Brian Gunney
Brian Gunney
This requests a small change in `conduit::relay::mpi::recv_using_schema` to make it work properly when the source is `MPI_ANY_SOURCE`. This currently doesn't work because the `MPI_Recv` doesn't necessarily catch the message found...
I'd like to have more flexible ways to complete `conduit::relay::mpi::Request`s from non-blocking communications. Specifically, 1. support the way applications store Requests and 2. provide a choice to use MPI_Waitsome and...
# Simplify communication loop in DistributedClosestPoint class - This PR is a refactoring - It does the following (modify list as needed): - Removes the "skip" messages in the communication...
We can simplify the logic that allows ranks to skip searches for query points from query partitions that are far away. We are currently sending short messages to let the...
DistributedClosestPoint should be extended to support multi-domain (domain overloading) for both the query mesh and the object mesh. Multi-domain specification is defined by the Conduit blueprint, which should be used...
`DistributedClosestDistance` has been used in 3D, but Axom only tests it in 2D. These steps constitute a simple extension for testing the search in 3D. 1. Extend the circle to...
# Support domain over/underloading in `DistributedClosestPoints` - This PR is a feature - It does the following (modify list as needed): - Modifes `DistributedClosestPoints` to support multidomain (or singledomain) blueprint...
The crash occurs when the code tries to access element 1 of a 1-element array. The problem is likely in special case of a `BVH` for one point. The code...
`UnstructuredMesh` crashes while allocating memory. Easiest way to reproduce is to reserve space. Without explicitly reserving space, it will still crash when when implicitly reserving. Crashes only in the docker...
`DistributedClosestPoint` is limited in what the exact form of incoming data. Ideally, it should be able to handle any valid blueprint mesh. At a minimum, it should detect valid data...