[cig-commits] r20399 - seismo/3D/SPECFEM3D/trunk

[dkomati1 at geodynamics.org]

Author: dkomati1
Date: 2012-06-22 13:32:34 -0700 (Fri, 22 Jun 2012)
New Revision: 20399

Modified:
   seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt
Log:
added two small problems that should be fixed in the "configure" scripts


Modified: seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt
===================================================================
--- seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt	2012-06-22 07:23:45 UTC (rev 20398)
+++ seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt	2012-06-22 20:32:34 UTC (rev 20399)
@@ -458,16 +458,33 @@
 - suggestion 28:
 ----------------
 
-Regarding the implicit time schemes that Zhinan has implemented in 2D, I agree that it would be great to put that in the official SVN version relatively soon to avoid losing the changes if we wait for too long. But I think we only need this in 2D for now, so let us not do it in 3D (at least for now in 2012). Let us just commit Zhinan's 2D version of the implicit routines to the official SVN code (making it off by default; the default should remain a purely explicit second-order Newmark scheme). 
+two small problems in the "configure" script (of SPECFEM2D, SPECFEM3D and also SPECFEM3D_GLOBE):
 
+* CFLAGS for the C compiler is hardwired in the different Makefiles (and set to -O2 or something simple like that); it is not defined in "flags.guess"; this is problematic when running on more than 2 GB of memory per core, which is not uncommon these days; in this case many compilers require users to add "-mcmodel=medium" to the Fortran and to the C compiler flags. In the current flags.guess this can be done for Fortran but not for C and thus we get an error message from the linker (with gfortran / gcc, with Intel ifort / icc, probably others as well); the solution would be to move CFLAGS from the Makefiles to flags.guess (only)
+
+* when we use Intel icc (for instance with "./configure FC=ifort CC=icc") the script seems to guess that GNU gcc is used instead:
+
+checking for gcc... icc
+checking whether we are using the GNU C compiler... yes
+checking for gcc... (cached) icc
+
+Obviously the script should print "checking for CC" instead, and "checking whether we are using the GNU C compiler" should return "no".
+
+Another option (at least in the meantime) would be to tell users in the manual to use this: "./configure FC=ifort CC=icc CFLAGS="-O2 -mcmodel=medium -shared-intel".
+
 - suggestion 30:
 ----------------
 
+Regarding the implicit time schemes that Zhinan has implemented in 2D, I agree that it would be great to put that in the official SVN version relatively soon to avoid losing the changes if we wait for too long. But I think we only need this in 2D for now, so let us not do it in 3D (at least for now in 2012). Let us just commit Zhinan's 2D version of the implicit routines to the official SVN code (making it off by default; the default should remain a purely explicit second-order Newmark scheme). 
+
+- suggestion 31:
+----------------
+
 Elliott Sales de Andrade and Qinya will implement a routine to interpolate and output the sensitiviy kernels on a topologically-regular grid
 instead of the non-structured SEM grid (this can be useful for many processing packages that require a topologically-regular grid
 of points as input).
 
-- suggestion 31:
+- suggestion 32:
 ----------------
 
 Convert the mesh database of the 2D code from ASCII to binary to produce smaller files?