Login

russian armor

CoH2 GPU/CPU usage

27 Jan 2016, 10:44 AM
#41
avatar of Trubbbel

Posts: 721

jump backJump back to quoted post27 Jan 2016, 10:34 AMCyanara
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.

Ok and I guess to do that means playing in windowed mode? Which in itself is going to make performance worse?
27 Jan 2016, 10:46 AM
#42
avatar of Cyanara

Posts: 769 | Subs: 1


Ok and I guess to do that means playing in windowed mode? Which in itself is going to make performance worse?


Nah, just run it in the background and alt-enter at some point to switch to it and then do a screenshot. It shows the last few minutes of performance on a graph.
27 Jan 2016, 13:08 PM
#43
avatar of Plaguer

Posts: 498


Ok and I guess to do that means playing in windowed mode? Which in itself is going to make performance worse?


Using more than one monitor is also possible, that's how I keep an eye on the performance atleast
27 Jan 2016, 17:01 PM
#44
avatar of kamk
Donator 11

Posts: 764


Ok and I guess to do that means playing in windowed mode? Which in itself is going to make performance worse?

Just let it run in the background, and set the refresh timer to 10 seconds (highest setting sadly).
Another fancy program to monitor temps and power consumption: http://www.aida64.com/downloads
(also set another refresh timer there - has a graph feature)

IMHO, GTX 960 is barely an improvement over a decent running (& clocked) 660 Ti.
27 Jan 2016, 18:02 PM
#45
avatar of Trubbbel

Posts: 721

Ok so here are the outputs from a 4v4 game

http://i.imgur.com/OGAfOwK.png?1




These processes normally run when I play. It's not much I Think
http://i.imgur.com/UY68ARQ.png

What do you think? It's working ok, I'm mostly looking to get max graphics out and wonder if it is any worth buying a new card or something.

Currently:
i5-3570k
GTX 660ti 2G
8GB RAM
27 Jan 2016, 18:12 PM
#46
avatar of kamk
Donator 11

Posts: 764

Ok so here are the outputs from a 4v4 game
-snip-
What do you think?

What are your graphic settings?
Results from the benchmark?
How's the load on the rest of your system? (process explorer)


Edit:
...
the game slows down and then suddenly all units accelerate lots.

My guess would be you got connectivity issues, or some background process causing problems.


Edit 2:
i think you should deactivate AA, and lower your Physics settings. Tamper around, i'd start with off.
27 Jan 2016, 18:33 PM
#47
avatar of Trubbbel

Posts: 721

jump backJump back to quoted post27 Jan 2016, 18:12 PMkamk

What are your graphic settings?
Results from the benchmark?
How's the load on the rest of your system? (process explorer)
Updated above

Edit 2:
i think you should deactivate AA, and lower your Physics settings. Tamper around, i'd start with off.

What does those do? Will it affect looks alot, details and such? Don't wanna sacrifice details if I don't have to.
27 Jan 2016, 19:41 PM
#48
avatar of kamk
Donator 11

Posts: 764

Just try it out.

If you want AA so badly, force FXAA in the nvidia driver.
28 Jan 2016, 00:41 AM
#49
avatar of Cyanara

Posts: 769 | Subs: 1

Ok so here are the outputs from a 4v4 game



GPU load is the issue here. If you look at it, you'll see that it's up and down all over the place. This means that it's not being fully used. If it was maxed out, it would be a fairly flat line at around the 50% mark.

At this point your CPU is almost certainly the limiting factor, and there's little you can do about that (except overclock it, if you have the know-how). You should definitely have room to increase your graphics settings, even though image quality and texture quality are the most important ones for making the game look nice and you've already got them fairly high. It's unlikely you'd notice higher textures on a 1080 monitor unless you zoomed in a lot.

Turning down/off physics will slightly improve CPU performance. In-game AA is demanding, but if your GPU has that much room to spare it probably doesn't matter.
28 Jan 2016, 01:10 AM
#50
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?


It's possible the in-built benchmark doesn't have the same CPU demands as the real game. I recommend trying it yourself with a large team game replay.
Seb
28 Jan 2016, 01:15 AM
#51
avatar of Seb
Admin Black Badge

Posts: 3709 | Subs: 2

The benchmark indeed doesn't have the same demand as a real game.

Unfortunately it is the only way to exactly reproduce a situation, so that performance impact can be compared from one test to another (after changing settings, etc).

The closest second would be to watch the same replay and auto-follow the same unit with the camera.

Now that I think about it, it might be possible for modders to create/script something so that we could have a better benchmark that is reproducible?
28 Jan 2016, 01:18 AM
#52
avatar of Cyanara

Posts: 769 | Subs: 1

jump backJump back to quoted post28 Jan 2016, 01:15 AMSeb

The closest second would be to watch the same replay and auto-follow the same unit with the camera.


Nah, just use the 'follow camera' option. That's what I did :) Copies what the player looks at.
Seb
28 Jan 2016, 01:19 AM
#53
avatar of Seb
Admin Black Badge

Posts: 3709 | Subs: 2

Oh I even forgot that existed now, good idea.
28 Jan 2016, 01:22 AM
#54
avatar of kamk
Donator 11

Posts: 764

Demos not broken anymore after patches?
28 Jan 2016, 01:29 AM
#55
avatar of Cyanara

Posts: 769 | Subs: 1

jump backJump back to quoted post28 Jan 2016, 01:22 AMkamk
Demos not broken anymore after patches?


Yeah, hard to benchmark changes in system demands from patch to patch. Still, that's a moot point as you can't go back to an old patch, even if you wanted to. We're only trying to establish the game's requirements as they currently stand. For that a replay does the job fine.
28 Jan 2016, 03:06 AM
#56
avatar of kamk
Donator 11

Posts: 764

jump backJump back to quoted post28 Jan 2016, 01:29 AMCyanara


Yeah, hard to benchmark changes in system demands from patch to patch. Still, that's a moot point as you can't go back to an old patch, even if you wanted to. We're only trying to establish the game's requirements at this point in time. For that a replay does the job fine.

Oh, don't get me wrong. The demo way is perfect to show performance degradation, and trying to find ways to avoid / postpone it.

For benchmarking, meaning comparison of raw performance with certain settings, personally i would suggest something that can be used more easily, especially in the future as well. The integrated test is more of a worst case scenario IMHO.

28 Jan 2016, 06:19 AM
#57
avatar of vietnamabc

Posts: 1063

Frankly, COH2 optimization is ultra bad, my computer runs Homeworld:Desert of Kharak on High even smoother than this game at Low setting.
28 Jan 2016, 07:43 AM
#58
avatar of Cyanara

Posts: 769 | Subs: 1

Looks like this is all old news: [url=http://forums.anandtech.com/showthread.php?t=2328902]http://forums.anandtech.com/showthread.php?t=2328902

I'm not a programmer, so I know little about those types of technologies, but people in that thread seem to point out some very poor choices by Relic in which technologies they used. It seems to be particularly bad for AMD CPUs, as COH2 is compiled with Intel C++ Compiler for optimisation with Intel products only.

They do seem to confirm that the game is multi-threaded, however.
28 Jan 2016, 09:39 AM
#59
avatar of Trubbbel

Posts: 721

jump backJump back to quoted post28 Jan 2016, 00:41 AMCyanara
At this point your CPU is almost certainly the limiting factor, and there's little you can do about that (except overclock it, if you have the know-how). You should definitely have room to increase your graphics settings, even though image quality and texture quality are the most important ones for making the game look nice and you've already got them fairly high. It's unlikely you'd notice higher textures on a 1080 monitor unless you zoomed in a lot.

Turning down/off physics will slightly improve CPU performance. In-game AA is demanding, but if your GPU has that much room to spare it probably doesn't matter.

jump backJump back to quoted post27 Jan 2016, 17:01 PMkamk
IMHO, GTX 960 is barely an improvement over a decent running (& clocked) 660 Ti.

Ok thanks for your input guys, I guess I'll just stick with the rig I have then with a GPU that doesn't max out, with a processor that ranks fairly highly for this type of game (single thread) and already using NVIDIA which seems to work better than AMD.

Overclocking and turning off physics then I guess if I want to try something, and maybe trying AA off. What does the setting physics do?
28 Jan 2016, 10:23 AM
#60
avatar of Plaguer

Posts: 498



Ok thanks for your input guys, I guess I'll just stick with the rig I have then with a GPU that doesn't max out, with a processor that ranks fairly highly for this type of game (single thread) and already using NVIDIA which seems to work better than AMD.

Overclocking and turning off physics then I guess if I want to try something, and maybe trying AA off. What does the setting physics do?


Some people have actually said that AMD works better than NVIDIA, I need to test this out as soon as I get to borrow my friends R9 390
1 user is browsing this thread: 1 guest

Livestreams

United Kingdom 277
Peru 25
New Zealand 5
unknown 3
United States 1

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

722 users are online: 722 guests
0 post in the last 24h
8 posts in the last week
34 posts in the last month
Registered members: 49107
Welcome our newest member, Falac851
Most online: 2043 users on 29 Oct 2023, 01:04 AM