[gmx-users] sgi -r12000 flag

[Erik Lindahl]

Arvid Soederhaell wrote:

Hi Arvid,

Try setting CFLAGS="-mips4 -O3 -r10000 -woff 1174" (and FFLAGS) before 
running the configure script, and then recompile.

Cheers,

Erik

>Hi gromacs users
>
>I'm trying to install gromacs 3.1.4 on an SGI origin having R12000 CPU's.
>The configure script detects the cpu type but I get the following message: 
>
>-----------------------------------------------------------------------
>MIPSpro Compilers: Version 7.2.1 checking whether cc accepts -r12000 
>-mips4 -O3 -OPT:IEEE_arithmetic=3 -OPT:rsqrt=ON -SWP:loop_overhead
>-INLINE:=ON -LNO:opt=1 -LNO:ou_further=3
>-OPT:Olimit=0:roundoff=3:alias=typed -woff 1174 -D__INLINE_INTRINSICS... 
>no 
>******************************************************************* *
>Sorry, these optimization settings don't seem to work for * * your C
>compiler. Use make CFLAGS=..., or edit the top Makefile. *
>*******************************************************************
>checking whether f77 accepts -r12000 -mips4 -O3 -OPT:IEEE_arithmetic=3
>-OPT:rsqrt=ON -SWP:loop_overhead -INLINE:=ON -LNO:opt=1 -LNO:ou_further=3
>-OPT:Olimit=0:roundoff=3:alias=typed -OPT:cray_ivdep=TRUE... f77 ERROR
>parsing -r12000:  unknown flag yes
>-------------------------------------------------------------------------
>
>The run ends with the message:
>
>-------------------------------------------------------------------------
>configure: error: No math library found
>-------------------------------------------------------------------------
>
>Since it says "ERROR parsing -r12000: unknown flag" i have edited the
>configure script and replaced the -r12000 by -r10000. This works
>relatively fine (quite afew warnings though) until the compilation reaches
>the "protonate.c" subroutine where the error messages are 
>
>------------------------------------------------------------------------
>protonate.c", line 31: error(1007): unrecognized token
>   */
>     ^
>
>
>
>"protonate.c", line 31: error(1166): expected a declaration
>   */
>     ^
>
>"protonate.c", line 32: error(1007): unrecognized token
>  static char *SRCID_protonate_c = "$Id: protonate.c,v 1.5.2.1 2002/04/01
>20:20:42 lindahl Exp $"; 
>
>^
>
>"protonate.c", line 32: error(1166): expected a declaration
>  static char *SRCID_protonate_c = "$Id: protonate.c,v 1.5.2.1 2002/04/01
>20:20:42 lindahl Exp $"; 
>
>^
>
>"protonate.c", line 33: warning(1014): extra text after expected end of
>          preprocessing directive
>  #include <math.h>
>--------------------------------------------------------------------------
>
>and so on... 
>
>I have seen that there have been problems with the -n32 flag for SGI's,
>but this does not seem to be related to that, or? Is it wrong or does it
>make the code substantially slower when using the -r10000 flag? Any ideas
>how to solve the problem?
>
>I have seen that there is a hack for sgi's in the configure script
>suggesting --disable-waterwater-loops and --disable-waterloop-unrolling. I
>dont really understand what that means, but it doesn't make a difference
>if i include them upun installing.
>
>(I believe that the FFTW is properly installed in single precision.)
> 
>Thanks in advance
>
>   Arvid Soederhaell
>
>Arvid Soederhaell, PhD                          Phone:+49-30-94793-231 
>Forschungsinstitut fuerMolekulare Pharmakologie Mob:+49-160-9835 8282
>Robert-Roessle-Strasse 10                       Fax::+49-30-94793-169   
>131 25 Berlin                                   (Mob:+46-70-723 33 63) 
>
>_______________________________________________
>gmx-users mailing list
>gmx-users at gromacs.org
>http://www.gromacs.org/mailman/listinfo/gmx-users
>Please don't post (un)subscribe requests to the list. Use the 
>www interface or send it to gmx-users-request at gromacs.org.
>  
>