Login

russian armor

CoH2 GPU/CPU usage

19 Jan 2016, 06:06 AM
#21
avatar of Cyanara

Posts: 769 | Subs: 1

I've also been noticing this same issue on my computer. Other games max my video card, but COH2 just gets slower and slower despite not even coming close to maxing out my CPU, RAM, or GPU. Even a fresh install of Windows made no difference.

The stuff about thread scheduling is interesting. I'll have to look into that more since my CPU is an old one with only a clock speed of 3.2Ghz.
19 Jan 2016, 15:07 PM
#22
avatar of kamk
Donator 11

Posts: 764

Well, achieving the 100%'ish isn't a problem with COH2... temporarily.

The benchmark:

Yay, 100%'ish load.

But an actual game:

23 minute game, Rzhev Winter, 3v3, medium settings with AA & vsync off
Started out with fluent 80-100 FPS, went to crap a little later. Probably would go down further over time.


30 minute game, City 17, 3v3, low settings with AA & vsync off
Same again, decent load at start, all fluent, and then drops FPS & load after a while.
Funny detail: GPU clocks down due boredom.



jump backJump back to quoted post19 Jan 2016, 01:02 AMSeb
...
My conclusion from this is and the fact that my load was not at 100% in the 1080p test, is that my GTX 970 is not the limiting factor in 1080p....
I can then assume the limiting factor for me is the performance of my CPU (now rather old i7 3770k even with a 130% OC), and more precisely its limit on a single core (because the game is not well enough multi-thread optimized).

Most likely. My CPU seems to be around 30% load all game long.

But let's be fair here for a second:
http://www.cpu-world.com/benchmarks/desktop_CPUs_single.html
https://www.cpubenchmark.net/singleThread.html
... and so on.
Your CPU in this regard is still pretty much as good as they get. My little Xeon doesn't have to hide either. So what kind of CPU setup do we have to own to keep up?
hmm... someone got a heavily OC'd nitrogen cooled dual core around? :snfPeter:



Nevertheless, the performance degradation is rather severe in this game, and sadly unavoidable atm.
- my 2 cents


Apropo, to answer the original question from OP:
you want more GPU load? Ramp up ze graphics. Will at least give you some load at start.
Seb
19 Jan 2016, 22:17 PM
#23
avatar of Seb
Admin Black Badge

Posts: 3709 | Subs: 2

Yes I think this is it really, mostly optimization issues. COH2 using way too much CPU for what it does, and not enough in different threads either.

As for the GPU usage decreasing over time, I would then assume the CPU (main thread) getting more and more taxed over time as there are units on the field, so that the GPU is even less needed in comparison.

That would make sense with the rest of the analysis.

And yes indeed, if you are in that case of low GPU usage, feel free to increase all the graphics related to GPU, especially AA.
Seb
20 Jan 2016, 00:45 AM
#24
avatar of Seb
Admin Black Badge

Posts: 3709 | Subs: 2

tl;dr no new conclusion (cpu => 1 core only, gpu mostly underused, smoke kills gpu)


More interesting graph with the awesome Process Explorer

Both graphs are in sync, over a spectate of 20min.



The colors on the CPU Usage are for each core. Pretty much only one core is used. The total never went above 25%.

This means my earlier comment about thread scheduling is somewhat false. Windows doesn't artificially make it look like the game is using more cores, it's just that it puts other running applications on the other cores. Or in case of i7, yeah it's actually putting each of the game 25% in 12.5% * 2 hyper-process (of the same core) or whatever the name. In practice the result is the same though and conclusion correct: only single core performance matters for this game at all, and having more than 2 cores is totally irrelevant except for streaming or other stuff next to it.


RAM 1.8 GB constant, the max for a 32 bit app... Yeah this is getting old.


IO at the start (loading) and end (saving replay). Not much to say, good that there is no IO during the game (having to load stuff from the disk would be pretty bad). I guess it's storing the additional data somewhere else in RAM and swapping it at runtime? I'm not really familiar with the mechanic to pre/reload more stuff when the 32bit limit is reached.




The 2 gaps to 0% in the early are when I paused the game (through the app to freeze the process).

GPU load typically 100% at the start and then gets lower. CPU can't get higher at the same time since it's already maxed on 1 core, but that means fps gets lower, and particularly goes below the critical threshold of 60fps, which is really where optimization should be aimed at.

Well, it goes below 60gps for my CPU that is. And that means the most recent such as 6700K (or older 4790K, how is it so good??), whilst not that much better, would maybe allow to keep it above the critical 60fps mark all the time, which would change everything, in regard to fluidity, and more importantly when considering streaming at a true 60fps which is currently impossible for this game otherwise. That said since I'm already at 4.35 GHz so this is not even sure.


What's interesting is the VRAM collapse from 1.9 GB at the start to then averaging 1.0 GB for the rest of the game. wtf?? This doesn't seem to match what GPU-Z reported for VRAM earlier though and will need more testing.


System GPU memory is always like 60MB whatever that changes.


Last note, multiple times the GPU load went back to 100%, it was when there was smoke (and sometimes I zoomed it on it). It killed my fps, fully loaded GPU, and significantly reduced CPU (we can see those better on the CPU graph), GPU being the main limit in this rare case.

Basically that is really bad, because it means I'd have to decrease AA to keep a constant framerate for the rare occurrence of smoke, when it's already not using my GPU the rest of the time. There really should be a separate setting to reduce only the smoke details.


Finally, running threads during the game :





The number of them doesn't matter so long as they depend on each other, or are dedicated to (too) small tasks. Total around 24% CPU at that time.
20 Jan 2016, 02:04 AM
#25
avatar of Cyanara

Posts: 769 | Subs: 1

Thanks for making me aware of this CPU stuff, Seb. It helps to clarify a lot.

I guess overclocking is the best way to go if someone has a lower-clocked CPU like myself. That said, it's fairly inexplicable that the game would have such increasing CPU needs as the game goes on. I can't say I know of any other games that have that issue, but I'll have to keep an eye out now. Definitely seems to be an optimisation issue.

In regards to measuring GPU usage, it's worth noting that they tend to have 2 engines: 3D and 2D. Therefore some tools such as Process Explorer will often show 50% as max usage in 3D games because that's all the game uses.

The reason it will occasionally go over 50% to 100% in COH2 is that particle effects seem to be mostly or entirely rendered by the 2D engine. So 50% 3D engine + 50% 2D engine = 100% usage.

However, I think the 2D engine is probably not meant to handle particularly intensive stuff. It's possibly oriented more at basic applications? Whatever the reason, the 2D engine on my GTX770 can't handle particle effects like smoke and fire in this game.
20 Jan 2016, 04:01 AM
#26
avatar of Cyanara

Posts: 769 | Subs: 1

Hmm. So I was trying to find out if there's any way to view how many active threads a particular process has. I didn't find anything, but I did find that Process Explorer lets me set CPU Affinity (which cores it is allowed to run on).

Removing 1 out 4 cores had a small or no visible impact on FPS.
Removing 2 out of 4 cores had significant impact.
Removing 3 out of 4 cores had a huge impact.

Individual core usage noticeably increases as you reduce the number of available cores, but still didn't get near 100%.

I also noticed that RelicCoH2.exe had a very low CPU priority by default, but raising it had no apparent impact. It might help to reduce lag moments though.

CPU Core Distribution


In lobby:


In loading screen:


In-game:


While I can't get over that strange 30% CPU usage barrier, the core distribution definitely changes enough to suggest that the game is technically multi-threaded. I don't really have enough knowledge with this stuff to make any better guesses though. I feel like some input from Relic would go a long way here.

20 Jan 2016, 10:39 AM
#27
avatar of kamk
Donator 11

Posts: 764

jump backJump back to quoted post20 Jan 2016, 00:45 AMSeb
...
RAM 1.8 GB constant, the max for a 32 bit app... Yeah this is getting old.
...
Well, it goes below 60gps for my CPU that is. And that means the most recent such as 6700K (or older 4790K, how is it so good??), whilst not that much better, would maybe allow to keep it above the critical 60fps mark all the time, which would change everything, in regard to fluidity, and more importantly when considering streaming at a true 60fps which is currently impossible for this game otherwise. That said since I'm already at 4.35 GHz so this is not even sure.
...
Basically that is really bad, because it means I'd have to decrease AA

I'm pretty sure COH2 has large adress awareness, and uses more.
Just for the benchmark i've seen a 2,4 GB baseline, and another 0,9 GB while benching, ~3,3 GB overall.

As i linked earlier: your CPU is pretty much as good as they get, single-threaded wise. There hasn't been any improvement for ages, and without a miracle there won't be any.

COH2s AA is quite bad, just force FXAA in your nvidia control panel and deactivate COH2s.


Anyways, some core scaling benchmarks:
FPS average from the COH2 benchmark. No special treatment.











































cores FPS avg. priority
1 25,56 high
2 36,48 high
3 36,56 high
4 36,98 high
4+4 37,01 high
4+4 36,92 normal
4+4 37,05 real

The "no special treatment" shows a little. (Edit: background tasks kept running, therefore interfering a little with the low core testing, as well as showing some difference in priority). Doesn't matter either way.
It basically caps out at 1,x cores, and isn't properly multi-threaded. So no news there either.


Back to square one:
...someone got a heavily OC'd nitrogen cooled dual core around? :snfPeter:
21 Jan 2016, 00:09 AM
#28
avatar of ausownage

Posts: 117

CoH2 is a resource hog because it is poorly optimized.

Relic have made little effort to increase CoH optimization in recent... years.
21 Jan 2016, 11:48 AM
#29
avatar of Cyanara

Posts: 769 | Subs: 1

For what it's worth, I figured I may as well have a go at disabling replays to see if that might have anything to do with it. Still seemed to be just as bad though.

The confusing part is why FPS can jump by about 20 when I look an empty part of the map if the GPU isn't the bottleneck.
Seb
21 Jan 2016, 15:07 PM
#30
avatar of Seb
Admin Black Badge

Posts: 3709 | Subs: 2

(educated guess) The CPU is calculating most of the animations. It doesn't need to do it for units that are not on the screen.
21 Jan 2016, 15:47 PM
#31
avatar of The amazing Chandler

Posts: 1355

CoH2 is a resource hog because it is poorly optimized.

Relic have made little effort to increase CoH optimization in recent... years.


Little to almost nothing i would say.
The community is saying this since release but Relic gives a sh... Instead they are giving us new overpriced and OP commanders (in one word "P2W"), yupiiiiii!
22 Jan 2016, 12:48 PM
#32
avatar of Cyanara

Posts: 769 | Subs: 1

I overlocked my CPU from 3.2Ghz to 3.8Ghz and played some 4v4. I started with about 110fps, and ended up with 50-60fps well into the game. Still not great to be losing over half my starting fps, but at least it was much more responsive than it usually is.

Edit:

I created a benchmark out of a 4v4 game on Steppes and used FRAPS from 2 minutes in, with 'follow camera' enabled on myself.

Benchmarks:


Steppes 4v4 700sec 3.2Ghz No physics

Frames Time (ms) Min Max Avg
36877 700000 31 77 52.681


Steppes 4v4 700sec 3.8Ghz No physics

Frames Time (ms) Min Max Avg
42447 700000 36 87 60.639


Steppes 4v4 700sec 3.2Ghz Max physics

Frames Time (ms) Min Max Avg
33391 700000 22 70 47.701


tldr;

Turning off physics = good
Overclocking CPU = even better
No magic bullets though.
23 Jan 2016, 05:48 AM
#33
avatar of ausownage

Posts: 117

Wow, you experienced a jump in FPS on that scale?

impressive. I am running 2500k @ 3.6ghz or so but I can go up to 4.5ghz. I'll have to give a try. Cheers.
23 Jan 2016, 07:03 AM
#34
avatar of Cyanara

Posts: 769 | Subs: 1

Did my own benchmarks for CPU affinity changes to RelicCoH2.exe:

Steppes 4v4 Replay - 1 Core enabled


Frames Time (ms) Min Max Avg
6735 200000 22 47 33.675


Steppes 4v4 Replay - 2 Cores enabled



Frames Time (ms) Min Max Avg
9307 200000 32 64 46.535


Steppes 4v4 Replay - 3 Cores enabled



Frames Time (ms) Min Max Avg
10843 200000 38 71 54.215


Steppes 4v4 Replay - 4 Cores enabled




Frames Time (ms) Min Max Avg
11777 200000 42 78 58.885
23 Jan 2016, 12:25 PM
#35
avatar of kamk
Donator 11

Posts: 764

Which CPU are you using? Those jumps after 2 cores are a little unexpected. Fraps interfering?


24 Jan 2016, 00:46 AM
#36
avatar of Cyanara

Posts: 769 | Subs: 1

jump backJump back to quoted post23 Jan 2016, 12:25 PMkamk
Which CPU are you using? Those jumps after 2 cores are a little unexpected. Fraps interfering?





Core i7 960
24 Jan 2016, 19:03 PM
#37
avatar of iignition

Posts: 22

I'm seeing around 40-50% CPU and around 30-40% GPU usage throughout a 1v1 with a steady 60 fps. I'm running a overclocked i5-3570k and GTX 970 on high settings.
25 Jan 2016, 02:53 AM
#38
avatar of Cyanara

Posts: 769 | Subs: 1

I'm seeing around 40-50% CPU and around 30-40% GPU usage throughout a 1v1 with a steady 60 fps. I'm running a overclocked i5-3570k and GTX 970 on high settings.


Yeah, 1v1 will play much better than the larger game modes. It seems to have a lot to do with how many units are in the game. So they all start off well, but 4v4 loses fps really fast.

If you're getting a steady 60fps, that would likely mean you have v-sync on. You can turn that off if you want to see how your total fps changes throughout the game.
27 Jan 2016, 09:50 AM
#39
avatar of Trubbbel

Posts: 721

What would you say is a great system for playing Coh2 then? I currently run

i5-3570k
GTX 660ti 2G
8GB RAM

I usually play 3v3 and 4v4 games and sometimes (as of last few months I feel) the game slows down and then suddenly all units accelerate lots. It doesn't happen too often but it happens, and sometimes it sort of lags a bit. I've done the ingame test for settings and put them as high as I could wih resonable performance because I want it to look good.

Are there any cost-effective upgrades that will bring significant improvement to my rig when playing CoH2? I did check on that single thread benchmark link and the i5-3570k seems to hold up really well so that doesn't seem to warrant a change. Does the GPU? Will it be Worth to go with GTX 960 Strix 4GB? I found a reasonable price for that card just today.
27 Jan 2016, 10:34 AM
#40
avatar of Cyanara

Posts: 769 | Subs: 1

What would you say is a great system for playing Coh2 then? I currently run
I usually play 3v3 and 4v4 games and sometimes (as of last few months I feel) the game slows down and then suddenly all units accelerate lots.


That is usually an Internet connectivity issue between you and the server. It loses connection briefly, and when it reconnects it has to play out all the 'steps' of the game to catch up with everyone else.

Obviously there's not much you can do about the line beyond your modem (besides call your ISP if you can prove connectivity issues), but you can update your modem's firmware if it's out of date. If you're playing over a wireless connection, there is also a lot of potential for interference and drop outs there. Check Google for guides on making sure your wireless signal isn't competing with your neighbours. You can also try updating any drivers for your wireless devices, and attempt to rule out wireless issues altogether by playing over a direct Ethernet cable for a while, if possible.

The issues discussed in this thread are more about a consistent decline in frames per second as a match progresses.


Are there any cost-effective upgrades that will bring significant improvement to my rig when playing CoH2? I did check on that single thread benchmark link and the i5-3570k seems to hold up really well so that doesn't seem to warrant a change. Does the GPU? Will it be Worth to go with GTX 960 Strix 4GB? I found a reasonable price for that card just today.


Based on the findings of this thread, you will almost certainly be limited by your CPU, and there doesn't seem to be anything any of us can do about it. There just aren't any CPUs out there that are as powerful as this game seems to require.

That said, more data is always good. Feel free to download a copy of either Process Explorer or GPU-Z and take a screenshot of your GPU's activity as you play, and copy it here for us to look at. If it is practically a flat line at 50% or 100%, then your GPU will be a limiting factor. If not, it's probably your CPU.

The one thing you will be limited by on your GPU is the VRAM. 2GB is definitely not enough to run maximum textures smoothly in 4v4, but it's hard to say if it really matters with the whole CPU issue already slowing things down. At any rate, Process Explorer will help you work out how much VRAM you are using at each texture setting.

Your RAM should be fine, as long as you're not really running anything else in the background (browsers are surprisingly bad).

1 user is browsing this thread: 1 guest

Ladders Top 10

  • #
    Steam Alias
    W
    L
    %
    Streak
Data provided by Relic Relic Entertainment

Replay highlight

VS
  • U.S. Forces flag cblanco ★
  • The British Forces flag 보드카 중대
  • Oberkommando West flag VonManteuffel
  • Ostheer flag Heartless Jäger
uploaded by XXxxHeartlessxxXX

Board Info

406 users are online: 406 guests
3 posts in the last 24h
4 posts in the last week
23 posts in the last month
Registered members: 48722
Welcome our newest member, asherllc
Most online: 2043 users on 29 Oct 2023, 01:04 AM