Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Wed, 2018-05-02 at 08:28 -0500, Satish Balay wrote: > On Tue, 1 May 2018, Drew Parsons wrote: > > > Looks like that does the trick. Running make in src/dm/f90-mod, it > > successfully compiles and archives petscdmmod.o and > > petscdmplexmod.o > > inside build-arch/lib/libpetsc.a > > Great!

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Tue, 1 May 2018, Drew Parsons wrote: > Looks like that does the trick. Running make in src/dm/f90-mod, it > successfully compiles and archives petscdmmod.o and petscdmplexmod.o > inside build-arch/lib/libpetsc.a Great! I've merged it to maint - so this change should be in petsc-3.9.2 BTW:

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Tue, 2018-05-01 at 17:41 +0800, Drew Parsons wrote: > On Tue, 2018-05-01 at 01:02 -0500, Satish Balay wrote: > > On Tue, 1 May 2018, Drew Parsons wrote: > > > > > Would it work to split petscdmmod.F into two pieces > ... > > Can you try: > > > >

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Tue, 2018-05-01 at 01:02 -0500, Satish Balay wrote: > On Tue, 1 May 2018, Drew Parsons wrote: > > > Would it work to split petscdmmod.F into two pieces, e.g. one with > > petscdmda and central petscdm and the other with petscdmplex? > > ... > Can you try: > >

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Tue, 1 May 2018, Drew Parsons wrote: > Would it work to split petscdmmod.F into two pieces, e.g. one with > petscdmda and central petscdm and the other with petscdmplex? > > Naively the two separate object files would then need around 1GB, > allowing mpif90 -c to succeed. This assumes that

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Sat, 2018-04-28 at 18:11 +0300, Adrian Bunk wrote: > > The failing command is: > mpif90 -c -g -O2 -fstack-protector-strong -fPIC -Wdate-time > -D_FORTIFY_SOURCE=2 -fdebug-prefix-map=/home/bunk/build/petsc- > 3.9.0+dfsg1=. -I/home/bunk/build/petsc-3.9.0+dfsg1/include >

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Sat, Apr 28, 2018 at 08:29:48AM -0500, Satish Balay wrote: > > Also we print a summary of compile flags up in the log: > > Using Fortran compile: mpif90 -c -g -O2 -fstack-protector-strong -fPIC > -Wdate-time -D_FORTIFY_SOURCE=2 > -fdebug-prefix-map=/<>/petsc-3.9.0+dfsg1=. >

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Sat, 28 Apr 2018, Matthew Knepley wrote: > On Sat, Apr 28, 2018 at 5:58 AM, Adrian Bunk wrote: > > mips and mipsel are the only architectures where we have only 2 GB > > address space for userspace, so when the compiler needs more than > > 2 GB of memory you are getting this

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

On Sat, Apr 28, 2018 at 5:58 AM, Adrian Bunk wrote: > On Fri, Apr 27, 2018 at 09:48:17AM +0800, Drew Parsons wrote: > > Source: petsc > > Version: 3.9.0+dfsg1-1exp4 > > Severity: normal > > > > PETSc 3.9 Fails To Build From Source on mips. PETSC 3.8 continues to > build fine. >

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

Thanks Satish, good to know it's not just us. Drew On Thu, 2018-04-26 at 21:14 -0500, Satish Balay wrote: > We've received a couple of similar reports on regular [amd64] builds > on VMs. And increasing memory allocated to the VM alleviated the > problem > > We don't know why gfortran is

Bug#897006: [petsc-maint] Bug#897006: petsc FTBFS on mips: f951: out of memory

We've received a couple of similar reports on regular [amd64] builds on VMs. And increasing memory allocated to the VM alleviated the problem We don't know why gfortran is requiring so much memory to compile these module files. Satish On Fri, 27 Apr 2018, Drew Parsons wrote: > Source: petsc >