Game Sluggish (Outside Areas)

[Palandus]

I'm curious Dr. Disaster do you have poor framerates in Outside Areas?

If you don't what kind of Rig do you have, specs wise?

[badhabit]

this discussion really made me ROFL please more arguing guys

Problem is, dr.disaster don't understand the difference between a program which spawns some threads and a real parallel program written with multicore in mind and therefore also utilizes all ressources in a cpu bound situation (as analyzed here). *sigh* or he is just trolling I'm not sure.

[Dr.Disaster]

I'm curious Dr. Disaster do you have poor framerates in Outside Areas?

If you don't what kind of Rig do you have, specs wise?

I got an i7-950 with an GTX 760. LoG2 runs in 1920x1080 with everything maxed. FPS are in general 60, running with vsync+tripplebuffer set via the nvidia driver and inside the game options.

Depending on the outside area i'm in and the direction i'm looking at FPS in LoG2 can drop down to around 40 which is actually very hard to notice without running Fraps or "debugInfo = true". Prime example of a drop spot is the entrance to Forgotten River from Twigroot Forest:

SpoilerShow

The utilization you can see in task manager is the game and ONLY(!) the game; nothing else! Most of the time i can't tell why FPS do drop at all because neither CPU nor GPU are maxed out. Now hitting ESC in such situations to pause the game causes my GPU to max out and FPS to go up for whatever is currently on screen. For example with the screen above hitting ESC gives me 50 FPS. It seems as if the game is deliberately trottling the GPU like it's trying to save resources.

this discussion really made me ROFL please more arguing guys

Problem is ...

Show will go on Drakkan

@badhabit: if you desire to utilizing all resources of your CPU a.k.a. max out all your cores .. go grab CPU Killer! Along with your definition it's a fine real parallel program.

Your theory crafting is all nice and dandy yet it leads nowhere. In your world hardware does not matter while in RL it does. We've just seen it again right here in the current threads. It did not matter to you what the problem or the hardware situation was, your sole solution to everything is: "use the ingame framerate limiter". Well up to now this "mantra" of yours did not help anyone who came here since it totally ignores the actual cause of the problem at hand.

[badhabit]

I'm curious Dr. Disaster do you have poor framerates in Outside Areas?

If you don't what kind of Rig do you have, specs wise?

I got an i7-950 with an GTX 760. LoG2 runs in 1920x1080 with everything maxed. FPS are in general 60, running with vsync+tripplebuffer set via the nvidia driver and inside the game options.

Depending on the outside area i'm in and the direction i'm looking at FPS in LoG2 can drop down to around 40 which is actually very hard to notice without running Fraps or "debugInfo = true". Prime example of a drop spot is the entrance to Forgotten River from Twigroot Forest:

SpoilerShow

The utilization you can see in task manager is the game and ONLY(!) the game; nothing else! Most of the time i can't tell why FPS do drop at all because neither CPU nor GPU are maxed out. Now hitting ESC in such situations to pause the game causes my GPU to max out and FPS to go up for whatever is currently on screen. For example with the screen above hitting ESC gives me 50 FPS. It seems as if the game is deliberately trottling the GPU like it's trying to save resources.

this discussion really made me ROFL please more arguing guys

Problem is ...

Show will go on Drakkan

@badhabit: if you desire to utilizing all resources of your CPU a.k.a. max out all your cores .. go grab CPU Killer! Along with your definition it's a fine real parallel program.

Your theory crafting is all nice and dandy yet it leads nowhere. In your world hardware does not matter while in RL it does. We've just seen it again right here in the current threads. It did not matter to you what the problem or the hardware situation was, your sole solution to everything is: "use the ingame framerate limiter". Well up to now this "mantra" of yours did not help anyone who came here since it totally ignores the actual cause of the problem at hand.

Thanks for this example: you reach not the aimed for 60fps, means you are limited performance wise by something. The GPU is not the limitation (as GPU-Z's GPU load is 87%) so it is the CPU. You have an CPU utilization of 21% which calculates for your 8core system to 1.68 cores. Means CPU bound while only using 1.68 cores of the available 4 physical( and 4 HT cores). Means, Log2 is NOT utilizing all CPU ressources at hand while being CPU bound. Means, not a proper multithreaded engine, the quad core recommendation of AH is very suspicious (or plainly wrong). (You CPU Killer argumentation is completely off topic)

[Dr.Disaster]

Thanks for this example: you reach not the aimed for 60fps, means you are limited performance wise by something. The GPU is not the limitation (as GPU-Z's GPU load is 87%) so it is the CPU ...

And right here you're already on the wrong track: none of my cores is maxed out; not even close to.

[badhabit]

Thanks for this example: you reach not the aimed for 60fps, means you are limited performance wise by something. The GPU is not the limitation (as GPU-Z's GPU load is 87%) so it is the CPU ...

And right here you're already on the wrong track: none of my cores is maxed out; not even close to.

No, you just don't understand what I'm talking about, but here is the simple variant:

Recipe:
Pin your Log2.exe exactly in this spot (entrance to Forgotten River from Twigroot Forest) to: 1, 2 (core 1+2), 3 (core 1,2,3), and 4 (core 1,2,3,4) CPU cores and report the associated frame rates. You can do this while Log2 is running. (And shut your browser down for this test)

If we will see a jump up in FPS by adding the third core my theory is wrong. If there is no jump you are wrong.

[Dr.Disaster]

Thanks for this example: you reach not the aimed for 60fps, means you are limited performance wise by something. The GPU is not the limitation (as GPU-Z's GPU load is 87%) so it is the CPU ...

And right here you're already on the wrong track: none of my cores is maxed out; not even close to.

No, you just don't understand what I'm talking about, but here is the simple variant:

Pin your Log2.exe exactly in this spot to: 1, 2 (core 1+2), 3 (core 1,2,3), and 4 (core 1,2,3,4) CPU cores and report the associated frame rates. You can do this while Log2 is running. (And shut your browser down for this test)
If we will see a jump up in FPS by adding the third core my theory is wrong. If there is no jump you are wrong.

sigh .. i already did that over a week ago

[badhabit]

Thanks for this example: you reach not the aimed for 60fps, means you are limited performance wise by something. The GPU is not the limitation (as GPU-Z's GPU load is 87%) so it is the CPU ...
And right here you're already on the wrong track: none of my cores is maxed out; not even close to.

No, you just don't understand what I'm talking about, but here is the simple variant:

Pin your Log2.exe exactly in this spot to: 1, 2 (core 1+2), 3 (core 1,2,3), and 4 (core 1,2,3,4) CPU cores and report the associated frame rates. You can do this while Log2 is running. (And shut your browser down for this test)
If we will see a jump up in FPS by adding the third core my theory is wrong. If there is no jump you are wrong.

sigh .. i already did that over a week ago

You did somethign else there (no FPS given), also in another spot (CPU bound there? no GPU load visible...) Do EXACTLY as I described it above.

(Or, if we take this statement from you seriously, log2 is not utilizing more than 2 cores.)

[Dr.Disaster]

You are able to read what i linked to so do it.

[badhabit]

You are able to read what i linked to so do it.

I did, and... now? Is this of any help? no, your measurement(s) is non-conclusive, you gave no GPU load and/or FPS. (you did a useful measurement here, FPS, CPU and GPU load seeable)

Of help is the test I recommended in the CPU bound spot you found.
But I guess you already did the test and you found the expected behaviour: no gain in FPS with more than 2 cores despite not achieving 60 fps. Conclusion: Engine is not multithreaded and the quad core recommendation of AH is wrong.