On Mon, Jan 13, 2014 at 5:37 AM, Richard Biener wrote:
> On Wed, 27 Nov 2013, Jakub Jelinek wrote:
>
>> On Wed, Nov 27, 2013 at 10:53:56AM +0100, Richard Biener wrote:
>> > Hmm. I'm still thinking that we should handle this during the regular
>> > transform step.
>>
>> I wonder if it can't be don
On Tue, 14 Jan 2014, Jakub Jelinek wrote:
> On Tue, Jan 14, 2014 at 10:01:06AM +0100, Richard Biener wrote:
> > Jakub, adding the new flag is ok with me.
>
> So like this?
Ok if it passes testing.
Thanks,
Richard.
> 2014-01-14 Jakub Jelinek
>
> * tree-vectorizer.h (struct _loop_vec_i
On Tue, Jan 14, 2014 at 10:01:06AM +0100, Richard Biener wrote:
> Jakub, adding the new flag is ok with me.
So like this?
2014-01-14 Jakub Jelinek
* tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
field.
(LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
On Tue, 14 Jan 2014, Richard Biener wrote:
> On Mon, 13 Jan 2014, Cong Hou wrote:
>
> > I noticed that LIM could not hoist vector invariant, and that is why
> > my first implementation tries to hoist them all.
>
> Yes, I filed PR59786 for this. I'll see if I can come up with
> a fix suitable fo
On Mon, 13 Jan 2014, Cong Hou wrote:
> I noticed that LIM could not hoist vector invariant, and that is why
> my first implementation tries to hoist them all.
Yes, I filed PR59786 for this. I'll see if I can come up with
a fix suitable for stage3.
> In addition, there are two disadvantages of h
I noticed that LIM could not hoist vector invariant, and that is why
my first implementation tries to hoist them all.
In addition, there are two disadvantages of hoisting invariant load +
lim method:
First, for some instructions the scalar version is faster than the
vector version, and in this ca
On Mon, Jan 13, 2014 at 02:37:38PM +0100, Richard Biener wrote:
> 2014-01-13 Richard Biener
>
> PR tree-optimization/58921
> PR tree-optimization/59006
> * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
> hoisting invariant stmts.
> * tree-vect-stmts.c
On Wed, 27 Nov 2013, Jakub Jelinek wrote:
> On Wed, Nov 27, 2013 at 10:53:56AM +0100, Richard Biener wrote:
> > Hmm. I'm still thinking that we should handle this during the regular
> > transform step.
>
> I wonder if it can't be done instead just in vectorizable_load,
> if LOOP_REQUIRES_VERSION
On Wed, Nov 27, 2013 at 1:53 AM, Richard Biener wrote:
> On Fri, 22 Nov 2013, Cong Hou wrote:
>
>> Hi
>>
>> Currently in GCC vectorization, some loop invariant may be detected
>> after aliasing checks, which can be hoisted outside of the loop. The
>> current method in GCC may break the information
On Wed, 27 Nov 2013, Jakub Jelinek wrote:
> On Wed, Nov 27, 2013 at 12:54:14PM +0100, Richard Biener wrote:
> > On Wed, 27 Nov 2013, Jakub Jelinek wrote:
> >
> > > On Wed, Nov 27, 2013 at 10:53:56AM +0100, Richard Biener wrote:
> > > > Hmm. I'm still thinking that we should handle this during th
On Wed, Nov 27, 2013 at 12:54:14PM +0100, Richard Biener wrote:
> On Wed, 27 Nov 2013, Jakub Jelinek wrote:
>
> > On Wed, Nov 27, 2013 at 10:53:56AM +0100, Richard Biener wrote:
> > > Hmm. I'm still thinking that we should handle this during the regular
> > > transform step.
> >
> > I wonder if
On Wed, 27 Nov 2013, Jakub Jelinek wrote:
> On Wed, Nov 27, 2013 at 10:53:56AM +0100, Richard Biener wrote:
> > Hmm. I'm still thinking that we should handle this during the regular
> > transform step.
>
> I wonder if it can't be done instead just in vectorizable_load,
> if LOOP_REQUIRES_VERSION
On Wed, Nov 27, 2013 at 10:53:56AM +0100, Richard Biener wrote:
> Hmm. I'm still thinking that we should handle this during the regular
> transform step.
I wonder if it can't be done instead just in vectorizable_load,
if LOOP_REQUIRES_VERSIONING_FOR_ALIAS (loop_vinfo) and the load is
invariant, j
On Fri, 22 Nov 2013, Cong Hou wrote:
> Hi
>
> Currently in GCC vectorization, some loop invariant may be detected
> after aliasing checks, which can be hoisted outside of the loop. The
> current method in GCC may break the information built during the
> analysis phase, causing some crash (see PR5
Hi
Currently in GCC vectorization, some loop invariant may be detected
after aliasing checks, which can be hoisted outside of the loop. The
current method in GCC may break the information built during the
analysis phase, causing some crash (see PR59006 and PR58921).
This patch improves the loop i
15 matches
Mail list logo
