http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38201
H.J. Lu hjl.tools at gmail dot com changed:
What|Removed |Added
Status|NEW |RESOLVED
--- Comment #16 from hjl dot tools at gmail dot com 2008-12-09 21:58
---
A patch is posted at
http://gcc.gnu.org/ml/gcc-patches/2008-12/msg00585.html
--
hjl dot tools at gmail dot com changed:
What|Removed |Added
--- Comment #15 from hjl dot tools at gmail dot com 2008-12-03 16:50
---
Simulator is fine. AVX executable can only run on simulator. If
there is a simulator which can run SSE5 and AVX, we will add
a new switch for it.
--
hjl dot tools at gmail dot com changed:
What
--- Comment #14 from pinskia at gcc dot gnu dot org 2008-11-23 18:28
---
But -mavx -msse5 will generate codes which won't run on any machines.
It could run on a simulator that has both (or a new processor which has not
come out yet). Or are there conflicts with the opcodes
--- Comment #10 from rguenth at gcc dot gnu dot org 2008-11-22 15:03
---
We should have -mfma to enable a fused multiply-add instruction that is
available
when enabling either -msse5 or -mavx. -mfma should not itself enable any
of the instruction set enabling features.
HJ, why did
--- Comment #11 from hjl dot tools at gmail dot com 2008-11-22 15:09
---
(In reply to comment #10)
We should have -mfma to enable a fused multiply-add instruction that is
available
when enabling either -msse5 or -mavx. -mfma should not itself enable any
of the instruction set
--- Comment #12 from hjl dot tools at gmail dot com 2008-11-22 15:15
---
Richard asked:
Why should it (-mavx -msse5) be disallowed if a user asks for it? Do we
disallow -msse4a -mssse4?
Reply:
-msse4a -mssse4 can generate code which runs if you check the feature
bit in CPUID before
--- Comment #13 from rguenth at gcc dot gnu dot org 2008-11-22 15:31
---
I see.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38201
--- Comment #8 from Joey dot ye at intel dot com 2008-11-21 12:00 ---
In short, set A={-favx, -ffma}, set B={-f3dnow, -f3dnowa, -fsse4a, -fsse5}. Any
option combination from both sets should be prohibited.
Please add more options into these set in case I missed any.
--
--- Comment #9 from hjl dot tools at gmail dot com 2008-11-21 13:35 ---
(In reply to comment #8)
In short, set A={-favx, -ffma}, set B={-f3dnow, -f3dnowa, -fsse4a, -fsse5}.
Any
It is -mXXX, not -fXXX.
option combination from both sets should be prohibited.
That is correct.
--
--- Comment #1 from dwarak dot rajagopal at amd dot com 2008-11-20 16:48
---
1) -msse5 includes -mfma switch (because fma is a part of sse5 instructions).
So having -msse5 -mfma is same as having just msse5, though you can just
have -fma (without -msse5).
2) -mavx -msse5 = Yes. This
--- Comment #2 from hjl dot tools at gmail dot com 2008-11-20 16:57 ---
(In reply to comment #1)
1) -msse5 includes -mfma switch (because fma is a part of sse5 instructions).
So having -msse5 -mfma is same as having just msse5, though you can just
have -fma (without -msse5).
Please
--- Comment #3 from hjl dot tools at gmail dot com 2008-11-20 18:52 ---
Since -mfma implies -mavx, we got
[EMAIL PROTECTED] gcc]$ cat f.c
double f;
void
foo (double x, double y, double z)
{
f = x * y + z;
}
[EMAIL PROTECTED] gcc]$ ./xgcc -B./ -O2 -mfma -msse5 f.c -S
--- Comment #4 from dwarak dot rajagopal at amd dot com 2008-11-20 19:35
---
Yes, you are right. -mfma -msse5 does not make sense. I mistook -mfma for
-mfused-madd and hence the confusion.
Hence these combinations (1 and 2) does not make sense.
Thanks,
Dwarak
--
--- Comment #5 from hjl dot tools at gmail dot com 2008-11-20 19:46 ---
(In reply to comment #4)
Yes, you are right. -mfma -msse5 does not make sense. I mistook -mfma for
-mfused-madd and hence the confusion.
Hence these combinations (1 and 2) does not make sense.
Should we
--- Comment #6 from dwarak dot rajagopal at amd dot com 2008-11-20 19:49
---
Should we disallow such combinations?
Yes.
- Dwarak
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38201
--- Comment #7 from hjl dot tools at gmail dot com 2008-11-20 21:29 ---
We have the same issue with -m3dnow, -m3dnowa and -msse4a.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38201
17 matches
Mail list logo