[postgis-devel] Costs, Parallelism and Regressions

[Paul Norman]

On 11/3/2017 12:27 PM, Paul Ramsey wrote:
> On the former issue, I'm not sure how much we trust the empirical cost 
> scoring? More importantly, if we can force parallelism by increasing 
> our costs, even beyond the empirical values, do we want to do that?
>
> Any other knowledge of parallelism or costing folks have and can link 
> to or share, would love to hear!

If our costs are right and the query costs are accurate, forcing 
parallelism in the face of a higher parallel cost would reduce performance.

Because perfect costs and planning is not possible, we might find it 
better to err on the low side or high side of what the data says is the 
best cost.