Phil Carns

make failure if Check not installed

1

If you compile libevfibers without first installing the "check" unit testing framework, then the build will fail as shown below. The package is easy to install to resolve the problem,...

**Is your feature request related to a problem? Please describe.** There are some code paths in Mercury that will allocate extra memory if needed (for example, to satisfy a request...

HG: safe mechanism to deregister an RPC while handles for that RPC are still in use

3

**Is your feature request related to a problem? Please describe.** Imagine a hypothetical scenario in which a service is periodically receiving a particular RPC type. The service then begins to...

NA OFI: inadvertent activation of psm3 instead of psm2

1

**Describe the bug** After upgrading libfabric (via external repo at https://github.com/mochi-hpc/mochi-spack-packages/tree/main/packages/libfabric) from 1.11.1 to 1.13.0, psm2 performance on the LCRC Bebop system dropped by an order of magnitude. Performance with...

NA OFI: support vector rma operations

22

**Is your feature request related to a problem? Please describe.** The na_ofi.c already uses libfabric API calls that support vectors for local and remote buffers (fi_writemsg() and fi_readmsg()), but it...

NA OFI: hg_test* hangs using verbs;ofi_rxm with concurrent handles

8

Examples include hg_test_rpc, hg_test_rpc_lat, and hg_test_write_bw, all of which work correctly until the phase of the test that uses multiple handles or concurrent RPCs. This is with libfabric 1.6 and...

If the client and server disagree on the maximum unexpected message size (for example, if each is choosing them at runtime based on interface settings) then it may cause communication...

If one ES consistently launches ULTs to execute on a different slave ES (for example, to offload specific calculations or I/O operations), then the per-ES stack queue on the latter...

The following test program can be used to illustrate the problem: https://gist.github.com/carns/57b1e124460f3b2371d5219c747d0f10 The test program should be executed with 4 processes. Each process writes a different (but adjacent) 100 byte...

We have observed some highly unusual access patterns being generated by collective write operations in ompio. Many thanks to @wkliao for helping to narrow down how to reproduce the behavior....