- No longer update global history in fetch stage.
- Only update global history when instruction is a branch.
- Does allow for the possibility of back-to-back branches to see
slightly different histories on subsequent executions.
- Updated in fetch speculatively.
* Updates gated off by cpu.resp.fire().
* BTB direction factored into history update.
- All branches update the BHT.
- Each instruction carries history; index into BHT is recomputed by
passing in mem_reg_pc.
BTB entries reference a small number of unique pages, so we separate the
storage of pages from indices. This makes much larger BTBs feasible. It's
easy to exacerbate cycle time this way, so one-hot encoding is used as needed.