Bugs item #1827627, was opened at 2007-11-07 16:49
Message generated for change (Comment added) made by stmane
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1827627&group_id=56967

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Build
Group: buildtools CVS Head
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Marcin Zukowski (e-r00)
Assigned to: Sjoerd Mullender (sjoerd)
Summary: --enable-bits doesn't work with icc

Initial Comment:
--enable-bits is still in configure and HowToStart.
However, it doesn't seem to work with icc. 
So either some fixes are needed to guarantee it does what a user expects, or 
prints a fail message, or it should be removed.


----------------------------------------------------------------------

>Comment By: Stefan Manegold (stmane)
Date: 2007-11-07 18:22

Message:
Logged In: YES 
user_id=572415
Originator: NO

[on our desktops at CWI]

$ icc --help | egrep '32|64'
           while also generating generic IA-32 code.  <codes> includes
              320 - gcc 3.2.x compatibility
-Wp64              print diagnostics for 64-bit porting
-W[no-]shorten-64-to-32
                   warn for values implicitly converted from a 64-bit to
                   a 32-bit type. Similar to -Wp64
           extended    - rounds intermediates in 64-bit (extended)
precision
-pc32         set internal FPU precision to 24 bit significand
-pc64         set internal FPU precision to 53 bit significand
-pc80         set internal FPU precision to 64 bit significand (DEFAULT)
-auto-ilp32            specify that the application cannot exceed a
32-bit
$ type -a icc
icc is /usr/local/bin/icc
$ ls -l /usr/local/bin/icc
/usr/local/bin/icc -> /soft/64/icc-10.0.023/bin/icc*
$ ls -l /soft/64/icc-10.0.023/bin/icc
/soft/64/icc-10.0.023/bin/icc*
$ ls -l /soft/32/icc-10.0.023/bin/icc
/soft/32/icc-10.0.023/bin/icc*
$ cat > hello.c
#include <stdlib.h>         
#include <stdio.h>

int main () {
printf("Hello\n");
return 0;
}
$ /soft/64/icc-10.0.023/bin/icc -o hello hello.c ; file hello ; ./hello
hello: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically
linked (uses shared libs), for GNU/Linux 2.6.9, not stripped
Hello
$ /soft/32/icc-10.0.023/bin/icc -o hello hello.c ; file hello ; ./hello
hello: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
dynamically linked (uses shared libs), for GNU/Linux 2.6.9, not stripped
Hello

In other words, icc does not seem to offer a commandline swithc to enable
32-bit compilation.
But .../configure CC=/soft/32/icc-10.0.023/bin/icc might do the job...


----------------------------------------------------------------------

Comment By: Marcin Zukowski (e-r00)
Date: 2007-11-07 18:00

Message:
Logged In: YES 
user_id=607094
Originator: YES

Of course:

[17:58:51 [EMAIL PROTECTED]
/export/scratch0/marcin/BUILD/MonetDB-icc-DBGXPROF-32-32/src/gdk/.libs 
$ grep 'CC=icc' ../../../config.log 
  $ /ufs/marcin/monet/MonetDB/configure CC=icc
--prefix=/ufs/marcin/local/monet-icc-DBGXPROF-32-32 --enable-bits=32
--enable-oid32 --disable-profile --without-java --without-bz2
--enable-debug --disable-optimize
$ file libbat_la-gdk_align.o 
libbat_la-gdk_align.o: ELF 64-bit LSB relocatable, x86-64, version 1
(GNU/Linux), not stripped


----------------------------------------------------------------------

Comment By: Fabian (mr-meltdown)
Date: 2007-11-07 17:44

Message:
Logged In: YES 
user_id=963970
Originator: NO

does file(1) actually agree that it is a 64-bits binary?

----------------------------------------------------------------------

Comment By: Stefan Manegold (stmane)
Date: 2007-11-07 17:27

Message:
Logged In: YES 
user_id=572415
Originator: NO

No, I didn't know that --- actually I did not even know what platform you
were using (I could have guessed but we do have more than one at CWI ...).

Given the CPU you specify, I now *guess* you're referring to our FC6
desktops at CWI and to icc 10.0, right?

Well, I am actually not sure whether (1) "our" icc 10.0 allows to produce
32-bit binaries on our 64-bit FC6 desktops, (2) the correct option to do
this is set (added to $CC) in buildtools/conf/MonetDB.m4, (3) our 64-bit
FC6 comes with all the 32-bit libraries to actually get a 32-bit MonetDB
compiled successfully, (4) it would work with gcc.

In any case I agree with Marcin that the current behavior of silently
ignoring --enable-bits=32 (with icc on our 64-bit FC6 machines) is not the
desired one.

(Who ever will take care of this one may consider the above as hint where
to start the investigation.)


----------------------------------------------------------------------

Comment By: Marcin Zukowski (e-r00)
Date: 2007-11-07 17:12

Message:
Logged In: YES 
user_id=607094
Originator: YES

I'm sure you know, but if you insist :)

On Athlon 64 X2, "--enable-bits=32" with icc it compiles into 64-bit
binaries.

I don't even know if icc supports that, but I was simply suprised to get a
64-bit thing.


----------------------------------------------------------------------

Comment By: Stefan Manegold (stmane)
Date: 2007-11-07 17:01

Message:
Logged In: YES 
user_id=572415
Originator: NO

"does not seem to work" as in (1) "does not have any effect" or (2)
"results in an error" or (3) "..."?

It --- to be precise one of "--enable-bits=32", "i--enable-bits=64" --- is
(IMHO) still supposed to work ...


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1827627&group_id=56967

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs

Reply via email to