https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
Richard Biener changed:
What|Removed |Added
Last reconfirmed|2012-10-06 07:54:57 |2021-2-11
--- Comment #30 from Richard
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #29 from Richard Biener ---
On the original testcase I now get
> ./f951 -quiet -Ofast t.f90 -march=core-avx2 -fopt-info-vec
t.f90:157:0: optimized: loop vectorized using 32 byte vectors
t.f90:158:0: optimized: basic block part vect
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #28 from rguenther at suse dot de ---
On Tue, 8 Nov 2016, ro at CeBiTec dot Uni-Bielefeld.DE wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
>
> --- Comment #27 from ro at CeBiTec dot Uni-Bielefeld.DE Uni-Bielefeld.DE>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #27 from ro at CeBiTec dot Uni-Bielefeld.DE ---
> --- Comment #26 from Richard Biener ---
> I believe that https://gcc.gnu.org/ml/gcc-patches/2016-11/msg00686.html will
> fix this (currently in testing).
While it fixes the failures
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #26 from Richard Biener ---
I believe that https://gcc.gnu.org/ml/gcc-patches/2016-11/msg00686.html will
fix this (currently in testing).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
Rainer Orth changed:
What|Removed |Added
CC||ro at gcc dot gnu.org
--- Comment #25 from
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #24 from Richard Biener ---
Oh, and the duplicate costing is from the repeated use of pol_z(:,0,kg) and
similar terms. It's counted only once during scalar cost compute and not at
all
if not all of the uses are vectorized.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #23 from Richard Biener ---
Author: rguenth
Date: Mon Nov 7 08:06:08 2016
New Revision: 241893
URL: https://gcc.gnu.org/viewcvs?rev=241893&root=gcc&view=rev
Log:
2016-11-07 Richard Biener
PR tree-optimization/37150
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #22 from Richard Biener ---
Not costing redundant permutations (using a too trival implementation but good
enough for this case):
t.f90:158:0: note: Cost model analysis:
Vector inside of basic block cost: 984
Vector prologue cost
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #21 from Richard Biener ---
Ok, so fixing the accounting to disregard obviously dead loads gets us to
t.f90:158:0: note: Cost model analysis:
Vector inside of basic block cost: 1224
Vector prologue cost: 0
Vector epilogue cost:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #20 from Dominique d'Humieres ---
> (maybe you can benchmark with -fno-vect-cost-model?)
Compiling the code in comment 0 with -fno-vect-cost-model leads to
% grep mulpd pr37150.s | wc -l
142
% grep mulsd pr37150.s | wc -l
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #19 from Richard Biener ---
We do find SLP opportunities but in the end fail to vectorize with AVX2
because of
t.f90:158:0: note: BB vectorization with gaps at the end of a load is not
supported
t.f90:158:0: note: not vectorized: rel
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #18 from Dominique d'Humieres ---
> Still a current issue?
The test in comment 0 is not vectorized with trunk (7.0) at r241833.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
Thomas Koenig changed:
What|Removed |Added
CC||tkoenig at gcc dot gnu.org
--- Comment #
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #16 from Richard Biener ---
(In reply to Joost VandeVondele from comment #15)
> Created attachment 29738 [details]
> maybe smaller testcase version ?
>
> Attached is what I think is roughly the smallest kernel of this type that we
>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
--- Comment #15 from Joost VandeVondele
2013-03-27 12:53:16 UTC ---
Created attachment 29738
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=29738
maybe smaller testcase version ?
Attached is what I think is roughly the smallest ker
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37150
Richard Biener changed:
What|Removed |Added
Status|NEW |ASSIGNED
AssignedTo|una
17 matches
Mail list logo
