Promised performance increases observed - with Patch 16.1 - updated 05.02.2015

[Leonidas II]

Wow guys , I have tested Patch 9beta and I am very surprised.I was playing at very high , now I have set my settings to extreme (except for unit detail : high , water :middle , shadows : middle )
Campaign map now runs absolutely fluid , before the patch it was droping to 30 FPS when selecting a unit or a city.
The battles have also improved very much . Just played a 40 vs 40 battle.Before the patch FPS began to drop like crazy , especially when I zoomed into a battle FPS dropped to 25 .
Now everything runs fluid , I can zoom in and it's running grate ! Even when units clash it's running well.And all this in a 40vs40 Battle and with even higher settings than I was using before the patch !

I have to admit that CA has done a good job . Seems like they have solved the problem and are improving performance now.

Edit : My results are including the update for patch9beta that came out today

[alQamar]

I would like to remind you all, I've set up this thread for discussion about performance degrees of Rome 2 in consideration with the patches. Just like Leo's posting #118, not for tech discussion, as such a thread exists in the tech section.
Thank you for your understanding.

Could I bother you Leo to run the benchmark with the 8.1 state and and patch 9 beta state on same settings, probably the higher ones you can use now as said? Thank you. I really need some direct comparisons and screenshots, same to Baleur please. I know you state the benchmark is not reflecting the game but it is the only reliable comparison we have.

[Leonidas II]

Did you even read the thread you linked? The guys is having MAJOR problems in getting several mods and settings to work simply BECAUSE he has gone over the memory limit.
Come on bro you make it to easy when you mess up like that

Please discuss it per private message guys , seriously.

[johnny_b_good]

Heard that today's hotfix increase fps ?! Anyone?

[Leonidas II]

Here are my test results.

System:
Intel Core I7-860 2.8 GHz ; Turbo : 3.6 GHz
AMD Radeon HD 7970 GHz;3GB
8GB RAM


Used settings :
Anti-Aliasing - on
Texture Quality - Ultra
Texture Filtering - Anistrophic 8x
Shadows - Extreme
Water - Middle
Sky - Extreme
Shader Model - 5
Particle Effects - Extreme
Vegetation Alpha - off
Unit size - ultra
Building Details - Extreme
Unit Details - High
Trees - Extreme
Grass - Extreme
Terain - Extreme
Depth of Field - High


Before Patch 9Beta :



Patch 9Beta :



That's the reason why I don't trust the benchmark ...

[A Barbarian]

Here are my test results.
That's the reason why I don't trust the benchmark ...

1) U r right, the optimizations in patch 9 are primarily about battles and the benchmark isn't a very good test for that
2) But u have reduced CPU load by using 'unit details high' instead of extreme. That's exactly the part which has been optimized in patch 9. If u would try extreme for patch 8.1 and patch 9, then u should see some difference, even in the benchmark.

[alQamar]

Here are my test results.

System:
Intel Core I7-860 2.8 GHz ; Turbo : 3.6 GHz
AMD Radeon HD 7970 GHz;3GB
8GB RAM


Used settings :
Anti-Aliasing - on
Texture Quality - Ultra
Texture Filtering - Anistrophic 8x
Shadows - Extreme
Water - Middle
Sky - Extreme
Shader Model - 5
Particle Effects - Extreme
Vegetation Alpha - off
Unit size - ultra
Building Details - Extreme
Unit Details - High
Trees - Extreme
Grass - Extreme
Terain - Extreme
Depth of Field - High


Before Patch 9Beta :



Patch 9Beta :



That's the reason why I don't trust the benchmark ...

but it is quite interesting that you experience a decrease of performance with patch 9, same as I do with 670GTX. But it is not that massive than I experience it.

[alQamar]

I just dl the latest 9 beta hotfix and on maxed out settings the performance decreased again.

release (3rd sept)
33 fps
patch 8.1 final
31.8
patch 9 beta
29.2
patch 9 beta hotfix
28.6


Shogon can you really play Rome 2 visibly assigned on 3 screens the same time? I am just wondering because I really get higher framerates on the preset you did your screenshots on, but a 1080p resolution.

[shogon]

60 Hz VSYNC...the devils work.

I haven't tried patch 9 because DeI isn't fully updated to it, and plus I want to win as Carthage before I update. But, I do get around 70+ frames with a single Titan playing 3240x1920 on 3 monitors @ 100 Hz Lightboost, with certain settings I can get the vram up to 3.4 GBs, vegetation alpha adds like 1 GB of vram. Game textures look like crap in some cases, some of the time the landscape just vanishes to this badly noticeable patch of out of focused land, happens in campaign and battle maps when I zoom in.

But that's what happens when you code the game for hardware based on technology years ago, and possibly laptops, they (CA) probably don't even know what GK110 is, let alone SLI.

[Leonidas II]

vegetation alpha adds like 1 GB of vram

What ?

[shogon]

What ?

This is Vegetation Alpha.


Here is the benchmark with it on.



Here it is with Vegetation Alpha off.



3240x1920 @ 100 Hz with 3 Asus VG248QE's in Surround
Intel 3970x @ 4.9
EVGA GTX Titan 1097/6500

Also, it not quite 1GB of VRAM difference between on and off for me, but it's close.

[Leonidas II]

Ok , as long as 3GB Vram are enough , I am happy

I simply forgot that you're playing in much higher resolution than Full HD , so a VRam usage over 3GB is not surprising on your computer.

[longvehicle]

Hey guys.I know its little bit off topic but,does Shogun 2 has the same CPU bottleneck when two army clash? I think Im going to buy Shogun 2 rather than Rome 2.

[shogon]

Shogon can you really play Rome 2 visibly assigned on 3 screens the same time? I am just wondering because I really get higher framerates on the preset you did your screenshots on, but a 1080p resolution.

Yup, I've been using 3 1080p (3240x1920 resolution) screens in portrait with Nvidia Surround. Luckily with the 600 series and later you can do such a thing with a single card. If I try using very high settings my fps goes down a lot so I'm stuck with high, and till DeI and other mods get updated to patch 9.

[alQamar]

@shogon To have a more comparable set, could I bug you to test the game on maxed out settings and extreme (probably with unlimited vram setting on) either with patch 8.1 and 9 beta hf#1? and repost the screenies?
that would be very helpful. As you seem to use msi afterburner could you please make a logfile for all those 4 benches?

I would need the following objects to be logged in the files:
Power, GPU usage, FB usage, BUS usage, core clock, memory usage, power limit, utilisation limit, framerate, cpu(x) usage, RAM usage, pagefile usage.

I asked Prussianprince to do a basic comparison aswell, as he reported to have improvements in performance, which is a quite opposite behaviour to my findings. I am really interested how much different hardware plays a role here.
The logfiles in your case are very helpful for me, as I really have no access to a 2+ GB VRAM card at the moment, so basically I would love to see how much Rome 2 will consume on your rig in a 1080p single monitor resolution.
Please disable mods and GEM (roll key) and antivirus (windows defender, ms security essentials) when running tests.


Thank you, 1000 times.

[shogon]

@alQamar,

Hopefully I did this right for you, I haven't tested it in my native resolution yet, but I shall in time.

Here are the settings (like I said, hopefully right)



Results with 8.1 patch,



Results with patch 9,



And the logs from MSI. 1.81 log.rar1.9 log.rar

If I messed up let me know haha, or if I have to try some different settings.

PC:

i7 3970x 4.9 GHz 1.46v HyperThreading disabled
EVGA GTX Titan 1097 MHz/7000 MHz 1.21v
Asus Rampage 4 Extreme
Samsung Eco Ram 2x4GB 2133 MHz 1.52v
Samsung 840 Pro 512GB SSD
Asus VG248QE @ 120 Hz

[alQamar]

Thanks shogon, I will look into the logfiles later. The settings were perfectly correct, except you missed to enable antialiasing, but that does not harm the result. It is interesting enough that also on your rig the performance decreases with patch 9 on extreme preset, as said in my OP, are reproducible. Fairly evident that CA is wrong about their assumptions this does not happen for all owners of capable hardware. Again I am waiting for results of other people too.

[shogon]

I have some more stuff for you to rummage through when you are able to, mainly I did 1.9 with AA at 1080p, my usual settings at 3240x1920 with 1.81 and 1.9, and then 3420x1920 maxed (with AA this time) with both 1.81 and 1.9.

So here we go,

Here is patch 1.9 with AA this time at 1920x1080p maxed



Usual settings at 3240x1920 with 1.81:

Results:

Now with 1.9:

Here are the results for 3240x1920 maxed out with AA enabled patch 1.81:

Now with patch 1.9:

Here are the log files for each run as well.

1.9AA.rar181_1920p_high.rar1.9[1902p_high].rar1.81_1920p_maxed.rar1.9[1920p] maxed.rar

Also here is the updated spec list for these benches, mainly the refresh rate to 100 Hz Lightboost (which shouldn't effect the scores at all) from 120 Hz Lightboost, and increase the monitor amount to 3 and having them in portrait Surround from Nvidia via 2 DVI + Display Port connections.
i7 3970x 4.9 GHz 1.46v HyperThreading disabled
EVGA GTX Titan 1097 MHz/7000 MHz 1.21v
Asus Rampage 4 Extreme
Samsung Eco Ram 2x4GB 2133 MHz 1.52v
Samsung 840 Pro 512GB SSD
3 Asus VG248QE (3240x1920) @ 100 Hz Portrait Surround

[alQamar]

Guys, I am getting crazy right now.

Patch 9 final is even slower than patch 9 beta. The performance decrease of Patch 9 final is now at -11% compared to Patch 8.1 final!!! WTF
Compared to release 3rd Sept. the lost performance is about -14,24%!

I am not saying there are no performance increases in particular scenes but at least in the benchmark they are not measureable but turning out opposite!
CA, stop confusing people with misleading patch logs or finally use your own ingame benchmarking product, to estaminate performance changes, please.

A big thank you towards Shogon his helpful submissions. I will hopefully find time to browse through the logfiles and analyze them in Excel.
To avoid getting this thread too technical (it has been created to check the bit of truth of CA performance increase advertisements) I will post them in the tech section


PATCH 9 Final, maxed out settings, no mods, 1080p, 670 GTX FTW, i7-2600 @ 4,6 GHz




PATCH 8.1 Final, maxed out settings, no mods, 1080p, 670 GTX FTW, i7-2600 @ 4,6 GHz

[Pesticide]

i have a 920 @ 3500 mhr and a 670 gtx, ill run some final beta benches too, but ye i gave up on the game since launch