chore: leaf aggregation tuning

[jonathanpwang]

towards INT-2940

Change DEFAULT_NUM_CHILDREN_LEAF=1 because I think the App VM is usually very wide and its better to do a 1-1 aggregation first to get to a Agg VM proof. This is configurable from the AggConfig.

Update the metric labels a little so that group scopes based on the aggregation layer.

[github-actions]

Benchmarks

group	app_log_blowup	app_total_cells_used	app_total_cycles	app_total_proof_time_ms	leaf_log_blowup	leaf_total_cells_used	leaf_total_cycles	leaf_total_proof_time_ms	max_segment_length	instance	alloc
ecrecover_program	2	15,230,037	290,016	(+16.0 [+0.7%]) 2,403.0	2	445,726,715	9,777,953	34,875.0	1048476	64cpu-linux-arm64	mimalloc
fibonacci_program	2	51,505,102	1,500,137	5,503.0	2	(-2,570 [-0.0%]) 128,966,671	(-334 [-0.0%]) 3,173,374	(+485.0 [+3.6%]) 13,836.0	1048476	64cpu-linux-arm64	mimalloc
regex_program	2	165,028,173	4,190,904	(-23.0 [-0.1%]) 15,800.0	2	(+5,730 [+0.0%]) 291,417,182	(+615 [+0.0%]) 6,525,792	(+649.0 [+2.5%]) 26,310.0	1048476	64cpu-linux-arm64	mimalloc
verify_fibair	2	(-360 [-0.0%]) 8,121,174	(+27 [+0.0%]) 195,357	(+4.0 [+0.3%]) 1,477.0	-	-	-	-	1048476	64cpu-linux-arm64	mimalloc

Commit: 3407330

Benchmark Workflow