On 20/05/2020 10:07, Gard Spreemann wrote: > Drew Parsons <[email protected]> writes: > >> If I understand correctly, it's dangerous to simply enable 64-bit in >> PETSc alone. It needs to be done all along the computational library >> stack.
In the case of openmpi, I've just done an experimental 64-bit build, 4.0.3-7 and uploaded to test if it builds correctly on all architectures. I'd be in favour of 64-bit computational stacks; some of the packages (pnetcdf, etc) already don't work with 32-bit; It should make is _possible_ to run on 32-bit, but accept that performance on 64-bit counts for our users. Taking a perf hit on 32-bit archs to make 64-bit archs faster is reasonable; the question becomes - can 32-bit archs handle 64-bit code (int64 etc), or are we dropping support for 32-bit archs ? > > -- Alastair McKinstry, email: [email protected], matrix: @alastair:sceal.ie, phone: 087-6847928 Green Party Councillor, Galway County Council
