAIL PROTECTED]>:
> On 2/19/08 2:27 PM, Fran Baena wrote:
> > Hi everybody,
> >
> > i am studing how gcc carries out Alias Representation and some questions
> appear.
> >
> > For instance, given this code portion:
> >
> > if ( ... )
> >p = &a;
> > else
>
On Feb 21, 2008, at 6:10 AM, FX wrote:
Use of long double math builtins on powerpc-darwin
My question is simple: are there any plans to fix?
I don't know of anyone working on it. The issue is trivial enough to
fix, if people want to fix it. Essentially, the various builtins need
to hav
On 2/21/08 1:13 PM, Fran Baena wrote:
2008/2/21, Diego Novillo <[EMAIL PROTECTED]>:
On 2/19/08 2:27 PM, Fran Baena wrote:
> Hi everybody,
>
> i am studing how gcc carries out Alias Representation and some questions
appear.
>
> For instance, given this code portion:
>
> if ( ... )
>
Zdenek, you committed changes to tree-tailcall.c but you didn't fully
convert the file. Was that a mis-commit? The file does not compile and
uses PHI_RESULT instead of gimple_phi_result.
Thanks. Diego.
> "If your change is to code that is not in a front end, or is to the C
> front end You must bootstrap all languages, not just C."
>
> Does that mean Ada, Objective-C++, and treelang, too? Or just the
> languages enabled by default?
Just the languages enabled by default will do.
Ben
Hi,
> Zdenek, you committed changes to tree-tailcall.c but you didn't fully
> convert the file. Was that a mis-commit? The file does not compile and
> uses PHI_RESULT instead of gimple_phi_result.
the file compiles for me; it indeed uses PHI_RESULT, but since
that is equivalent to DEF_FROM_PT
If I have a CALL_EXPR how can I find out information about the call
site? For example, the filename and line number.
thanks,
-- dasarath
On Thu, Feb 21, 2008 at 1:45 PM, Zdenek Dvorak <[EMAIL PROTECTED]> wrote:
> Hi,
>
>
> > Zdenek, you committed changes to tree-tailcall.c but you didn't fully
> > convert the file. Was that a mis-commit? The file does not compile and
> > uses PHI_RESULT instead of gimple_phi_result.
>
> the fi
> "Dasarath" == Dasarath Weeratunge <[EMAIL PROTECTED]> writes:
Dasarath> If I have a CALL_EXPR how can I find out information about the call
Dasarath> site? For example, the filename and line number.
See EXPR_HAS_LOCATION, EXPR_LOCUS, etc.
In general I recommend reading tree.def and skimming
> The issue is trivial enough to fix, if people want to fix it.
> Essentially, the various builtins need to have different linkage names,
> not just _sin.
Those who kow how to fix it don't want to fix it, and those who want
to fix it don't how to fix it! Would it be possible to break this
vicious
Snapshot gcc-4.3-20080221 is now available on
ftp://gcc.gnu.org/pub/gcc/snapshots/4.3-20080221/
and on various mirrors, see http://gcc.gnu.org/mirrors.html for details.
This snapshot has been generated from the GCC 4.3 SVN branch
with the following options: svn://gcc.gnu.org/svn/gcc/branches
> I would vote for -ffast-math to disable it.
Please don't. I think parentheses should be obeyed in FORTRAN.
-ffast-math groups several useful optimization for usual codes
and I don't see why users should have to remember their names
if they want to keep mandatory parentheses.
Now I also think on
Dominique Dhumieres wrote:
I would vote for -ffast-math to disable it.
Please don't. I think parentheses should be obeyed in FORTRAN.
Does the Fortran standard say that (I assume that despite using
the name for the 66 language, you really mean Fortran :-)
-ffast-math groups several useful o
Tobias Burnus wrote:
According to the GCC 4.4 Release Criteria,
http://gcc.gnu.org/gcc-4.4/criteria.html, only C and C++ are primary
languages. And thus only C and C++ regressions can be release critical.
I propose to add Fortran to these languages.
Thank you for the suggestion.
My opinion,
Please do not check-in anything in the branch until I'm done figuring
out the various regressions in the last couple of days.
Folks, please run all tests with all languages to make sure no new
regressions are introduced by your patch (I think some of these
regressions are coming from one of
On Thu, 21 Feb 2008, Mike Stump wrote:
> On Feb 21, 2008, at 6:10 AM, FX wrote:
> > Use of long double math builtins on powerpc-darwin
>
> > My question is simple: are there any plans to fix?
>
> I don't know of anyone working on it.
Are current Darwin maintainers working on fixing anything in th
Paolo,
As you suggested, I moved the extra fwprop nearer combine, just after
split - but it failed to propagate anything.
The reason is that immediately post split the data flow is reflecting
cross dependencies between Word and subreg U/D. So the USE of just 1
QImode subreg of SImode regis
Naveen H.S. wrote:
> SH2A, SH4 and SH4A FPU targets support mode switching i.e. switching
> between single precision and double precision. Double-precision FPU
> is also available for the above mentioned targets.
>
> The option "-mfmovd" is enabled by default for SH2A which generates
> "fmov.d"
29 matches
Mail list logo
