[petsc-users] configure error on Titan with Intel

[Barry Smith]

> On May 11, 2015, at 6:55 PM, Mark Adams <mfadams at lbl.gov> wrote:
> 
> More Titan fun.  I have built for PGI on Titan but tests are failing to build: -lhwloc?

  Mark,

  What test is this? Did "make tests" run correctly?  How did you run this thing below, remember if you cut and paste everything including what you typed and the directory that your are in that helps. Did you happen to do this make test on a different login node or with a different configuration? 

> 
> cc -mp -fast  -o ex56 ex56.o -L/autofs/na3_home1/adams/petsc_master_opt64idx_pgi/lib  -lpetsc -Wl,-rpath,/opt/pgi/14.2.0/linux86/5.1/lib -L/opt/pgi/14.2.0/linux86/5.1/lib -llapack -lblas -Wl,-rpath,/autofs/na3_home1/adams/petsc_master_opt64idx_pgi/lib -L/autofs/na3_home1/adams/petsc_master_opt64idx_pgi/lib -lparmetis -lmetis -lhwloc -lssl -lcrypto -ldl -lstdc++
> /usr/bin/ld: cannot find -lhwloc
> make: [ex56] Error 2 (ignored)

  Notes: ./configure ran fine and detected -lhwloc in some standard system install location under normal circumstances it couldn't just disappear for a different example.


> /
> 
> On Fri, May 8, 2015 at 5:51 PM, Jed Brown <jed at jedbrown.org> wrote:
> Satish Balay <balay at mcs.anl.gov> writes:
> 
> > Also - Perhaps with intel compilers - the recommended blas is
> > something other than ACML? [like MKL?]. Something to check..
> 
> MKL (and anything built using the Intel compiler) has the "run slower on
> AMD" feature.  For any given operation, ACML won't necessarily be
> faster, but at least it's not intentionally crippled.  For most of what
> PETSc does, any such MKL/ACML difference is irrelevant.
> 
> <configure.log><make.log>