Re-examine HyperDrive params that depend on the size of aggregations

[arajasek]

#6515 revealed an issue (since fixed) in proof de/serialization that was leading to things being bigger than expected.

We need to reassess values

• We need to rerun gas calculations to see if there's a significant change in validation time resulting from the fix (a slowdown is expected, but unclear how large)
• If the change is small, we proceed with our current values
• If the change is large, we recompute gas values (takes ~2 days)
• If we recompute gas values, we also need to revisit bounds on aggregation batch sizes in actors

If anyone knows of any other values similarly affected, please flag below

[jennijuju]

small is defined as within 20% by @Kubuxu

[jennijuju]

Also the balancer/gas discount value?

[Kubuxu]

The result is 6% slower. Closing.