Right now, pack/unpack
a) use memcpy, even if a native type is available
b) convert mask arguments of kind 1 and 2 to default
logical kind, which wastes cycles and memory
Case b) is actually true of all intrinsic functions with mask.
--
Summary: performance of pack/unpack
Since revision r126615:
2007-07-12 Andreas Schwab [EMAIL PROTECTED]
* gengtype-lex.l: Allow declarations to be indented.
Bootstrap of gcc on arm-netbsdelf has failed because
build/gengtype /work/rearnsha/gnusrc/gcc/trunk/gcc gtyp-input.list
There are major failures in the lapack 3.1.1 tests with gcc 2.95.3 and in the
full range up to gcc 4.2.1. The problem seems to appear on 32-bit i386, when
the routines chgeqz and zhgeqz are compiled with -O2.
With 2.95.3 I get:
cgd.out:
Matrix order=8, type=15, seed=2866, 214, 1,1861,
--- Comment #19 from s_j_newbury at yahoo dot co dot uk 2007-08-03 13:52
---
Created an attachment (id=14015)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14015action=view)
EABI sys_cacheflush support
Use EABI syscall method for sys_cacheflush when __ARM_EABI__ is defined
--
--- Comment #18 from s_j_newbury at yahoo dot co dot uk 2007-08-03 12:56
---
(In reply to comment #17)
Re comment #16: I've tested this, and it seems to work. What failures do you
expect?
/* How to make a trampoline. */
#define FFI_INIT_TRAMPOLINE(TRAMP,FUN,CTX) \
({ unsigned
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||pinskia at gcc dot gnu dot
|
--- Comment #98 from sliwa at cft dot edu dot pl 2007-08-03 12:09 ---
*** Bug 32976 has been marked as a duplicate of this bug. ***
--
sliwa at cft dot edu dot pl changed:
What|Removed |Added
--- Comment #1 from sliwa at cft dot edu dot pl 2007-08-03 12:09 ---
*** This bug has been marked as a duplicate of 323 ***
--
sliwa at cft dot edu dot pl changed:
What|Removed |Added
--- Comment #17 from aph at gcc dot gnu dot org 2007-08-03 12:02 ---
Re comment #16: I've tested this, and it seems to work. What failures do you
expect?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31325
--- Comment #16 from s_j_newbury at yahoo dot co dot uk 2007-08-03 11:52
---
(In reply to comment #14)
Actually, forget that last message. Most of these patches seem to be gcc 4.2
based and the libffi and gij patches are already done.
Andrew: I spotted your new patch here
--- Comment #7 from dir at lanl dot gov 2007-08-03 14:03 ---
I was looking in the CrashReported log for something else and I noticed that
ecj1 was hanging because it was crashing. Here is the crash log trace back -
**
Host Name: dir
Date/Time: 2007-07-13 10:52:37.582
--- Comment #5 from dnovillo at gcc dot gnu dot org 2007-08-03 13:12
---
Fixed. http://gcc.gnu.org/ml/gcc-patches/2007-08/msg00174.html.
--
dnovillo at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #1 from belyshev at depni dot sinp dot msu dot ru 2007-08-03
15:06 ---
Fails since:
r121570 | hubicka | 2007-02-04 16:01:13 +0300 (Sun, 04 Feb 2007) | 7 lines
* ipa-inline.c (try_inline): Improve debug output; work on already
inline edges too.
--- Comment #5 from pthaugen at us dot ibm dot com 2007-08-03 15:30 ---
This patch gives us an additional 2-3% overall on CPU2000, running on Power6.
Facerec was the big winner with about 40% improvement, there were a few
improvements in the 5-10% range and a few degradations in the
As of 20070730, all gfortran tests fail on alpha-dec-osf4.0f:
25729:./PR19754_2.exe: /sbin/loader: Error: Unresolved symbol in
/amnt/sequoia/export/vol/gcc/obj/alpha/gcc-4.3.0-20070730/4.0f-gcc/alpha-dec-osf4.0f/./libgfortran/.libs/libgfortran.so.3:
vsnprintf
25729:./PR19754_2.exe: /sbin/loader:
--- Comment #4 from bergner at gcc dot gnu dot org 2007-08-03 14:43 ---
Andrew's patch from Comment #3 bootstrapped and regtested with no regressions
on powerpc64-linux running the testsuite in both 32-bit and 64-bit modes.
I can also confirm that it fixes the test case I posted above.
--- Comment #4 from dnovillo at gcc dot gnu dot org 2007-08-03 13:08
---
Subject: Bug 31521
Author: dnovillo
Date: Fri Aug 3 13:08:29 2007
New Revision: 127180
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127180
Log:
2007-08-03 Andrew Pinski [EMAIL PROTECTED]
--- Comment #1 from fxcoudert at gcc dot gnu dot org 2007-08-03 16:25
---
I'm in Edinbourg airport right now, so I'll make this quick:
--- value_4.f90 2007-08-03 18:24:42.0 +0200
+++ /home/fxcoudert/devel/trunk/gcc/testsuite/gfortran.dg/value_4.f90
2007-07-03
// Compile with -O1 -finline-functions -fipa-cp
// reduced from ada/adadecode.c
static int f0 (char *s0, char *s1)
{
return __builtin_strlen (s0) __builtin_strlen (s1);
}
int f1 (char *s, int j)
{
if (f0 (s, ))
return 1;
return j;
}
void f2 (char *s)
{
f1 (s, 0);
}
//
bug.c:16:
Followed by manual, optional trailing text can be used to give more information
in the warning
message(http://gcc.gnu.org/onlinedocs/gcc-3.4.6/cpp/Pragmas.html#Pragmas -
'#pragma GCC dependency' section). However, proprecessor gives warning messages
when there is optional trailing text. And this
--- Comment #2 from schwab at suse dot de 2007-08-03 14:29 ---
Created an attachment (id=14017)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14017action=view)
Patch for gengtype-lex.l
Please try this patch. Ignoring bs-nl should be good enough here.
--
--- Comment #20 from aph at gcc dot gnu dot org 2007-08-03 14:33 ---
Please post patches to [EMAIL PROTECTED] Please provide a ChangeLog.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31325
--- Comment #1 from schwab at suse dot de 2007-08-03 14:29 ---
Created an attachment (id=14016)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14016action=view)
Patch for gengtype-lex.l
Please try this patch. Ignoring bs-nl should be good enough here.
--
schwab at suse dot de
--- Comment #2 from ro at gcc dot gnu dot org 2007-08-03 14:23 ---
Fixed by this patch:
http://gcc.gnu.org/ml/java-patches/2007-q3/msg00103.html
--
ro at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #4 from pinskia at gcc dot gnu dot org 2007-08-03 12:17 ---
%m means it can represent it as memory access which means access it via a
register, this is not a bug.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||pinskia at gcc dot gnu dot
|
Since revision r126615:
2007-07-12 Andreas Schwab [EMAIL PROTECTED]
* gengtype-lex.l: Allow declarations to be indented.
Bootstrap of gcc on arm-netbsdelf has failed because
build/gengtype /work/rearnsha/gnusrc/gcc/trunk/gcc gtyp-input.list
--- Comment #3 from pinskia at gcc dot gnu dot org 2007-08-03 17:27 ---
*** Bug 32978 has been marked as a duplicate of this bug. ***
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32973
--- Comment #1 from pinskia at gcc dot gnu dot org 2007-08-03 17:27 ---
*** This bug has been marked as a duplicate of 32973 ***
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #1 from tkoenig at gcc dot gnu dot org 2007-08-03 17:29 ---
*sigh*
Oh well, I broke it, so I'd better fix it.
--
tkoenig at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #10 from pinskia at gcc dot gnu dot org 2007-08-03 19:31
---
(In reply to comment #9)
Andrew, do you want to convert to sizetype or to ptrdiff_t? Does it matter?
sizetype is correct for POINTER_PLUS_EXPR. We don't have a ptrdiff_t
internally inside GCC as far as I know,
--- Comment #8 from patchapp at dberlin dot org 2007-08-03 19:40 ---
Subject: Bug number PR target/30315
A patch for this bug has been added to the patch tracker.
The mailing list url for the patch is
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg00204.html
--
Several compilers have an generic intrinsic function ISNAN():
interface
logical function isnan(x)
real, intent(in) :: x
end function isnan
end interface
Example for compilers implementing it: ifort, g95, sunf95, sgi.
Example for a program using it: http://www.libatoms.org/
Note: Fortran
--- Comment #1 from kargl at gcc dot gnu dot org 2007-08-03 19:50 ---
If we're going to implement isnan, then we should implement all of the
FP classification functions (isfinite, isinf, isnan, isnormal, and maybe
fpclassify).
--
kargl at gcc dot gnu dot org changed:
--- Comment #16 from pluto at agmk dot net 2007-08-03 19:55 ---
(In reply to comment #15)
as you can see there is a semicolon after call_do_IRQ(void)
and following asm statement isn't treated as a function body.
in this way -O1 -f{no-}unit-at-a-time accidentally produces
It would be useful if gfortran could implement the following two intrinsics:
a) The Gamma function GAMMA, specific name DGAMMA
b) The logarithm of the Gamma function LGAMMA, specific names ALGAMA, DLGAMA
They are supported by a couple of compilers such as g95, xlf95.
It would be nice if the
--- Comment #2 from burnus at gcc dot gnu dot org 2007-08-03 20:25 ---
If we're going to implement isnan, then we should implement all of the
FP classification functions (isfinite, isinf, isnan, isnormal, and maybe
fpclassify).
I'm not sure if this is needed. For instance, isnan()
--- Comment #4 from burnus at gcc dot gnu dot org 2007-08-03 20:45 ---
I then believe that we do not need isnan. A person can use
either IEEE_* when it becomes available or bind(c). I don't
see why isnan is special in comparison to isinf.
I don't see it either, but other compilers
--- Comment #3 from sgk at troutmask dot apl dot washington dot edu
2007-08-03 20:33 ---
Subject: Re: Implement vendor-specific ISNAN() intrinsic function
On Fri, Aug 03, 2007 at 08:25:08PM -, burnus at gcc dot gnu dot org wrote:
If we're going to implement isnan, then we
--- Comment #2 from fxcoudert at gcc dot gnu dot org 2007-08-03 21:09
---
(In reply to comment #1)
*sigh*
Yeah; I was thinking of organising a non-C99 targets have to die *right*
*now* day, would you join me? ;-)
--
fxcoudert at gcc dot gnu dot org changed:
What
--- Comment #3 from tkoenig at gcc dot gnu dot org 2007-08-03 21:09 ---
Created an attachment (id=14018)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14018action=view)
patch
This should fix it, and also provide at least some safeguards
against buffer overrun in case we ever
--- Comment #5 from fxcoudert at gcc dot gnu dot org 2007-08-03 21:17
---
(In reply to comment #3)
patch
Looks OK to me.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32977
--- Comment #2 from fxcoudert at gcc dot gnu dot org 2007-08-03 21:28
---
Subject: Bug 32955
Author: fxcoudert
Date: Fri Aug 3 21:28:48 2007
New Revision: 127186
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127186
Log:
PR fortran/32955
*
according to www.linux-foundation.org/spec/refspecs/elf/x86_64-SysV-psABI.pdf
the registers should be in the order rax, rbx, rcx, rdx, gcc puts them in rax,
rdx, rcx, rbx.
--
Summary: config/i386/i386.h rbx and rdx dwarf register numbers
are switched.
--- Comment #1 from pinskia at gcc dot gnu dot org 2007-08-03 21:51 ---
You have to becareful when fixing this though. As the unwinder depends on the
order and if you change it you change the ABI.
--
pinskia at gcc dot gnu dot org changed:
What|Removed
--- Comment #9 from fxcoudert at gcc dot gnu dot org 2007-08-03 21:27
---
Well, we'll see if I broke any non-C99 target :)
--
fxcoudert at gcc dot gnu dot org changed:
What|Removed |Added
procedure bad is
a : Long_Float := 2.0;
begin
for i in 1 .. Integer(Long_Float'Truncation(a)) loop
null;
end loop;
end bad;
bad.adb:4:37: unimplemented attribute
gnatmake: bad.adb compilation error
procedure good is
begin
for i in 1 ..
--- Comment #4 from tkoenig at gcc dot gnu dot org 2007-08-03 21:12 ---
(In reply to comment #2)
(In reply to comment #1)
*sigh*
Yeah; I was thinking of organising a non-C99 targets have to die *right*
*now* day, would you join me? ;-)
Of course. You get the dynamite? :-)
--- Comment #8 from fxcoudert at gcc dot gnu dot org 2007-08-03 21:26
---
Subject: Bug 31202
Author: fxcoudert
Date: Fri Aug 3 21:26:10 2007
New Revision: 127185
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127185
Log:
PR fortran/31202
* f95-lang.c
--- Comment #3 from fxcoudert at gcc dot gnu dot org 2007-08-03 21:30
---
Fixed.
--
fxcoudert at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #2 from sje at cup dot hp dot com 2007-08-03 22:05 ---
Created an attachment (id=14019)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14019action=view)
New test case
The first test case only fails on HP-UX in 32 bit mode. This new test case
fails on HP-UX and on Linux
--- Comment #6 from patchapp at dberlin dot org 2007-08-03 22:05 ---
Subject: Bug number PR 32977
A patch for this bug has been added to the patch tracker.
The mailing list url for the patch is
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg00217.html
--
--
fxcoudert at gcc dot gnu dot org changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever Confirmed|0 |1
Last
--- Comment #5 from fxcoudert at gcc dot gnu dot org 2007-08-03 22:07
---
I'd go for implementing isnan as an extension, and only isnan. (until we get
the IEEE module). The following patch does just that:
Index: intrinsic.c
--- Comment #8 from tkoenig at gcc dot gnu dot org 2007-08-03 22:10 ---
Fixed, closing.
--
tkoenig at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #3 from sje at cup dot hp dot com 2007-08-03 22:10 ---
I no longer think this bug is due to how I convert types. I don't think there
is necessarily a bug in the division code but that the increased floating point
register pressure is triggering an existing bug. The failure
--- Comment #2 from jakub at gcc dot gnu dot org 2007-08-03 22:13 ---
You are looking at obsolete version of the psABI.
Please see
http://www.x86-64.org/viewvc/trunk/x86-64-ABI/low-level-sys-info.tex?revision=214view=markup
General Purpose Register RAX 0 \RAX\\
General Purpose
--- Comment #7 from tkoenig at gcc dot gnu dot org 2007-08-03 22:09 ---
Subject: Bug 32977
Author: tkoenig
Date: Fri Aug 3 22:09:10 2007
New Revision: 127187
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127187
Log:
2007-08-03 Thomas Koenig [EMAIL PROTECTED]
PR
--- Comment #4 from rearnsha at gcc dot gnu dot org 2007-08-03 22:35
---
That looks like it solves the problem.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32973
--- Comment #4 from jwakely dot gcc at gmail dot com 2007-08-03 23:30
---
re-confirmed with latest 4.3 snapshot:
tests$ ~/gcc/4.x/bin/g++ -v
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: /home/jwakely/src/gcc-4.3-20070727/configure
--- Comment #9 from pinskia at gcc dot gnu dot org 2007-08-03 23:36 ---
Subject: Bug 32935
Author: pinskia
Date: Fri Aug 3 23:36:05 2007
New Revision: 127190
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127190
Log:
2007-08-03 Andrew Pinski [EMAIL PROTECTED]
PR
--- Comment #10 from pinskia at gcc dot gnu dot org 2007-08-03 23:38
---
Fixed.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #4 from pinskia at gcc dot gnu dot org 2007-08-04 00:06 ---
Subject: Bug 32304
Author: pinskia
Date: Sat Aug 4 00:05:56 2007
New Revision: 127191
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127191
Log:
2007-08-03 Andrew Pinski [EMAIL PROTECTED]
PR
--- Comment #5 from pinskia at gcc dot gnu dot org 2007-08-04 00:06 ---
Fixed.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Status|ASSIGNED
--- Comment #11 from pinskia at gcc dot gnu dot org 2007-08-04 00:33
---
Subject: Bug 32399
Author: pinskia
Date: Sat Aug 4 00:33:31 2007
New Revision: 127196
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127196
Log:
2007-08-03 Andrew Pinski [EMAIL PROTECTED]
PR
--- Comment #12 from pinskia at gcc dot gnu dot org 2007-08-04 00:35
---
Fixed.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #7 from pinskia at gcc dot gnu dot org 2007-08-04 00:43 ---
Note with mismatch checking turned on (which is done by default at stage1), we
get a different ICE (which is not related to the changes done by pointer plus:
t.c: In function 'crashGcc':
t.c:6: error: type mismatch
--- Comment #8 from pinskia at gcc dot gnu dot org 2007-08-04 01:17 ---
Here is a full patch which fixes the type mismatch type ICE also:
Index: fold-const.c
===
--- ../../gcc/fold-const.c (revision 127190)
+++
: *** gcc: No such file or directory. Stop.
--
Summary: gcc-4.3-20070803 bootstrap failure
Product: gcc
Version: 4.3.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: ada
AssignedTo: unassigned at gcc
--- Comment #1 from pinskia at gcc dot gnu dot org 2007-08-04 01:50 ---
*** This bug has been marked as a duplicate of 32941 ***
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #3 from pinskia at gcc dot gnu dot org 2007-08-04 01:50 ---
*** Bug 32983 has been marked as a duplicate of this bug. ***
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||pinskia at gcc dot gnu dot
|
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||pinskia at gcc dot gnu dot
|
--- Comment #9 from pinskia at gcc dot gnu dot org 2007-08-04 05:21 ---
Subject: Bug 32780
Author: pinskia
Date: Sat Aug 4 05:21:30 2007
New Revision: 127199
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=127199
Log:
2007-08-04 Andrew Pinski [EMAIL PROTECTED]
PR
--- Comment #10 from pinskia at gcc dot gnu dot org 2007-08-04 05:22
---
Fixed.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
75 matches
Mail list logo