I guess you can just create links [from mpicc to ccache] instead of these wrapper scripts in these locations - with the same ccache performance?
And with the scripts [mpicc/mpicxx links to ccache] in PATH - you don't need --with-mpi-dir [and mpiexec/include/lib work-around] Also is there an advantage to have mpicc linked to ccache instead of gcc? [I guess having both linked to ccache will result in duplicate ccache searches for the same command?] Satish On Fri, 17 Jan 2020, Jed Brown wrote: > To use system MPI, I made a directory with one-line scripts > > $ rg . ~/usr/ccache/mpich/ > /home/jed/usr/ccache/mpich/bin/mpiexec > 1:#!/bin/sh > 3:/opt/mpich/bin/mpiexec "$@" > > /home/jed/usr/ccache/mpich/bin/mpicc > 1:#!/bin/dash > 3:ccache /opt/mpich/bin/mpicc "$@" > > /home/jed/usr/ccache/mpich/bin/mpicxx > 1:#!/bin/dash > 3:ccache /opt/mpich/bin/mpicxx "$@" > > > ln -s /opt/mpich/include ~/usr/ccache/mpich/include > mkdir ~/usr/ccache/mpich/lib # Configure wants (or wanted) this > directory to exist, but it can be empty > > > Then --with-mpi-dir=$HOME/usr/ccache/mpich is all you need. Works with > all packages and I haven't touched it in nearly a decade. > > Matthew Knepley <[email protected]> writes: > > > I configured MPI with CC="ccache gcc" > > > > knepley/feature-dm-remove-hybrid *$:/PETSc3/petsc/petsc-dev$ > > /PETSc3/petsc/bin/mpicc -show > > /Users/knepley/MacSoftware/bin/ccache gcc -Qunused-arguments > > -fstack-protector -Qunused-arguments -g3 -Wl,-flat_namespace > > -I/PETSc3/petsc/include -L/PETSc3/petsc/lib -lpmpich -lmpic > > h -lopa -lmpl -lpthread > > > > and then used --with-mpi-dir. This is the way to go if you do multiple > > ARCHes. Have not had a problem on OSX or Linux. Works with all packages. > > > > Matt > > > > On Fri, Jan 17, 2020 at 11:29 AM Patrick Sanan <[email protected]> > > wrote: > > > >> I'm shamefully not using ccache. How do I do it? Is it as simple as > >> ./configure --with-cc="ccache gcc" --with-cxx="ccache g++"? Works on OS X > >> and various Linuxes? Any known issue with external packages or otherwise? > > > > > > > > -- > > What most experimenters take for granted before they begin their > > experiments is infinitely more interesting than any results to which their > > experiments lead. > > -- Norbert Wiener > > > > https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/> >
