This PR optimizes the start times for block producing. Before this PR, a producer always starts producing each block at the time point of the multiple of config::block_interval_us (i.e. 500 ms) and spend at most cpu-effort-percent (i.e. 80 by default, so it is 400 ms out of 500 ms) to produce each block. Therefore it leaves 100 ms idle time in between each block producing. In this PR, we leave no gap in between block producing so that the last block in each production round can reach the next producer earlier so that the next producer can start producing on schedule.

Removed configuration options:

• produce-time-offset-us
• last-block-time
• last-block-cpu-percent

The schema for Runtime_Options has changed to

struct runtime_options {
      std::optional<int32_t>   max_transaction_time;
      std::optional<int32_t>   max_irreversible_block_age;
      std::optional<int32_t>   cpu_effort_us;
      std::optional<int32_t>   max_scheduled_transaction_time_per_block_ms;
      std::optional<int32_t>   subjective_cpu_leeway_us;
      std::optional<double>    incoming_defer_ratio;
      std::optional<uint32_t>  greylist_limit;
   };

Resolves #726