Tomasulo: Instantiate all 6 RS types in wrapper, add multi-RS tests

  Add MUL_RS (depth 4), MEM_RS (depth 8), FP_RS (depth 6), FMUL_RS
  (depth 4), and FDIV_RS (depth 2) to tomasulo_wrapper alongside the
  existing INT_RS. Dispatch routing decodes rs_type to per-RS valid
  signals; o_rs_full is muxed by targeted RS type. ICARUS path keeps
  single INT_RS with tie-offs for new outputs.

  Update cocotb model and interface for per-RS dispatch, issue, status
  and CDB broadcast to all RS instances. Add 7 new integration tests:
  dispatch routing, CDB wakeup across all RS, per-RS full independence
  flush coherence, partial flush, FMUL_RS 3-source FMA with rounding
  mode, and 300-cycle random multi-RS stress test.

Author: adambagley