Multi-element packet stacking.

[Tonyroid]

Part 2: Why my packet stacker sucked and how to make a good one.

 

[Lilalaunekuh]

5 hours ago, avc15 said:

- It will still send small packets on occasion.

The idea to work using reservoirs got many flaws and isn´t the best one, BUT:

This isn´t one of the flaws.

=> You can guarantee a stream of always filled packets (1000g).

[But you need to sort your packets, reducing the need for a reservoir in the first place.]

[TLW]

5 hours ago, Lilalaunekuh said:

[But you need to sort your packets, reducing the need for a reservoir in the first place.]

If you've sorted the packets already, you don't need a reservoir to do packet-stacking. The issue is that this is rather large and finicky. (Even with 7 gasses that's, what, 14x6 tiles at a minimum? And 7 manual setups of specific gasses? All to combine two pipes.)

17 hours ago, Yunru said:

Only when the tank's full, which is kinda the point... If you're trying to say what I think you are? 

I'm not trying to say what you think I'm trying to say.

 

If your output pipe isn't backing up, then a reservoir won't consolidate packets.

This causes it to be unsuitable for most of the cases where you want to combine packets in the first place.

Want to prevent small packets changing temperature too quickly? This isn't suitable.

Want to prevent CPU load from a long run of a mixed gas pipe? This isn't suitable.

Want to be able to merge pipes in a sane manner? This isn't suitable.

 

Take another look at my example, please. Something that actually packet-stacked would have full throughput. Using a reservoir does not. This example has just over half throughput.

[Lilalaunekuh]

2 minutes ago, TLW said:

This causes it to be unsuitable for most of the cases where you want to combine packets in the first place.

Sure, I hope we are all on the same side here.

 

3 minutes ago, TLW said:

Something that actually packet-stacked would have full throughput.

Ok, that is a point we could discuss but it would be more a meta discussion.

 

4 minutes ago, TLW said:

Using a reservoir does not. (have full throughput)

It´s big, it´s clunky and maybe many other things ...

But you are able to use it to achieve full throughput.

[Yunru]

12 hours ago, TLW said:

If you've sorted the packets already, you don't need a reservoir to do packet-stacking. The issue is that this is rather large and finicky. (Even with 7 gasses that's, what, 14x6 tiles at a minimum? And 7 manual setups of specific gasses? All to combine two pipes.)

I'm not trying to say what you think I'm trying to say.

 

If your output pipe isn't backing up, then a reservoir won't consolidate packets.

This causes it to be unsuitable for most of the cases where you want to combine packets in the first place.

Want to prevent small packets changing temperature too quickly? This isn't suitable.

Want to prevent CPU load from a long run of a mixed gas pipe? This isn't suitable.

Want to be able to merge pipes in a sane manner? This isn't suitable.

 

Take another look at my example, please. Something that actually packet-stacked would have full throughput. Using a reservoir does not. This example has just over half throughput.

Then no, I'm not saying what you think I'm saying. You link the overflow sensor to an mechanised airlock that disables the tank. Then it only sends packets (which will naturally be the largest possible) when full.