Separate xbar and replace en/rdy with val/rdy

[tancheng]

This PR:

• Closes [P0] Proceed ctrl signal only both xbar and FU have done their job together #53
• Closes [P1] Systolic array simulation #22
• Closes [P0] Xbar proceeding should also depend on whether the opt can be consumed #10

---

Contributes to bugs:

• Separates xbar: [P0] Synthesis-friendly tile design #13
  • Need synthesis @yyan7223 to verify the improvement and debug.
• Rreplacing en/rdy with val/rdy: [P2] Cleanup and update #39
• [P1] Data arrival-dependent execution and cycle-based execution #26

---

Related bugs:

• [P0] Reduce all FUs to one out #51
• [P1] Predicate should be associated with specific control signal #27
• [P3] Allow parameterizable number of registers serving as the input operands for operation/computation #52

[yyan7223]

Cool!!!