opt: add another bytecode circuit to improve super circuit capacity

[DreamWuGit]

Description

details of this PR targets to implement

• evm circuit
  • add one more bytecode table (name bytecode_table_1)
  • modify SameContextGadget&CommonErrorGadget to do conditionally lookups base new field is_first_bytecode_table
  • implement algorithm find_two_closest_subset to dynamic split bytecodes into two lists
  • update opcodes gadget since SameContextGadget::assign_exec_step changes to include block, call.
  • modify dynamic copy table lookup with new field is_first_bytecode_table
• copy circuit
  • copy table: add new flag column indicates which bytecode table is used.
  • do conditionally lookup bytecode table along with new flag column: is_first_bytecode_table
  • constrain is_first_bytecode_table: bool type & not changed in one copy event .
  • modify assign to pass bytecode_map to get which bytecode circuit is belong to.
• super circuit
  • add new bytecode circuit and assign correct part of bytecode sequences.
  • update ccc rows in helpermin_num_rows_block_subcircuits by BytecodeCircuit::min_num_rows_block
  • add tests for multi bytecodes instances so that can be group into two sub circuit.
• misc
  • add rust feature dual_bytecode to enable/disable two bytecode circuit side by side functionality.
  • testing
    • CI tests include new feature
    • mock prover integration tests
    • test tool
  • WIP: after testing then refactor or optimize some codes (opcode_lookup etc.)

[lispc]

change assign_exec_step input to also include block and call. Inside SameContextGadget, alloc the cell and do assignment.

[DreamWuGit]

change assign_exec_step input to also include block and call. Inside SameContextGadget, alloc the cell and do assignment.

after included block and call , exec_step no need to change as got enough information already.

[z2trillion]

I think it is much cleaner to add a in_first_bytecode_table boolean field to Bytecode. With that, we won't need to pass around bytecode_map anymore. I have a prototype PR here: #1402, to make this more clear.

[DreamWuGit]

I think it is much cleaner to add a in_first_bytecode_table boolean field to Bytecode. With that, we won't need to pass around bytecode_map anymore. I have a prototype PR here: #1402, to make this more clear.

looking at it today.

[DreamWuGit]

add comment for stable sort helper explicitly (5e840fa)

[roynalnaruto]

Preliminary check done. I have some suggestions, some nit-picks and some opinions. Let me know what your thoughts are

[roynalnaruto]

Looks good! :)