Energy Net improvements

Loool that compact reactor would not even possible xD

Choco, just a question, do you have any documentation on how to make GUI + tile entities with working sideness please ? It's for my evil plans... ( create a small fusion reactor mod working with experimental and classic , as Speiger doesn't want to include it in the main mod ), here is how it will work ( note that not existing items will be custom ) :

- Extract electrolysed cell into deuterium
-place deuterium in the 2 slots in the reactor ( there is a total of 3, 2 input of D, 1 output of plasma, the fusion would be passive, so it creates plasma over time only using deuterium, but I am gonna make the fusion reactor very expensive, requiring beacons for passive ignition xD )
-use 4 plasma cells by placing them on the first slot in a plasma extractor and place an empty energy Crystal ( must be this one ) in the second slot, and get your fully charged energy Crystal out of the reaction ( note that this is proof of concept, I want to make them require constant EU draw but I don't think this is possible ).

No way, since the first beta of Ic2 classic from Immibis and the last version from Speiger, i always did elect Ic2Classic, and trust me, i'll always choose it.In my personal opinion when ic2E devs started to implement their ideas, wanted to make a full new Industrialcraft, but they did 2 things wrong:

1)Instead of focus on a certain point and hardly work on it ( make something that is stable, and works fine) , what they did is just implement little stuff, filling Ic2 of random features, and when they started to work on that big point that needed attention and stabilization, they saw themselves frightened for the difficulty that the job was going to take, so they leave it. After sometime, they decide to get back on track on that big point, finding themselves in the same situation.


2)The little and random implementations were added just to make a Difference from the original Industrialcraft2 to their version, features that don't were requested by the community, or were what the mod neededIt's ok, some people do like IndustrialcraftExperimental, and after all this discution, well, we should pay attention to the word 'experimental', but in my opinion, Ic2Experimental never had to exist.LONG LIFE TO 16x16 TEXTURES!

Also a note: They could come every time i release a version and say we do not want this anymore...
They could not prevent the current version because they gave me permissions to release it but further work could be stopped by that (since i release it on github which has his reasons(partly control))

Excuse me i Necro an old post, but I've noticed that barely any progress has been made with the E-net for a very long time. Having a bit of experience with tile logic in a few of my projects, I can offer a few alternatives, some that might even improve the previous methods of e-net calculations.

One such method is to simply allow each cable to ONLY accept up to it's maximum potential through any one point. This would allow a physical limit, and give incentive to upgrade to higher tier cables, as well as remove the "iffy" logic of "add one more generator, everything goes 'splodey". This method would simply force users to either use better cables, or design smarter designs. (Something that AE2 is currently promoting, I might add)

Speaking of which, that too is another method that could be used for the new e-net. It would create "channels" of each cable, and simply use this as an array to designate emitters and receivers, at which I would recommend adding a calculation to split emitter's output so that all emitters output into the same "channel" equally. That would remove the issue of draining one storage before another.

Now, for the mechanics of the e-net, many features could be implemented easily. I for one, would recommend that a flat energy loss does not get used. The reason is that it would be a disadvantage, and there would be no incentive to build smarter wiring. Everything would just be a mess. Tons of cables means large arrays, and more calculations. If the player is encouraged to build smarter, then the game will run alot smoother.

First mechanic: Energy loss is calculated as such: (WARNING - MATH AHEAD)
Loss = floor(CableCount/ConfigCluster)
Where Loss is the percentage loss of EU
CableCount is the amount of connected cables
and ConfigCluster is the value at which 1% energy is lost for every x number of connected cables

This calculation means that for every "cluster" of wires
(clusters get calculated by every touching wire, gets re-calculated on block updates)
there is a percentage loss for every packet transfer, starting at the point that the "cluster" contains x connected cables. This, of course, means that super long cables incur a percentage penalty when they're chained together, but simply adding a transformer or energy storage block resets the loss back to 0.

It isn't a perfect fix, as it doesn't take into account the distance between inputs and outputs, but until a better solution is made, it's probably the best option.

I think you misunderstood a couple of things, although I see the consideration.

The cables themselves actually won't tick to search for connections. That is only done whenever there is a change in that particular line.
The calculation would only be in accepting eu from emitters and transferring the eu to recievers. And as for the "parallel" eu that we see in sending multiple packets, that's effectively what this is doing. Except instead of sending the individual packets, they're rather divided up *to reach* the cable maximum current. This also means that if you're drawing 33 eu on a tin cable, you will reach a deficit, even with 3 batboxes powering one line.

The percentage loss is an incentive to /not/ make huge cables, and in doing so, make smaller (and less cpu intensive) power systems.
A pretty happy median is about 16 blocks. At that point, making a cable 100 blocks longhas a 6% loss. Even at 128/t with this, you're losing 8eu.
Could be more, but that's up to the balance team

20 cables of 10 wires apiece in theory is less laggy than 2 cables of 100 wires apiece for calculating connected tiles.
For transferring eu, only the emitters and receivers get calculated.
For a game mechanic, cables could be made to only connect to the same type of cable, allowing for more compact wiring.

As for the percentage loss, remember that it is only 1% for every number of cables. Thinking back, didn't we have something similar? I remember solar panels being useless after 5 cables of distance.
Higher tiers of cable /should/ have a higher loss, if designed improperly. Besides, if you're maxing out a gold cable with 512eu/t, then you can afford to lose 5eu/t for connecting a machine up to 60 blocks away. Honestly, that's generous.

Has player given any consideration to simply adopting Greg's model?

I really like player's design, but it has a couple problems:
1) he hasn't been able to optimize it to his satisfaction, afaik
2) it may not be intuitive to players (although to be fair I have yet to see how feedback is to be presented to the player)

Greg's model is really simple to understand, has functional loss, and seems to perform satisfactorily as far as I can tell. It also dodges the age-old problem of wiring your entire base with fiber to avoid loss, which is nice.

Quote from Chocohead

Yet people make posts about "help why does this happen?!" and the like. The old IC2 system was simple, and anything about the new system being complicated or anything is just speculation. Player already fixed the fibre your base up problem by making loss per cable anyway, so you'd literally only be lowering the loss per block using fibre rather than avoiding it completely.

Do you still see those posts anymore?

I remember seeing them all the time on old IC2, but I don't remember the last time I saw such an inquiry on the GT thread. I've never seen it in our modpack thread.

I'll be honest, I frequently found the classic IC2 system hard to wrap my brain around, and some things seemed frankly counter-intuitive. But in GT, it all makes sense and happens the way I'd expect: machines can request X amps of energy. If the wiring can't handle X amps, it burns. If the voltage on one of those amps is too high, the machine explodes.

Don't ask me to clarify IC2 un-intuitiveness, because I forget details. Maybe you were able to carry as many packets as you wanted on a line? I forget.

Quote from Natesky9

20 cables of 10 wires apiece in theory is less laggy than 2 cables of 100 wires apiece for calculating connected tiles.

Well... No... Even if you place 100K of IC2 cables they do not lag at all. (thinking that the enet of Exp is not counting)
These cables have maybe tileEntity data but they are not ticking TileEntities, they simply represent only the cableType, connection info, cableData for Enet.
Nothing else is done in that... (They tick if you foam them but even after they did dry up the do not tick anymore...)

You idea is... Well sorry crap...

First adding the idea of channels to cables is totally useless because electricity do not care about how you want the energy to flow... It flowes always the easiest way...
Also for what do you need channels? Flowing differend amounts of EU through the same cable? Why that is useless and also creates more lag then you think because each channel would have his own path which would mean you gain a lot of path and 9x more calculations which is way more laggy. Even with the Classic network which would simply say: "i am away fockers have fun with your crap"...

And also saying that you could build with 8-9 channels smaller and smother(performance) is a lie... Even the cake would be less a lie then that... (Portal)
You simply could with (uncomplication/IC2Classic) send multible packets of low energy through the Enet and have even a better result of sending power without explosions... or you put in transformer upgrades... Same solution...

Well your Math has one mind lost. You said all connected cables. You know that then the Powerloss would be way bigger then it is currently...
Second that is how powerloss works in (uncomplication/IC2Classic). It builds simply path (when ever you place a electric cable/machine/generator) from Sender to Receiver and back and calculates the Energy loss to it,
and every time energy get sended through that preset path it decreases the energy by the already calculated loss... (Good thing by the way is only the Sender Create lag so you do not need to care about placing receivers or cables)

And that to write took me 10 minutes which could have used to fix 3 bugs for classic... thanks man...