SoCET_Half_Precision_FPU

SystemVerilog implementation of half precision floating point unit

Useful links

• IEEE 754 Calculator
• IEEE 754 Float Toy

Module status

Multiplication:

• Functional for most numbers I've checked by hand (Including infinities, qNans/sNans, and sub-normals).
  • Current testing is plugging numbers in by hand so have only tested ~10 of each type
• Khoi is writing testbench so that we can finalize module.

Addition:

• Possibly done
• Need to confirm weird edge cases (Inf + NaN, etc)

Subtraction: N/A

Rounding: N/A

TODO:

• RISCV Business repository
  • Make a folder in there and put this code there
• Allow the user to choose the rounding mode
• Add inputs which accrues flags
  • Things like divide by zero, or other exceptions