public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
* c++/7542: compiling for powerpc Altivec
@ 2002-08-08  5:16 eyal
  0 siblings, 0 replies; 2+ messages in thread
From: eyal @ 2002-08-08  5:16 UTC (permalink / raw)
  To: gcc-gnats


>Number:         7542
>Category:       c++
>Synopsis:       compiling for powerpc Altivec
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Thu Aug 08 05:06:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     eyal@mobileye.com
>Release:        gcc-3.1
>Organization:
>Environment:
linux , powerpc 
>Description:
I tried to compile Altivec code - and used vec_msum(__vector signed char c1, __vector unsigned char c2, __vector signed int i1) ... 

the compiler change this to the assembler code vmsumubm - which uses 2 UNsigned char vectors instead of vmsummbm which uses one SIGNES and one UNSIGND !!

this bug causes the signed operand to be treated as unsigned !!
>How-To-Repeat:
the following compilation:
g++ -S -Wall -pedantic -ansi -Wno-long-long -DNDEBUG -O6 -DALTIVEC -maltivec testZevel.cpp

the g++ version is :
Reading specs from /opt/gcc-3.1/lib/gcc-lib/i686-pc-linux-gnu/3.1/specs
Configured with: ../gcc-3.1/configure --prefix=/opt/gcc-3.1 --enable-shared --enable-threads --enable-languages=c,c++
Thread model: posix
gcc version 3.1
>Fix:
need to be find fast so if you have something please let me know !!!
>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: application/octet-stream; name="testZevel.cpp"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="testZevel.cpp"

I2luY2x1ZGUgPGFsdGl2ZWMuaD4KCl9fdmVjdG9yIHNpZ25lZCBpbnQgdGVzdFpldmVsKCkKeyAg
ICAKICBfX3ZlY3RvciBzaWduZWQgY2hhciB2c2MxID0gey0xNiwtMTUsLTE0LC0xMywtMTIsLTEx
LC0xMCwtOSwtOCwtNywtNiwtNSwtNCwtMywtMiwtMX07CiAgX192ZWN0b3IgdW5zaWduZWQgY2hh
ciB2dWMxID0gezEsMiwzLDQsNSw2LDcsOCw5LDEwLDExLDEyLDEzLDE0LDE1LDE2fTsKICAKICAg
IAogIF9fdmVjdG9yIHNpZ25lZCBpbnQgdnNpMSA9IHswLDAsMCwwfTsKICBfX3ZlY3RvciBzaWdu
ZWQgaW50IHZzaTI7CiAgCiAgdnNpMiA9IHZlY19tc3VtKHZzYzEsdnVjMSx2c2kxKTsKCiAgcmV0
dXJuIHZzaTI7CiAgICAKfQogIAo=


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: c++/7542: compiling for powerpc Altivec
@ 2002-09-11 12:03 dje
  0 siblings, 0 replies; 2+ messages in thread
From: dje @ 2002-09-11 12:03 UTC (permalink / raw)
  To: dje, eyal, gcc-bugs, gcc-prs, nobody

Synopsis: compiling for powerpc Altivec

Responsible-Changed-From-To: unassigned->dje
Responsible-Changed-By: dje
Responsible-Changed-When: Wed Sep 11 12:03:45 2002
Responsible-Changed-Why:
    PowerPC
State-Changed-From-To: open->closed
State-Changed-By: dje
State-Changed-When: Wed Sep 11 12:03:45 2002
State-Changed-Why:
    Fixed in GCC 3.2

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=7542


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2002-09-11 19:03 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-08-08  5:16 c++/7542: compiling for powerpc Altivec eyal
2002-09-11 12:03 dje

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).