I would like to try to leverage https://github.com/stillwater-sc/universal to go from IEEE to posit format

Aloha. Did you get any help from unum listserv? "So that I make sure I understand this, does the "float fp16tofp32(fp16 p) " function actually mean "posit16tofp32" (and vice versa for fp32tofp16)? I want to demonstrate an application using the Nvidia/Mellanox DPA in the Bluefield-3 nic [1] and convert ieee floating point to posit in the network path when doing global sums and other collective operations. Any other references for IEEE 64 bit floating point to 32 bit posit would also be appreciated. [1] https://docs.nvidia.com/doca/sdk/dpa+subsystem/index.html

Did anyone respond to that, I have not checked the list