[PATCH 1/3] Speed up and fix genautomata

[Vladimir N. Makarov]

On 12/04/2009 01:45 PM, Segher Boessenkool wrote:
> The algoritm used to find min_issue_delay_vect is incorrect; it
> will not always consider paths that take more steps to arrive at
> the same node.
>
> Instead, use a simple standard breadth-first algorithm, that is a
> whole lot faster as well, especially on big complex automata.
>
> 2009-12-04  Segher Boessenkool<segher@kernel.crashing.org>
>
> 	* genautomata.c (curr_state_pass_num): Delete.
> 	(min_issue_delay_pass_states): Delete.
> 	(min_issue_delay): Delete.
> 	(initiate_min_issue_delay_pass_states): Delete.
> 	(output_min_issue_delay_table): Compute min_issue_delay_vect
> 	using a breadth-first search variant.
>    
The patch is ok for the trunk.  Thanks for the patch, Segher.