Page 2 of 2

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 11:32 am
by Factorian
MeduSalem wrote:
Factorian wrote:
MeduSalem wrote: One thing the Devs should really think about is adding an option to the menu of a splitter to adjust which side of the incoming belt is prefered when there's congestion... I hate building a factory design around the fact that it prefers to take from the left incoming belt. I just hate it. It should either be 50:50 chance which side gets taken, the same as with the output or it should be toggle-able if it should prefer left or right or 50:50.

Maybe they should add a smart splitter that lets you control this kind of thing, maybe even by hooking it up with red and green wires to say 'prefer left if this is full, right if this is below 12' ect would be really useful in making 'smarter' factories
That would actually be an interesting idea. A smart splitter that's as fast as an express splitter... and allows you to select on which conditions it takes from which incoming belt and on which condition it outputs to a specific outgoing belt. That way you don't have to rely on slow smart inserters if you want to split a belt on a specific condition. Such a smart splitter could use a processing unit to be crafted or something since the Express Splitter already uses Advanced Circuits.
I think the smart splitter would have to be slightly slower than the express splitter for balance reasons (but slightly faster than a red splitter), the same way a smart inserter is slightly slower than a fast inserter. Being able to hook it up to the logistic networks through cables would be a worthy payoff for the speed thing though

EDIT: It turns out its been suggested before about the devs aren't keen on implementing changes to the splitters

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 11:40 am
by MeduSalem
Maybe true... for balancing reasons. I could live with that... xD

But on another note... I thought the Smart Inserters are equally as fast as the Fast Inserters? At least on the Wiki it states for both a Speed of 2 and 3.5 for rotations/sec. But well the Smart Inserter uses more power. Which is something I think splitters could use as well but then again this is a can of worms that has been opened a million times already.

EDIT: Well too bad then if they don't want to add such a thing as a smart splitter... Because I'm not keen of using robots as a master solution to everything. xD

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 11:43 am
by Factorian
MeduSalem wrote:Maybe true... for balancing reasons. I could live with that... xD

But on another note... I thought the Smart Inserters are equally as fast as the Fast Inserters? At least on the Wiki it states for both a Speed of 2 and 3.5 for rotations/sec. But well the Smart Inserter uses more power. Which is something I think splitters could use as well but then again this is a can of worms that has been opened a million times already.
I dunno, according to the first post in this thread smart is a little slower / has less of a throughput (which makes sense when you realise it has to run a check for each item to see if it is in its filter list each time and therefore SHOULD be slower than an inserter that just flings everything it can get its claw on): https://forums.factorio.com/forum/vie ... f=5&t=5400

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 11:47 am
by MeduSalem
Factorian wrote:
MeduSalem wrote:Maybe true... for balancing reasons. I could live with that... xD

But on another note... I thought the Smart Inserters are equally as fast as the Fast Inserters? At least on the Wiki it states for both a Speed of 2 and 3.5 for rotations/sec. But well the Smart Inserter uses more power. Which is something I think splitters could use as well but then again this is a can of worms that has been opened a million times already.
I dunno, according to the first post in this thread smart is a little slower / has less of a throughput (which makes sense when you realise it has to run a check for each item to see if it is in its filter list each time and therefore SHOULD be slower than an inserter that just flings everything it can get its claw on): https://forums.factorio.com/forum/vie ... f=5&t=5400
But that efficiency test isn't really comparing their working speed... It's comparing their energy usage because the Smart Inserters use more power and therefore they run out of accumulator energy or fuel in the other method faster than the Fast inserters which will continue to move items.

The method to check if they are equally as fast would be if you let them move a specific amount of items.... power them on all at the same time and then look which one of them is finished faster. I guess it will be the fast inserter because of no filter checks... but I imagine that the time discrepancy is marginally since it would only depend on how well the code for the CPU is optimized. oO

I'll test that... Brb.

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 11:56 am
by Factorian
MeduSalem wrote: The method to check if they are equally as fast would be if you let them move a specific amount of items.... power them on all at the same time and then look which one of them is finished faster. I guess it will be the fast inserter because of no filter checks... but I imagine that the time discrepancy is marginally since it would only depend on how well the code for the CPU is optimized. oO

I'll test that... Brb.
That's a very good point I didn't consider that. Yeha please let us know the results of that test! I'm interested in how big of a difference it makes

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 12:06 pm
by MeduSalem
Factorian wrote:
MeduSalem wrote: The method to check if they are equally as fast would be if you let them move a specific amount of items.... power them on all at the same time and then look which one of them is finished faster. I guess it will be the fast inserter because of no filter checks... but I imagine that the time discrepancy is marginally since it would only depend on how well the code for the CPU is optimized. oO

I'll test that... Brb.
That's a very good point I didn't consider that. Yeha please let us know the results of that test! I'm interested in how big of a difference it makes
Let both move a whole chest full of Iron ore... and they finished at the exact same time even it took them nearly 9 minutes.

Seems like they are equally as fast... and that there's no discrepancy due to more code being involved for Smart Inserters... seems like they are working Tick-based so they are never getting asynchronous in their movement due to more code being involved for anything. If things slow down due to CPU cap then they are slowing down equally.

[edit] On a second thought it would be very bad if they could get asynchronous due to code lag. I imagine that would have bad consequences for the upcoming Multiplayer part... so it is a good sign that stuff is working deterministic in that terms. ^^

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 5:42 pm
by Factorian
MeduSalem wrote:
Factorian wrote:
MeduSalem wrote: The method to check if they are equally as fast would be if you let them move a specific amount of items.... power them on all at the same time and then look which one of them is finished faster. I guess it will be the fast inserter because of no filter checks... but I imagine that the time discrepancy is marginally since it would only depend on how well the code for the CPU is optimized. oO

I'll test that... Brb.
That's a very good point I didn't consider that. Yeha please let us know the results of that test! I'm interested in how big of a difference it makes
Let both move a whole chest full of Iron ore... and they finished at the exact same time even it took them nearly 9 minutes.

Seems like they are equally as fast... and that there's no discrepancy due to more code being involved for Smart Inserters... seems like they are working Tick-based so they are never getting asynchronous in their movement due to more code being involved for anything. If things slow down due to CPU cap then they are slowing down equally.

[edit] On a second thought it would be very bad if they could get asynchronous due to code lag. I imagine that would have bad consequences for the upcoming Multiplayer part... so it is a good sign that stuff is working deterministic in that terms. ^^
Ah cool, i've been operating under the assumption that smart was slower to account for its advantages. glad to know i dont need to worry, thanks for sharing!

Re: Using a main conveyor track.

Posted: Fri Aug 29, 2014 6:13 pm
by n9103
Factorian wrote:I dunno, according to the first post in this thread smart is a little slower / has less of a throughput (which makes sense when you realise it has to run a check for each item to see if it is in its filter list each time and therefore SHOULD be slower than an inserter that just flings everything it can get its claw on)
This immediately popped into mind when I read this sentence.
I can't be helped. :lol: