mroussel@alchemy.chem.utoronto.ca (Marc Roussel) (04/02/91)
I can't seem to get gprof to work with f77 programs. Here are the relevant parts of my makefile: FFLAGS = -g -pg -C $$FLOAT LDFLAGS = -Y L,$$LIBDIR -lode -lpg -lncarg -lncarg_gks -lncarg_loc -lutchem siminit.o: siminit.f declare.h common.h f77 $(FFLAGS) -c siminit.f create_consts.o: create_consts.f declare.h common.h f77 $(FFLAGS) -c create_consts.f sim4: sim4.o siminit.o create_consts.o f77 -g -pg sim4.o siminit.o create_consts.o -o sim4 $(LDFLAGS) The program compiles and executes without complaint (and as far as I can tell, it produces correct results) and leaves a gmon.out file behind. Unfortunately when I try to run gprof on the result I get the following error message: [nllookup] binary search fails??? Segmentation fault Does anyone have any idea what's causing this? Is it my fault or is it just these !@#@!% Apollo's again? We are running SR10.3.p with the BSD Unix stuff. The compiler is rev. 10.8. Marc R. Roussel mroussel@alchemy.chem.utoronto.ca
rtp1@quads.uchicago.edu (raymond thomas pierrehumbert) (04/02/91)
I also have never gotten gprof to work. I'm running 10.1p on a DN10k, rev 10.7p of the compilers (I tried to order rev 10.8p, but first the sales rep didn't know about availability, and then lost my P.O. Just an aside). Things compile with -pg, and gmon.out gets left behind, but when I run gprof, the results are vacuous; none of my own procedures are included in the call graph, only some cryptic system-generated ones). .