Porting from SLES9/EM64T to SLES9/AMD64
Hi, After porting a C++ application from SLES9/EM64T architecture to SLES9/AMD64, it fails to work. Apparently, looks like a stack corruption is happening. It works perfect on SLES9/EM64T arch. Any idea what might have gone wrong? Env is same on both boxes and gcc used was 3.3.2 to build the application. Any input is highly appreciated. Thanks Rajib
On Sunday 03 September 2006 15:32, Rajib Majumder wrote:
Hi,
After porting a C++ application from SLES9/EM64T architecture to SLES9/AMD64, it fails to work. Apparently, looks like a stack corruption is happening. It works perfect on SLES9/EM64T arch.
Any idea what might have gone wrong?
Env is same on both boxes and gcc used was 3.3.2 to build the application.
Any input is highly appreciated.
It would help if you posted a code snippet that exhibited the problem. Except for extensions like SSE3 and such like, EM64T and AMD64 should be compatible with each other, so unless you're using highly platform specific optimisations, you shouldn't need to do any particular porting My guess is that you have a bug in your program, that invokes undefined behaviour, and that this happens to behave "correctly" on EM64T but causes a crash on AMD64 Then again, it could be a bug in gcc, but without a code example of the issue it's hard to say
On Sunday 03 September 2006 15:32, Rajib Majumder wrote:
Hi,
After porting a C++ application from SLES9/EM64T architecture to SLES9/AMD64, it fails to work. Apparently, looks like a stack corruption is happening. It works perfect on SLES9/EM64T arch.
Any idea what might have gone wrong?
Env is same on both boxes and gcc used was 3.3.2 to build the application.
Any input is highly appreciated.
It would help if you posted a code snippet that exhibited the problem. Except for extensions like SSE3 and such like, EM64T and AMD64 should be compatible with each other, so unless you're using highly platform specific optimisations, you shouldn't need to do any particular porting
My guess is that you have a bug in your program, that invokes undefined behaviour, and that this happens to behave "correctly" on EM64T but causes a crash on AMD64
Then again, it could be a bug in gcc, but without a code example of the issue it's hard to say I agree 100%. It would be good if you also supplied the compiler
On Sun, 3 Sep 2006 15:46:21 +0200 Anders Johansson <andjoh@rydsbo.net> wrote: options that you used. -- Jerry Feldman <gaf@blu.org> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9
participants (3)
-
Anders Johansson -
Jerry Feldman -
Rajib Majumder