[MITgcm-devel] More on pgf77

[Chris Hill]

Please feel free to remove it.

Chris 

> -----Original Message-----
> From: mitgcm-devel-bounces at mitgcm.org 
> [mailto:mitgcm-devel-bounces at mitgcm.org] On Behalf Of Patrick Heimbach
> Sent: Friday, September 24, 2004 5:24 PM
> To: MITgcm-devel at mitgcm.org; Ed Hill
> Cc: MITgcm-devel
> Subject: [MITgcm-devel] More on pgf77
> 
> 
> Hi again,
> 
> reason why I tested pgf77 on faulks is that compilation 
> crashed on a different linux box at Harvard (and there's a 
> fix, see end of mail)
> 
> > uname
> Linux mashav 2.4.20-8 under RH9
> > pgf77 -V
> pgf77 5.2-1
> Copyright 1989-2000, The Portland Group, Inc.
> Copyright 2000-2004, STMicroelectronics, Inc.
> 
> with the following (weird?) error:
> > make
> pgf77 -byteswapio -r8 -Mnodclchk -Mextend -tp p6 -v -O2 
> -Munroll -Mvect=cachesize:512000,transform -c mom_vi_u_coriolis_c4.f
>  
> /usr/local/pgi/linux86/5.2/bin/p3/pgftn 
> mom_vi_u_coriolis_c4.f -opt 2 -x 119 0xa10000 -x 119 0x100000 
> -x 122 0x40 -x 123 0x1000 -x 127 4 -x 119 0x40000000 -vect 48 
> -y 34 16 -x 34 0x8 -x 32 512000 -x 16 12 -y 19 8 -y 35 0 -x 39 0x21 -x
> 51 0x20 -x 124 0x1401 -astype 0 -stdinc
> /usr/local/pgi/linux86/5.2/include:/usr/local/include:/usr/lib
> /gcc-lib/i386-redhat-linux/3.2.2/include:/usr/lib/gcc-lib/i386
-redhat-linux/3.2.2/include:/usr/include
> -def unix -def __unix -def __unix__ -def linux -def __linux 
> -def __linux__ -def __inline__= -def i386 -def __i386 -def 
> __i386__ -def __NO_MATH_INLINES -def
> linux86 -def __THROW= -cmdline '+pgf77 mom_vi_u_coriolis_c4.f 
> -byteswapio -r8 -Mnodclchk -Mextend -tp p6 -v -O2 -Munroll 
> -Mvect=cachesize:512000,transform -c' -x 125 2 -nodclchk 
> -extend -x 124 1 -x 124 0x28 -x 80 0x300 -y 80 0x1000 -x
> 9 1 -y 10 0 -asm /tmp/pgf77aaaaayCpai.s
> PGFTN-F-0000-Internal compiler error. dirv_permute: map is 
> wrong       2
> (mom_vi_u_coriolis_c4.f: 2903)
> pgf77-Fatal-fcomp completed with exit code 1
>  
> Unlinking /tmp/pgf77aaaaayCpai.s
> make: *** [mom_vi_u_coriolis_c4.o] Error 2
> 
> 
> 
> If I remove the FOPTIM flags  
> -Mvect=cachesize:512000,transform it compiles succesfully.
> So we should maybe comment that bit out in the corresponding optfile?
> 
> -Patrick
> 
> --------------------------------------------------------
> Patrick Heimbach   Massachusetts Institute of Technology
> FON: +1/617/253-5259                  EAPS, Room 54-1518
> FAX: +1/617/253-4464             77 Massachusetts Avenue
> mailto:heimbach at mit.edu               Cambridge MA 02139
> http://www.mit.edu/~heimbach/                        USA
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-devel
>