http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
Uroš Bizjak ubizjak at gmail dot com changed:
What|Removed |Added
Status|NEW |ASSIGNED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #14 from H.J. Lu hjl.tools at gmail dot com ---
(In reply to Uroš Bizjak from comment #13)
Patch:
--cut here--
Index: config/i386/i386.c
===
--- config/i386/i386.c
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #15 from Thiago Macieira thiago at kde dot org ---
(In reply to H.J. Lu from comment #14)
I think HLE is the part of TSX.
It is and should be removed from the list.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #16 from Uroš Bizjak ubizjak at gmail dot com ---
(In reply to Thiago Macieira from comment #15)
(In reply to H.J. Lu from comment #14)
I think HLE is the part of TSX.
It is and should be removed from the list.
OK.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #17 from Jakub Jelinek jakub at gcc dot gnu.org ---
I'm not sure about that, after all we enable PTA_HLE even for generic.
The point is whether it might be beneficial to emit the HLE prefixes or not, on
CPUs that never have HLE
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #18 from Uroš Bizjak ubizjak at gmail dot com ---
(In reply to Jakub Jelinek from comment #17)
I'm not sure about that, after all we enable PTA_HLE even for generic.
The point is whether it might be beneficial to emit the HLE
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #19 from Thiago Macieira thiago at kde dot org ---
The prefix can be emitted for any CPU, you don't need a flag for that. However,
you cannot emit the XTEST instruction unless the CPU has HLE or RTM.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #20 from Uroš Bizjak ubizjak at gmail dot com ---
(In reply to Thiago Macieira from comment #19)
The prefix can be emitted for any CPU, you don't need a flag for that.
However, you cannot emit the XTEST instruction unless the CPU has
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #21 from uros at gcc dot gnu.org ---
Author: uros
Date: Thu May 8 19:13:10 2014
New Revision: 210234
URL: http://gcc.gnu.org/viewcvs?rev=210234root=gccview=rev
Log:
PR target/59952
* config/i386/i386.c (PTA_HASWELL): Remove
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #22 from uros at gcc dot gnu.org ---
Author: uros
Date: Thu May 8 19:23:45 2014
New Revision: 210236
URL: http://gcc.gnu.org/viewcvs?rev=210236root=gccview=rev
Log:
PR target/59952
* config/i386/i386.c (PTA_HASWELL): Remove
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #23 from uros at gcc dot gnu.org ---
Author: uros
Date: Thu May 8 19:32:34 2014
New Revision: 210237
URL: http://gcc.gnu.org/viewcvs?rev=210237root=gccview=rev
Log:
PR target/59952
* config/i386/i386.c
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
Uroš Bizjak ubizjak at gmail dot com changed:
What|Removed |Added
Target||x86
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #12 from Thiago Macieira thiago at kde dot org ---
GCC 4.9.0 got released with -march=haswell still enabling RTM and HLE, even
though there are Haswell parts without TSX.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #11 from Yukhin Kirill kirill.yukhin at intel dot com ---
(In reply to Yukhin Kirill from comment #9)
(In reply to Jakub Jelinek from comment #6)
Prerelease samples shouldn't count, people using those just can avoid using
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #2 from Thiago Macieira thiago at kde dot org ---
The Ks and the Rs don't have it. The plain 4770 as well as 4770S do have it. I
guess we were too eager with TSX when the announcement came out 2 years ago: it
sounded like all Haswell
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #3 from Mikael Pettersson mikpelinux at gmail dot com ---
There are also Haswells that lack BMI2. I updated our dynamic binary
instrumentation engine for AVX2 about a year ago, but while our Haswell box at
the time had AVX2 it didn't
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #4 from Thiago Macieira thiago at kde dot org ---
(In reply to Mikael Pettersson from comment #3)
There are also Haswells that lack BMI2. I updated our dynamic binary
instrumentation engine for AVX2 about a year ago, but while our
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #5 from Mikael Pettersson mikpelinux at gmail dot com ---
(In reply to Thiago Macieira from comment #4)
(In reply to Mikael Pettersson from comment #3)
There are also Haswells that lack BMI2. I updated our dynamic binary
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #6 from Jakub Jelinek jakub at gcc dot gnu.org ---
Prerelease samples shouldn't count, people using those just can avoid using
-march=haswell and use -march=ivybridge -mavx2 or similar instead. Can anyone
from Intel verify if all
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #7 from Thiago Macieira thiago at kde dot org ---
(In reply to Jakub Jelinek from comment #6)
Prerelease samples shouldn't count, people using those just can avoid using
-march=haswell and use -march=ivybridge -mavx2 or similar
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #8 from Thiago Macieira thiago at kde dot org ---
(In reply to Thiago Macieira from comment #7)
(In reply to Jakub Jelinek from comment #6)
Prerelease samples shouldn't count, people using those just can avoid using
-march=haswell
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #9 from Yukhin Kirill kirill.yukhin at intel dot com ---
(In reply to Jakub Jelinek from comment #6)
Prerelease samples shouldn't count, people using those just can avoid using
-march=haswell and use -march=ivybridge -mavx2 or
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
--- Comment #10 from Thiago Macieira thiago at kde dot org ---
(In reply to Yukhin Kirill from comment #9)
I am checking and will get back to. I though all AVX2 parts had BMI1/2
The evidence seems to support that. I did check with a few
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59952
Jakub Jelinek jakub at gcc dot gnu.org changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last
24 matches
Mail list logo