Lucian Petrica

To reproduce, run tests against an emulator session, with 8 ranks Some outputs slightly different from expected: ``` [1,7]:1th item is incorrect! (5035.578613 != 5035.579102) [1,7]:2th item is incorrect! (-5832.367676...

Assuming #3 is fixed and we can receive from multiple nodes simultaneously, add driver/firmware support for configuring and operating multiple communicators simultaneously.

Interface with non-FPGA hosts

1

To have FPGAs talk to non-FPGA hosts we need a software implementation of the ACCL collectives protocol on top of TCP initially, then on top of RDMA.

Use PLRAM as exchange memory

2

Currently we have a small exchange memory mapped both in the Microblaze and host address spaces to hold configuration data. The size of this memory is limiting, e.g. some users...

Add support for hierarchical collectives within the confines of fanin == 1. Some examples: - hierarchical rings for (all)gather, (all)reduce, scatter-reduce - hierarchical trees for broadcast and scatter

The [request/completion structures](https://github.com/fpgasystems/Coyote/blob/master/scripts/wr_hdl/template_gen/lynx_pkg_tmplt.txt#L424C1-L433C12) contain several stale fields which should be cleaned up. Also mode should be merged into opcode as it doesn't currently provide any additional information. Specifically, the following...