https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #30 from Richard Biener ---
Author: rguenth
Date: Tue Feb 19 12:46:48 2019
New Revision: 269015
URL: https://gcc.gnu.org/viewcvs?rev=269015=gcc=rev
Log:
2019-02-19 Richard Biener
PR middle-end/88074
* toplev.c
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #29 from Jakub Jelinek ---
Author: jakub
Date: Sat Feb 16 18:46:04 2019
New Revision: 268962
URL: https://gcc.gnu.org/viewcvs?rev=268962=gcc=rev
Log:
PR middle-end/88074
* simplify.c
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #28 from Jakub Jelinek ---
Created attachment 45676
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45676=edit
gcc9-pr88074.patch
Incremental patch to tweak norm2 simplification. If e->value.real**2 would
overflow or be close
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Richard Biener changed:
What|Removed |Added
Status|ASSIGNED|NEW
CC|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Jakub Jelinek changed:
What|Removed |Added
CC||burnus at gcc dot gnu.org,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Richard Biener changed:
What|Removed |Added
Priority|P3 |P2
--- Comment #25 from Richard Biener
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Richard Biener changed:
What|Removed |Added
Target Milestone|7.4 |7.5
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #24 from rguenther at suse dot de ---
On Wed, 21 Nov 2018, joseph at codesourcery dot com wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
>
> --- Comment #23 from joseph at codesourcery dot com dot com> ---
> And, yes,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #23 from joseph at codesourcery dot com ---
And, yes, at least one extra bit in emin is needed for that sticky
rounding code to work (because the user's source code may have a decimal
constant that is slightly over half the least
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #22 from joseph at codesourcery dot com ---
On Wed, 21 Nov 2018, rguenther at suse dot de wrote:
> /* Nonzero value, possibly overflowing or underflowing. */
> mpfr_init2 (m, SIGNIFICAND_BITS);
> inexact =
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Richard Biener changed:
What|Removed |Added
Attachment #45039|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #20 from rguenther at suse dot de ---
On Tue, 20 Nov 2018, joseph at codesourcery dot com wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
>
> --- Comment #19 from joseph at codesourcery dot com dot com> ---
> On Tue,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #19 from joseph at codesourcery dot com ---
On Tue, 20 Nov 2018, rguenth at gcc dot gnu.org wrote:
> if (fmt->emin < min_exp)
> min_exp = fmt->emin - fmt->p + 1;
> so somehow the formula fmt->emin
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #18 from Richard Biener ---
Author: rguenth
Date: Tue Nov 20 12:18:58 2018
New Revision: 266313
URL: https://gcc.gnu.org/viewcvs?rev=266313=gcc=rev
Log:
2018-11-20 Richard Biener
PR tree-optimization/88074
*
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #17 from Richard Biener ---
Created attachment 45039
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45039=edit
not working patch
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Richard Biener changed:
What|Removed |Added
CC||jsm28 at gcc dot gnu.org
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #15 from Richard Biener ---
(In reply to jos...@codesourcery.com from comment #13)
> MPFR deals with larger exponent ranges for intermediate computations
> itself (MPFR functions generally set the maximum possible exponent range
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #14 from Richard Biener ---
(In reply to jos...@codesourcery.com from comment #12)
> Note that such issues are not unique to ctan.
>
> E.g., compiling __builtin_jn (10, 1e10) will produce such a hang.
> The difference is
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #13 from joseph at codesourcery dot com ---
MPFR deals with larger exponent ranges for intermediate computations
itself (MPFR functions generally set the maximum possible exponent range
internally). MPC generally doesn't seem to,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #12 from joseph at codesourcery dot com ---
Note that such issues are not unique to ctan.
E.g., compiling __builtin_jn (10, 1e10) will produce such a hang.
The difference is that in the ctan case the additional algorithm
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #11 from Richard Biener ---
A GCC-side fix is to limit the exponent range mpfr uses (for TFmode this
is -16381 to 16384). Not sure if mpfr or mpc need larger exponent ranges
for intermediate computations...
Lazily initialized in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #10 from Richard Biener ---
OK, so it seems to be about large imagparts only and only if the realpart is
non-zero. Thus in GCC we could stop folding ctan for imagpart (z) > 1000 for
example. Like sin and cos tan is _usually_
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #9 from Richard Biener ---
It has something to do with large real/imagparts, the actual numbers are not so
interesting. Somebody why knows how complex tan behaves might suggest a
workaround for GCC to drop to Inf/NaN at some point?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #8 from Richard Biener ---
See gmp-bugs followup. (all other mails still wait for moderators)
https://gmplib.org/list-archives/gmp-bugs/2018-November/004440.html
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #7 from Andrew Pinski ---
(In reply to Richard Biener from comment #6)
> but I'll try what we recommend as prerequesites now.
This is what I used and it causes the same issue:
GNU C++11 (Marvell Development Version) version 7.3.0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #6 from Richard Biener ---
Reported to the GMP mailinglist, not sure which libaries fault it is.
To complete library version list, GCC reports
GNU C17 (SUSE Linux) version 8.2.1 20180831 [gcc-8-branch revision 264010]
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #5 from Richard Biener ---
C testcase:
#include
int main()
{
_Complex double x;
__real x = 3.09126495087690770626068115234375e+8;
__imag x = -4.045689747817175388336181640625e+8;
volatile _Complex double y = ctan (x);
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #4 from Richard Biener ---
I'm using mpc 1.0.2.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
--- Comment #3 from Richard Biener ---
Hmm, we never return from __gmpn_sqr_basecase_x86_64 ...
#0 0x7734b92b in __gmpn_addlsh2_n_x86_64 ()
from /usr/lib64/libgmp.so.10
#1 0x004c in ?? ()
#2 0x7fff2c50 in ?? ()
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Richard Biener changed:
What|Removed |Added
Status|NEW |ASSIGNED
Known to work|
30 matches
Mail list logo
