Create a bunch of example CFUs operations

[mithro]

Here is a list of example CFU operations the we thought would be interesting;

• [x] Count of number of 1s
• [ ] Integer Multiple
  • [ ] 32bit×32bit == 32bit
  • [ ] 32bit×32bit == 64bit
• [ ] Integer Divide
  • Rounding modes?
• [ ] Multiply + Accumulate - rs1×rs2+ra → rd, ra -- ra is internal state?
• [ ] Divide + Remainder -- two output / result values.
• [ ] Alpha blending of 32bit argb colors
• [ ] SWAR / MMX style operations
  • [ ] two 32-bit integers,
  • [ ] four 16-bit integers, or
  • [ ] eight 8-bit integers at once
• [ ] Saturation arithmetic

We should create implementations of each of the CFU operations and check they work with our CFU interface.

[grayresearch]

Added Popcount Added MulAcc (acc += 32b x 32b) which is different from the above

32b x 32b -> 64b might be done with CFU_REQ_DATA_W=32 where CFU_RESP_DATA_W=64, or might be done withCFU_RESP_DATA_W=32 and CFU_RESP_OUTPUTS = 2.

This begs the question of how to handle CFU_RESP_OUTPUTS>1. Does the CPU accept the tuple of two outputs in one cycle, or does it receive them sequentially one at a time?