Hi John,
The OpenMP Tools Workgroup has been discussing a very similar issue for OpenMP threads and TARGET regions (e.g.,regions offloaded to GPUs). Tony Tye at AMD (who is also a DWARF committee member) proposed a model where the compiler can generate DWARF that
would tell a consumer (most lik
On 12/29/20 2:14 PM, Bishop, John E via Dwarf-Discuss wrote:
Fortran defines the TEAM_TYPE as "processor dependent", which means
implementation is up to the compiler/run-time implementation. It could be a pointer to a
structure; it could be a key to a database, it could be a small integer whic
> So what about these structures makes them difficult to describe in DWARF?
> Let each producer define structures that work for them.
Fortran defines the TEAM_TYPE as "processor dependent", which means
implementation is up to the compiler/run-time implementation. It could be a
pointer to a str
On 12/28/20 11:52 AM, Bishop, John E via Dwarf-Discuss wrote:
Different implementations of these features may use very different
methods (e.g. MPI, shmem, real shared memory, home-brew message passing,
RPC,..). This means that the structures representing coarray concepts
like coarrays, teams, ev
I am no longer a committee member, but I am still involved with helping make
Fortran debuggable. I ask for comments on the possible extension described
below.
Fortran now has an extensive set of features to support parallel program using
a shared-memory paradigm. This was introduced in the 2
