For -flto-partition=balanced, if user does not specify the --param
lto-partitions and --param lto-min-partition,
will GCC partition the program?
If yes what is the criteria?
Where can we get more help on using the --param lto-partitions and
--param lto-min-partition?
We want to know how the pa
On Thu, May 30, 2013 at 2:46 AM, Dehao Chen wrote:
> Hi,
>
> In tree-vect-loop.c, it limits the vectorization only to loops that have 2
> BBs:
>
> /* Inner-most loop. We currently require that the number of BBs is
> exactly 2 (the header and latch). Vectorizable inner-most loops
On Fri, May 31, 2013 at 9:21 AM, Swati wrote:
> For -flto-partition=balanced, if user does not specify the --param
> lto-partitions and --param lto-min-partition,
> will GCC partition the program?
> If yes what is the criteria?
It will partition the program into lto-partitions partitions unless
t
Hi,
During studying ivopt pass, I found the cost of preloop calculations
are inaccurately calculated in many scenarios.
There are two kinds of preloop calculations: base of candidates and
invariant part of iv use representation.
For base of iv candidates, it is calculated as below:
cost_base = fo
On Fri, May 31, 2013 at 10:20:01AM +0200, Richard Biener wrote:
> The limit is there because a loop with more than one basic-block with code
> necessarily has to have conditionally executed BBs and eventually PHI nodes
> at merge points.
>
> Now, it may be that we properly determine if we can hand
The GNU Compiler Collection version 4.8.1 has been released.
GCC 4.8.1 is the first bug-fix release containing important fixes for
regressions and serious bugs in GCC 4.8.0 with over 91 bugs fixed since
the previous release. Support for C++11 ref-qualifiers has been added
to GCC 4.8.1, making G++
Status
==
GCC 4.8.1 has been released, the branch is again open for regression
bugfixes and documentation fixes. GCC 4.8.2 could be tentatively released
in early September.
Quality Data
Priority # Change from Last Report
--- ---
On 31 May 2013 12:16, Jakub Jelinek wrote:
> The GNU Compiler Collection version 4.8.1 has been released.
>
> GCC 4.8.1 is the first bug-fix release containing important fixes for
> regressions and serious bugs in GCC 4.8.0 with over 91 bugs fixed since
> the previous release. Support for C++11 re
On 05/31/2013 10:20 AM, Richard Biener wrote:
So - I doubt that you both do not get any ICEs and more performance.
I added the second suggested patch:
Index: tree-vect-loop-manip.c
===
--- tree-vect-loop-manip.c (revision 19
On Fri, May 31, 2013 at 3:21 PM, Toon Moene wrote:
> On 05/31/2013 10:20 AM, Richard Biener wrote:
>
>> So - I doubt that you both do not get any ICEs and more performance.
>
>
> I added the second suggested patch:
>
> Index: tree-vect-loop-manip.c
> ===
On Fri, May 31, 2013 at 03:21:51PM +0200, Toon Moene wrote:
> SUBROUTINE XYZ(A, B, N)
> DIMENSION A(N), B(N)
> DO I = 1, N
>IF (A(I) > 0.0) THEN
> A(I) = B(I) / A(I)
>ELSE
> A(I) = B(I)
>ENDIF
> ENDDO
> END
Well, in this case (with -Ofast) it is just the case that ifcvt
or
On 05/31/2013 03:41 PM, Jakub Jelinek wrote:
On Fri, May 31, 2013 at 03:21:51PM +0200, Toon Moene wrote:
SUBROUTINE XYZ(A, B, N)
DIMENSION A(N), B(N)
DO I = 1, N
IF (A(I)> 0.0) THEN
A(I) = B(I) / A(I)
ELSE
A(I) = B(I)
ENDIF
ENDDO
END
Well, in this case (with -Ofas
On Fri, May 31, 2013 at 3:48 PM, Toon Moene wrote:
> On 05/31/2013 03:41 PM, Jakub Jelinek wrote:
>
>> On Fri, May 31, 2013 at 03:21:51PM +0200, Toon Moene wrote:
>
>
>>> SUBROUTINE XYZ(A, B, N)
>>> DIMENSION A(N), B(N)
>>> DO I = 1, N
>>> IF (A(I)> 0.0) THEN
>>>A(I) = B(I) / A(I)
>>>
On Fri, May 31, 2013 at 03:48:59PM +0200, Toon Moene wrote:
> >If you rewrite the above into:
> >SUBROUTINE XYZ(A, B, N)
> >DIMENSION A(N), B(N)
> >DO I = 1, N
> >C = B(I)
> >IF (A(I)> 0.0) THEN
> > A(I) = C / A(I)
> >ELSE
> > A(I) = C
> >ENDIF
> >ENDDO
> >END
> >
> >th
On 05/31/2013 03:54 PM, Jakub Jelinek wrote:
> I wrote:
But this "inner loop" has at least 3 basic blocks - so what does the
"loop->num_nodes != 2" test exactly codify ?
With the above testcase it has just 2.
Before ifcvt pass it still has 4:
Ah, I missed that subtle part. So my example is
On 05/30/2013 04:30 PM, Ryan Johnson wrote:
> Is there a way for libgcc_s to interpose on dlopen/dlclose if (and only if)
> those are present? If so, the wrappers could increment an atomic version
> counter, which would be plenty accurate for invalidating an object header
> cache, without requiring
On Fri, May 31, 2013 at 6:54 AM, Jakub Jelinek wrote:
> On Fri, May 31, 2013 at 03:48:59PM +0200, Toon Moene wrote:
>> >If you rewrite the above into:
>> >SUBROUTINE XYZ(A, B, N)
>> >DIMENSION A(N), B(N)
>> >DO I = 1, N
>> >C = B(I)
>> >IF (A(I)> 0.0) THEN
>> > A(I) = C / A(I)
>> >
yes, loop vectorizer relies on early passes to straighten out control
flow (unswitch, index splitting, loop distribution, ifcvt etc). Intel
ICC is pretty good at it. For the following simple made up case, icc
vectorizes the loop.
int a[1];
int b[1];
int foo (int n)
{
int i;
for (i
Hi list!
I was having a discussion with a coworker earlier today about whether
it's safe to say that:
> MyClass some_function()
> {
> MyClass k(/*possibly some arguments here*/);
>
> // Fill up k somehow.
>
> return k;
> }
will never call MyClass' copy constructor, assuming recent ve
19 matches
Mail list logo
