The Factorio Benchmark Website

test-000060 : Does priming inserters matter when the inserter is part of a DI chain?

Factorio Version 1.1.107

The TLDR

Priming is worth it even in the DI with a chest handoff case.

The Question

Does priming^[1] inserters matter when the inserter is part of a DI^[1] chain? Specifically, the case where the DI chain has backpressure.

When the inserter pulling from the source assembler crafts, if it is not primed then the inserter appears to wake for a tick. It could be that this contributes a significant effect to performance.

The Test

Test consists of two maps, both of which contain designs to produce rocket fuel. The solid fuel is created, then inserted into a chest, then inserted into the rocket fuel assembler.

Both designs are cloned using the region cloner mod to a large amount of production.

Each map gets benchmarked using the inbuilt Factorio benchmark. 3 runs are done, after which the lowest timing tick is taken.

The Data

• Priming inserters is clearly better, although the effect is small
• Most ticks appear to have greater performance, around 10% much more so.
• The dips (aka times when performance is higher) have not been closely examined, though it's likely due to all clones being in the same state of their production. In real maps this is unlikely to be the case.

All maps will be uploaded here.

Bonus test

flame_Sla pointed out that the inserter between chest and rocket fuel assembler could be set to 11 items instead of 12. That would lead to natural priming of the solid fuel to chest assembler for 11 out of 12 swings.

Somewhat surpsingly, that did perform better than the 12:12 case without priming. As expected, did perform worse than the primed case, though may still be a good middle ground to avoid the tedium of setting up priming for inserters.

Closing

Although tedious, making sure inserters are primed appears to be worthwhile if playing for maximum UPS efficiency.