Concerning Unit sizes

[Yura]

Hey guys, so far I have just been playing with the "normal" unit size settings, however now in a new campaign I wanted to try the biggest one. I was looking forward to see some awesome battles and maybe I am just biased from my "normal unit size" playthrough and used to the smaller units, but! it seems to me that battles with such huge units usually just become soooo tiresomely long, as only the front line fights with the rest being stuck in the back (even if there is enough space for them to "disperse" more).

Also, I find it very difficult to even deploy my units in a defending siege as they are just sooo many to fit in the forts / cities... anyone having the same feeling? Thanks for any thoughts! :-)

[LeadStarDude]

Normal (medium) or even low unit size works best on siege battles for sure. The game plays much better that way. The AI will have less pathfinding problems on lower unit size settings.

[Tomislav]

I always play on Ultra unit sizes. It feels weird with smaller units to me.

[Desmond]

I feel they are to few on ultra......

[Migz - DH]

I guess it's a good thing CA provided a lot of unit size options for people then.

[LeadStarDude]

I know everyone likes bigger armies, but bigger armies is what causes many of the siege battle bugs. Try playing a siege battle once with Ultra unit size & then again on medium. I am sure you will notice that the AI seems smarter with fewer units to process for.

[Shadow of the SPQR]

Tbh, too many men/units makes it a bit of a cluster♥♥♥♥ for me.

[Kayeka]

Originally posted by CommodusIV:

Tbh, too many men/units makes it a bit of a cluster♥♥♥♥ for me.

True, but too few makes Total War look like Minor Skirmish. I play with Large unit sizes myself.

[Hannibal Barca]

Originally posted by LeadStarDude:

Normal (medium) or even low unit size works best on siege battles for sure. The game plays much better that way. The AI will have less pathfinding problems on lower unit size settings.

pathfinding has nothing to do with unit size, the units run on tracks behind the scene 100 units is represented by a point on that track. Bad coding is the pathfidning issue. If you had said too many units bog down the processor so it doesnt respond fast enough then maybe but this game is suppose to be able to run on dual core processor.

[LeadStarDude]

Originally posted by DeathMarch:

Originally posted by LeadStarDude:

Normal (medium) or even low unit size works best on siege battles for sure. The game plays much better that way. The AI will have less pathfinding problems on lower unit size settings.

pathfinding has nothing to do with unit size, the units run on tracks behind the scene 100 units is represented by a point on that track. Bad coding is the pathfidning issue. If you had said too many units bog down the processor so it doesnt respond fast enough then maybe but this game is suppose to be able to run on dual core processor.

I think the seige battle problems are caused by a game engine flaw that bottlenecks the CPU. If the CPU is being overloaded by AI processing then the game engine is more likely to pick the first pathfinding option available which will most likely be the wrong option. The AI is completely on the CPU end & so is the unit pathfinding. The main reason so many people have problems with siege battles is because there are more unit paths available (pathfinding options). That is a lot more work for the CPU to do when processing for more units (larger unit sizes) on screen. Basically the higher number of AI units coupled with more pathfinding options available during siege battles creates a type of CPU bottleneck that causes either stuttering, or it causes AI to basically not know what to do soon enough so it just sits there or goes the wrong way. There is a lot more for the game engine to figure out during siege battles than on normal open field battles, & that stresses the CPU a lot more than normal. That is why CA suggests a high end i5 or i7 CPU to run this game on the highest settings.

CA recommends Haswell to play Rome II for a reason. They utilized some of the new features Haswell offers that Sandy Bridge doesn't such as the improved front-end and memory controller.

Haswell Performance Compared to Ivy Bridge:
Approximately 8% better vector processing performance.
Up to 6% faster single-threaded performance.
6% faster multi-threaded performance.
A 6% increase in sequential CPU performance (eight execution ports per core versus six).

New features:
Wider core: fourth ALU, third AGU, second branch prediction unit, deeper buffers, higher cache bandwidth, improved front-end and memory controller
New instructions[22] (HNI, includes Advanced Vector Extensions 2 (AVX2), gather, BMI1, BMI2, ABM and FMA3 support).[23]
The instruction decode queue, which holds instructions after they have been decoded, is no longer statically partitioned between the two threads that each core can service

I have not been having any issues with seige battles with my 4.0GHz Haswell i7. Any of you that do have problems should try lowering Unit Size.

[Hannibal Barca]

explain how this problems occurs with under 500 units on the map. I ve never had bottle neck problems with my 8 core 4.3ghz processor. I still see siege issues with limited number of unit on screen. CA themselves said its not a cpu issue if it was they would have fixed it by now. Its coding and coding is alot harder to fix than resource managment , I dont know if you know this but steam stats say vast majority of people that play Rome2 have 4core cpu's so why on earth would Sega make a game that only runs on the high end cpus.

[Migz - DH]

Originally posted by DeathMarch:

I dont know if you know this but steam stats say vast majority of people that play Rome2 have 4core cpu's

Where did you find that stat?

[Hannibal Barca]

http://store.steampowered.com/hwsurvey/cpus/

[Migz - DH]

Originally posted by DeathMarch:

http://store.steampowered.com/hwsurvey/cpus/

Ya, I thought so. That's not statistics for Rome 2 players specifcally; that's all of Steam users, so it's not useful in this situation, and the comment I quoted in post #11 should be removed from post #10 imo.

[Hannibal Barca]

0.28% have 8 core cpu , your going to me they all play Rome2? how many of those are haswell? lets say .14% , lets make a game .14% of people can run?
or for 40% of the people have 4 core.

[Migz - DH]

I'm not saying anything about what CPUs Rome 2 players use, nor do I care.

I'm saying that you cannot use that statistics page you linked to prove your comment above, because it's inapplicable.