Is the Fluid Rework still planned?

[DarkShadow44]

Hello,

I hope this wasn't already asked, but I couldn't find an official statement for this.
In FFF-274 a new fluid-system was proposed.
Is that still on the roadmap, definitely scrapped, or "possible sometime in the future but no promises"?

Regards,
Dark

[NotRexButCaesar]

It is pretty much scrapped. I belive Rseding said something along the lines of "the guy who wrote the fluid system is gone, and no one knows how it works anymore"

[jodokus31]

IIRC Some significant performance improvements were implemented.

[Koub]

viewtopic.php?p=520955#p520955
But I don't lose hope because :
viewtopic.php?p=475106#p475106

Actually, the fluids rework was on two aspects : performance wise, and gameplay wise.
- performance wise, it has been done, fluid network performance is very significantly better than it was in 0.16 for example.
- gemeplay wise, there are still issues. It's playable, but for whoever tries to squeeze the maximum theoretical throughput from one's fluid network, oddities start to appear.

[DarkShadow44]

I see, thanks! I asked because I saw that the fluids are still distributed kinda random across junctions. Not sure if they have different flow speed already.

[Yoyobuae]

.

[DarkShadow44]

Usually if pipes are so empty that you start seeing these issues there's likely some problem other causing that (ie. not enough production, fluid flow problems, too messy/complex pipe system which hides potential production and fluid flow problems).

Yes, I usually address those "issues" by a slight overproduction, when the pipes are completely full it's not big of a deal.

[NotRexButCaesar]

95194
A more recent "no"

[FasterJump]

I haven't run huge nuclear plant designs since years, but I remember it was a hassle because of the low throughput of pipes (had to run 4 sets of 8 parallel pipes IIRC, so 32 pipes. A lot of things have been announced about fluid, few made it to the core game.

I'm curious if it's any better now (throughput, FPS, symmetry, etc) compared to few years ago, do we still have strange behavior such as unequal split of liquid at junctions? Is there a lot of players experiencing frustration in today's game? I was especially excited about the implementation of different fluid density, which would have solved the throughput limitation I experienced in my 1RPM oil processing area.

It seems like fluid update is the last impossible challenge remaining unsolved.

Edit: Thanks AmericanPatriot for the detailed answer!

[NotRexButCaesar]

I haven't run huge nuclear plant designs since years, but I remember it was a hassle because of the low throughput of pipes (had to run 4 sets of 8 parallel pipes IIRC, so 32 pipes. A lot of things have been announced about fluid, few made it to the core game.

Pipe throughput does depend on distance between pumps. You can carry enough water for 10 water pumps in a line of fluid pumps and tanks.

I'm curious if it's any better now (throughput, FPS, symmetry, etc) compared to few years ago

There were many optimizations added. It is way faster UPS wise. Throughput should be pretty much the same.

do we still have strange behavior such as unequal split of liquid at junctions?

Yes

Is there a lot of players experiencing frustration in today's game?

Maybe? I can't speak for others, but I don't fine the fluid system debilitating.

I was especially excited about the implementation of different fluid density, which would have solved the throughput limitation I experienced in my 1RPM oil processing area.

No, that FFF system was not added.

It seems like fluid update is the last impossible challenge remaining unsolved.

You might say so. Especially now that we have train limits.

[Durabys]

Hello.
Nowhere did I see how this is going get tackled but we all were once promised UPS/FPS friendly Liquid Mechanics rework. So what happened?

[Bilka]

Hello.
Nowhere did I see how this is going get tackled but we all were once promised UPS/FPS friendly Liquid Mechanics rework. So what happened?

Fluid UPS optimisations (multithreading) have been in the game since 0.17: https://factorio.com/blog/post/017-stable https://factorio.com/blog/post/fff-271

[Drury]

Hello.
Nowhere did I see how this is going get tackled but we all were once promised UPS/FPS friendly Liquid Mechanics rework. So what happened?

I think you may be referring to Dominik's attempt at overhauling the fluid system? He left the company before he could finish.

[Durabys]

Hello.
Nowhere did I see how this is going get tackled but we all were once promised UPS/FPS friendly Liquid Mechanics rework. So what happened?

Fluid UPS optimisations (multithreading) have been in the game since 0.17: https://factorio.com/blog/post/017-stable https://factorio.com/blog/post/fff-271

I was talking about what Dominik had been working on. Do you intend to have one more look at it?

[Yoyobuae]

.

[Zistack]

I just fired up the game to test fluids 'cause I saw in the last FFF that they said that they no longer had any known bugs or internal tasks. I thought it was odd that they never mentioned the new fluid system. I was disappointed to see that the old fluid system is still there. I didn't realize that they had scrapped it internally. That was probably the one feature of the game that I was most looking forward to.

I find this really disappointing. The current system makes optimizing complex fluid distribution networks (or even some relatively simple ones) basically impossible, since the behavior is so inconsistent and unpredictable. The current system is still usable, but you basically have to over-build by a factor of 2 or more in order to have any confidence, which is going to unnecessarily eat up your CPU time.

[Silari]

Hello.
Nowhere did I see how this is going get tackled but we all were once promised UPS/FPS friendly Liquid Mechanics rework. So what happened?

Fluid UPS optimisations (multithreading) have been in the game since 0.17: https://factorio.com/blog/post/017-stable https://factorio.com/blog/post/fff-271

I was talking about what Dominik had been working on. Do you intend to have one more look at it?

They already said a while ago the fluid rework was shelved. I believe it was in the 1.0 date announcement FFF, but it's been a bit so it might be somewhere else.

I was talking about what Dominik had been working on. Do you intend to have one more look at it?

The fluid system is arguably the most UPS performant part of factorio.

The UPS improvements were only a part of the planned changes. It was more about getting liquid flow to be more predictable - like making pipe build order not matter.

[Yoyobuae]

.

[Yoyobuae]

.

[SoShootMe]

The current system makes optimizing complex fluid distribution networks (or even some relatively simple ones) basically impossible, since the behavior is so inconsistent and unpredictable.

Show me one of these "impossible" to optimize fluid distribution networks and I will show you what's wrong with it.

The fact that achievable flow is dependent on build order etc is well known. It means that, when attempting to optimise a design, you may have something that behaves as you intend, but when you replicate it somewhere else, it no longer does.

The only thing "wrong" with such a design is that it requires more flow than is achievable in the worst case. Due to the "inconsistent and unpredictable" behaviour (the reason a worst case exists), you can't easily tell when this may arise, and the practical solution is to be conservative and hence no longer aim for optimal.