Promised performance increases observed - with Patch 16.1 - updated 05.02.2015

[billydilly]

That's with Patch 10.

When I use custom settings the FPS is around 50 at the start of battles. But when I zoom all the way in on the fighting, the FPS drops as low as 15-19.

[alQamar]

Could I bug you to post screens of patch 9.1 with the same settings (high, very high), please?

[billydilly]

Yes, I'll do it tonight.

[Falconwatch]

Just a question for those that have R2, is the game engine now optimized for quad cores to benefit performance?

And how many CPU cores does the older Warscape engines support?

[stevehoos]

Just a question for those that have R2, is the game engine now optimized for quad cores to benefit performance?

And how many CPU cores does the older Warscape engines support?

The games runs like poo.

[alQamar]

Hi falcon this is strictly a matter of view.
I found out that Rome 2 does utilize an AMD octa core significantly better than Shogun 2 did.
Also hyperthreadimg on Intel CPUs is helpful if you have less than 3 cores.
Nevertheless the performance is not scalable enough. There is no significant difference to have a 8 core CPU or a quad core with or without hyperthreading.

I've posted detailed reports for a quad core in the tech support forum.the thread is named "puzzle Rome performance" or sorts.

[Emperor Domitianus]

Despite the patch I have noticed no difference:

my rig is :

i7 4.5 Ghz
GTX 780 Ti
16Gb ram 2133

I assume that everything is tied to the hw. Sincerely I think that they have messed up something trying to improve performace so that lowend pc could set hgher settings in the game.

[alQamar]

Strange domi there should be slight positive differences to see between 9.1 and 10 beta.

Please refer the table of Prussian prince and my data in the op

[Emperor Domitianus]

Strange domi there should be slight positive differences to see between 9.1 and 10 beta.

Please refer the table of Prussian prince and my data in the op

They have solved some strange behaiour with floating buildings and units skins problems with large battles but I haven't seen much difference in performance. I haven't run any benchmark and I refer to the real experience playing.
This may be because I have double your ram and a 780 Ti which is higher than your 670.
If you can point me what to look at I will have a closer look at that.

[stevehoos]

They have solved some strange behaiour with floating buildings and units skins problems with large battles but I haven't seen much difference in performance. I haven't run any benchmark and I refer to the real experience playing.
This may be because I have double your ram and a 780 Ti which is higher than your 670.
If you can point me what to look at I will have a closer look at that.

Carthage still has floating buildings, so I guess they missed that one.

[alQamar]

I only rely on the benchmark as we have no other reproduceable scene yet. Given the data of other 780ti or Titans more vram than 1.9 will not be allocated on a 1080p resolution and max settings

[alQamar]

Hi guys,
It is hard to admit the new results. I summarize them in the OP.


As I started this thread once on totalwar.com and a copy here, I was upset that people with probably lower hardware consequently being forced to run Rome 2 on lower settings (extreme to high preset) got visible performance upgrades (up to 10%) but people running the game on extreme settings or even maxed out possible settings suffered decreases of performance. I finally stopped my silence about that with patch 8.1 when CA ignored all warnings completely that there are significant peformance issues brought by a small hotfix (8.1) on patch 8.


Now and today we could be happy patch 10 arrives. The beta already gave us an idea that CA somehow managed to achieve 3 things

- they lowered the quality visibly (LOD issues)
- managed to improve the loosy performance on maxed out and extreme settings, so much that is was at least the same as we had on release
- degrading previous performance gains on lower settings

Now then Patch 10 finally arrived, they turned the table.

Somehow they managed to kill all progress on lowered graphical presets. That means, all progress this part of users had till patch 8.1 now melted away. Compared to patch 8.1 and patch 10 on
high preset people with similar hardware (which is not the worst at all) will suffer 17% performance loss! This is same inacceptable as it was before for people on extreme and higher settings.

I have even found out that the LOD issues are so much present now, I barely can stand it to watch the benchmark running even on full settings! LOD issues are everywhere.

When I ran the benchmark on HIGH preset with patch 10, I do not have even a bit lowered performance compared to patch 9.1 final, but the most itching thing is the now visible downgrade of quality like mentioned in the linked thread about LOD issues by LuciferHawk.


Fans of Total War I am left completely speechless.

The ignorance CA handles this topic is outstanding, and clearly it is really hard to stay calm. Some people many not understand the issue, as people on facebook already blamed me for raging about plus / minus 1 FPS, but is the core issue that is hurtening.

We read about performance increases in the patch notes, that cannot be reproduced in an objective situation the benchmark offers. Despite we can see the different happening and that is what makes me quite dissappointed, frustrated and angry.

I do not know if you care about this particular topic as I do, as long the game seems playable for you and you close your eyes ignoring the LOD issues and sorts, but for me this is a betrayal, a matter of false information again, yet CA will stand above this all, commenting anything on these matters.

I also really do not know how they dare to arse their people, buying their DLCs and their game this way.

I additionally to all do not have any personal dispute with CA, just frankly reflecting figures here and I cannot find words soon on what is happening on the minds over at Horsham.
If they not even fix the performance issues honestly, when then the downgrading. How does it make technically sense to downgrade the quality and not even gaining more FPS in this case I mentioned above?

It's becoming all senseless.

some comparisons
high preset:
patch 9 avg performance 157 fps
patch 10 avg performance 132 fps


very high preset:
patch 8.1 avg performance 88,8 fps
patch 10 avg performance 79,9 fps


ultra preset
patch 8.1 beta avg performance 67,5 fps
patch 10 avg performance 63,1 fps


Shameful display.

[BigEl_nobody]

Ok here is my reply:

I was really into the performance thing as well from the very start of the release and one of the reasons why this game is long dead for me now is the fact that i had my best Performance of this game with Patch 2.... and that is to date...

Not only didn´t they increased the performance from Patch 2 on to later ones, it went below and added weird half-frozen mini spikes so to say that would especially occour at the beginning of the benchmark.

So, yes, I completely see the same results as you alQamar. This game is a coded mess, completely unlike the very polished performance of Shogun 2, that only broke until extreme circumstances for me.

[Epic28]

You are seriously complaining about 60fps on ultra? The most you need is 30.

On a PS3 from 2005 sure. On a top of the line PC in 2014? 60 FPS is the norm.

-Edit

But yes, you missed the point. CA releases a patch that says performance increases, and yet some folks lose FPS during our benchmarks? Well that is an issue.

Although its become fairly evident CA's optimization for the game is sadly geared towards people with low tier PCs.

[alQamar]

Hi lads, thanks for appreciating my work on TWC. I am happy to help you where I can. I played a night battle yesterday (screenshots, partly with GEM enabled) and at least on maxed out settings it is far from being playable as I had FPS down to 6 fps when zooming in forests. I am quite aware CA told me more than one time why vegetation alpha is draining the hardware so much and my system is quite to weak for that. Even a 780TI and Titan won't handle it very smoothly. However I just want to inform you that the performance in battles is quite similar than it was in patch 9. LuciferHawk PM me that he noticed a performance gain in a certain condition. If the game lags the input lag is gone, so it seems the UI and mouse are now independent from the overall battle speed. I think I can confirm this, as the commanding of unit was never restricted or laggy even on low FPS. So at least some certain performance issues appear now to be solved.

[alQamar]

Hello everyone, I found some time to run the benchmarks again and to update the figures in the OP for anyone interested. The hotfix that came unexpected today did not fix any issues in the benchmark, such as LoD or other things still present there I mentioned before, but the peformance is a little bit better now compared to patch 10 final again.
Seen from the point of view we had in September we are still sub par compared to the performance and quality we had. It's hard to say if CA can manage real performance increases for this game e.g. 10-20% while positively adjusting the quality issues.

[tomFoolery]

Yeah, I undoubtedly had much, much better video-audio performance in the first month or two after release than I do now after CA's patches that "fixed" things.

[alQamar]

AMD or NVIDIA Brig?

[Reggum]

Does any ones game crash when they change there graphic settings? (mine does)

[Reggum]

Does any ones game crash when they change there graphic settings? (mine does)

Don't no how upload pics, but I got 32.4fps on the benchmark. Max Settings.

Running i5 2500k 4.2ghz cpu
sli 560ti 2gb
8gb ram.