Hey together, as it currently looks like are we more or less forced to re-ban the thermal expansion item conduits and minefactory reloaded rednet stuff. I think I just let the pictures talk and you know that there is no easy fix in view atm. I will leave this open for discussion over the weekend. MineFactory Reloaded - Rednet ThermalExpansion - Item Conduits ThermalExpansion - Item Conduits with pneumatic upgrade RotaryCraft and Applied Energistics for comprehension *everything is profiled about more than 8 minutes The forge ore dictionary is not build for that many accesses it could be improved but thermal expansion is the source of the problem as they miss-use the ore dictionary and don't cache the stuff for example. The MineFactory Reloaded Rednet stuff uses linked lists which is not the best as choice, as our devs told me they should be using hash sets maybe they will add this in the future. Unfortunately the mfr rednet stuff is the only one with computercraft support , for everything else there should be a replacement.
This part has me rather upset I must admit, even though I'm not currently using it in my build on M+ I do have a number of things I was considering doing with ComputerCraft that I would need setBundledOutput for, so it will be a real shame if this has to go... have you contacted MFR's developer about this issue yet? As for the TE item conduits, it would be interesting to see a direct comparison between them and the enderIO item conduits in an otherwise identical large-scale design, cause I'd hate to have to go through the trouble of replacing all my TE item conduits with enderIO item conduits only to find they're just as big of a problem...? Unfortunately there are parts of my build that would be really hard to do without the use of an item conduit of either type cause XU item transfer nodes (despite being my preferred choice for general moving of items) and BC pipes both have significant limitations that would require a far more complicated setup to make parts of my system still work :/
We did contact the MFR dev, but he didn't do anything yet. I also don't think he will do soon as he's currently updating to 1.7. As the mod is open source I tried to patch it myself, but it seems that the compilation for 1.6 isn't working any more...
I can't give you a an exact comprehension between TE and EnderIO, as the problem is that the main issue of TE is the pneumatic upgrade with the ore dictionary and ender io gets problematic when the system gets big. So if you have multiple small systems, ender io is your friend. If you have one big system, it is thermal expansion if we keep the ore dictionary by side. So in order to be able to get some stats we could compare we would need two servers with the same setup from multiple peoples in multiple sizes and use cases, kinda impossible. But if we look back to the time before we reset as we banned te conduits because of another problems the ender io conduits did a very good job at the beginning and got worse from time to time when the systems got bigger and players started to bundle them but I still can't compare them as it really depends on the systems and another point is that the enderio conduits slowed down all sorts of machines and it is kinda hard to put all this together into one big number. According to performance the best but also expansive choice is applied energistics. According to your situation, the logistic pipes might be a good way too. We already tried to solve the TE ore dic lookup, but it failed, maybe we get another idea which we can try, but as this mod is obfuscated it is kinda hard to achieve things even if we know exactly what needs to be changed.
Ender IO are the best thing to happen to machine MC item piping... When used as a single block. When the get bundled is when the issue arises. Is there any way we could ban the placing of 'conduit bundle'? I'm telling you, would help significantly. ESPECIALLY once Ender IO comes out of Beta for monster (next patch).
Well id go for ender io. - ender io has higher bandwidth leading to less facing-connection, leading to less conduits used - if you want reactorcraft, your options of power storage are very limited, even the power bank gets small with rc. Still the large powerbank should be less hazardous than an army of resonant cells and conduits on them - ender io has the power monitor to simply shut off power production, the alternative beeing cc .. which i think woul again use more power And we need at least one of the two since the most rotary setups dont work with extra utilities. I got some large XU pipes with many energy nodes and upgrades if you want to profile them btw
the problem is that opis tells us only have of the truth, it tells us the update time per item but not per whole system or things which happen next to it.
So what do the numbers in those pictures represent exactly? Are we looking at ms spent per tick? Or for the entire 8 minutes it was running? Also, do we know for sure that if a number is high, what sort of processing time per block we're talking about? If a million lycanite mobs are spawned and running around the world (for example) it would only seem natural for the lycanite update process to be very high. Likewise, if people have 30k TE item conduits running all over the place, it may be super efficiently coded, but the update time might still be enormous. Is that taken into account? (I haven't used Opis or whatever tool you're using there, so this is just a curiosity thing.)
all this is taken into account and our devs looked into it, too. The ore dictionary which is accessed by te every tick is not written for this and the mfr rednet uses linked lists where it has to go through all, instead of hashsets which would be ways faster. We look into the code before we go to these actions. However the only interesting number for you is the % and I can ensure you that there were not many lycan mobs on the map, maybe 1/10 compared to vanilla mods, but unfortunately I did not make a screenshots from the entities, but ofc I checked them. Else I would just have used aperf to limit them like the pixies.
THanks guys for taking the time to explain to the community the situation and thought process you guys go through for a big decision on a mod. I"we" appreciate it.
Good news for you all: I fixed the rednet bug, but the fix can't be applied in all modpacks as some use an older version of MFR. So I can't update them just now, but I'll discuss with @Slind14 how to distribute the fix.