Support NeoGAF

[Chance Hale]

Found fix for my audio issues. Game doesn't like 24bit 192000hz sound. Changing it to a bit lower fixes the scratchiness.

Everything was super slowed down and robotic for me, first time I've ever seen that. So weird!

 

[red731]

I feel lucky with my performance
i5 2500k @ 4,3
8Gb Ram
7970 (as crossfire is not working just a single one)

Running at 1440p with everything maxed except shadow map resolution, which takes away most fps for me, I am running at 40-60fps except on rooftops where you can see more of the city.

Whoa Feef. I was expecting that it will run like bad on my setup reading everything -

i7 3770k@4.4
16GB ram
7970 GHz 3GB

So you say that you are allright with the performace....hmmmmm
And at 1440, damn.

(btw why are people writing p behind resolution? like, you won't be running it at 1440i - everyone is proggressive hence the p.....and px are pixels - brain fart)

 

[jfoul]

1920x1080 with everything at max settings.

CPU: i7 4770K @ 4.4Ghz
GPU: EVGA GTX970 FTW+
MEM: 16GB

I've played for 6 hours without any issues.

 

[JambiBum]

I can't seem to play the multiplayer at all. Crashes after like five to ten minutes of playing. This happen to anyone else?

 

[derFeef]

Whoa Feef. I was expecting that it will run like bad on my setup reading everything -

i7 3770k@4.4
16GB ram
7970 GHz 3GB

So you say that you are allright with the performace....hmmmmm
And at 1440, damn.

(btw why are people writing p behind resolution? like, you won't be running it at 1440i - everyone is proggressive hence the p.....and px are pixels - brain fart)

You will do fine! Hopefully we can get some AMD drivers soon and it will run even better.

 

[red731]

You will do fine! Hopefully we can get some AMD drivers soon and it will run even better.

Hopefuly they will release another revision of Omega drivers soon - I still haven't installed them since I am waiting for new drivers that could drop next month (read that somewhere from someone on guru3d).

Also..digital asking price is too much, I would buy DL physical next month so all is good.

Thanks Feef. You have a jolly day

 

[Vuze]

lol, game doesn't take my input anymore...
Played just "fine" yesterday, now it only boots up in windowed mode and doesn't respond to keyboard input nor controller input...

Alt+Enter doesn't work either... Steam overlay works so it's the game. Tried rebooting the PC, game multiple times, Steam...

 

[KKRT00]

[

This just doesnt make sense. Unless game heavily uses AXV2 performance, which still doesnt explain some gaps.

 

[TheLegendOfMart]

This game is weird, no matter what I do changing all the settings down or up in 1080p I get around 30-35fps when I'm up high.

The constant fluctuating between 30 and 60 is annoying.

 

[Seanspeed]

This just doesnt make sense. Unless game heavily uses AXV2 performance, which still doesnt explain some gaps.

It wouldn't be the first game where Haswell seems to have a noticeable advantage.

But no, it wouldn't explain the difference between the i3 and i5 Haswells.

EDIT: Actually, clock speed difference and the fact that the game doesn't tend to use more than 2 cores very well on Intels might account for the difference seen.

 

[BreezyLimbo]

This game is weird, no matter what I do changing all the settings down or up in 1080p I get around 30-35fps when I'm up high.

The constant fluctuating between 30 and 60 is annoying.

Try lowering View Distance down to the complete minimum?

 

[icecold1983]

It wouldn't be the first game where Haswell seems to have a noticeable advantage.

But no, it wouldn't explain the difference between the i3 and i5 Haswells.

EDIT: Actually, clock speed difference and the fact that the game doesn't tend to use more than 2 cores very well on Intels might account for the difference seen.

which other games have you seen haswell have such a large perf increase over the 2600k?

 

[Gormenghast]

I am not sure why he's SLI usage isn't hitting max. Mine is hitting 99% on SLI Titans. Wondering, if he's using the correct SLI Bits profile.

1- SLI Titans are far less powerful than 980s
2- It's all about CPU, so if you have higher usage it means you have higher single thread performance.

Notice how in the video if he turns all settings down GPU usage drops to 40%. That basically means a HUGE CPU bottleneck.

His hex core CPU is basically completely useless since this game is all about single thread performance.

 

[Embearded]

Sometimes i feel like developers intentionally make their games run like crap on AMD products. Seeing Nvidia logo everywhere in the menu also makes me believe it even more...

My specs:
AMD 8350 @4.00GHz
7870 OC @1050MHz 2GB
8GB RAM @ 1866MHz

I tried the game and my settings were:
1080p 16:9
Vsync ON
Texture - Shadow - Foliage = Medium
Nividia things OFF
AA ON
Motion Blur OFF (i dont like it)
View Distance = defualt

Inside the tower i had 45-60+ fps and when i got out of the elevator BAM 15fps!
WTF???

 

[Kezen]

This just doesnt make sense. Unless game heavily uses AXV2 performance, which still doesnt explain some gaps.

Indeed the gap between Sandy/Ivy and Haswell is pretty striking.

Techland are apparently planning on fixing some performance issues :
http://steamcommunity.com/app/239140/discussions/0/613948093895564127/

 

[Gormenghast]

Indeed the gap between Sandy/Ivy and Haswell is pretty striking.

Techland are apparently planning on fixing some performance issues :
http://steamcommunity.com/app/239140/discussions/0/613948093895564127/

You can already know they are doing it wrong when they ask for "dxdiag" config files when it's blatantly obvious the problem is not related to specific hardware and it's just due to very bad threading code.

 

[LabouredSubterfuge]

Indeed the gap between Sandy/Ivy and Haswell is pretty striking.

Techland are apparently planning on fixing some performance issues :
http://steamcommunity.com/app/239140/discussions/0/613948093895564127/

Sounds more like bugs and glitches than performance, but one can hope.

 

[MaLDo]

You can already know they are doing it wrong when they ask for "dxdiag" config files when it's blatantly obvious the problem is not related to specific hardware and it's just due to very bad threading code.

But I can play with locked 60 fps in 2 GB gpus. So I guess an specific bug fix could help in a lot of performance problems.

 

[riflen]

Sometimes i feel like developers intentionally make their games run like crap on AMD products. Seeing Nvidia logo everywhere in the menu also makes me believe it even more...

My specs:
AMD 8350 @4.00GHz
7870 OC @1050MHz 2GB
8GB RAM @ 1866MHz

There is another theory that fits the facts; amd CPUs have long since been eclipsed by Intel's. Your CPU is beaten by some i3 CPUs in games.
Also your GPU is 3 years old and this game looks like it benefits from > 8GB RAM.

 

[Smash88]

I think GameGPU should really stop being used as a benchmark, they are inconsistent and unreliable. Why is this the norm on Neogaf anyways?

 

[Seanspeed]

which other games have you seen haswell have such a large perf increase over the 2600k?

Still trying to find it. It was something somewhat recent. Not games, just a game.

EDIT: Here we go:

 

[Gormenghast]

But I can play with locked 60 fps in 2 GB gpus. So I guess an specific bug fix could help in a lot of performance problems.

He can too. It all depends on view range and CPU. If you're not saying what you're using then other observations are useless.

The bottom line is:
- This game is considerably CPU bound
- What matters is solely single thread CPU performance
- Core usage goes proportionally to view distance

So, the FPS you'll have on your screen will depend on your CPU single thread performance and view range. That's all that matters here. SLI, 980s or 970s are not a factor in this game.

 

[icecold1983]

I think GameGPU should really stop being used as a benchmark, they are inconsistent and unreliable. Why is this the norm on Neogaf anyways?

their benches pretty much align with everyones performance and andys report on geforce.com

Still trying to find it. It was something somewhat recent. Not games, just a game.

ok, let me know. this is the first time i remember seeing such a dramatic gap

 

[Kezen]

You can already know they are doing it wrong when they ask for "dxdiag" config files when it's blatantly obvious the problem is not related to specific hardware and it's just due to very bad threading code.

Yeah, the problem clearly does not lie in the user's hardware this time around. The game should run better on my PC and many other's.

Sounds more like bugs and glitches than performance, but one can hope.

Haven't run into bugs or jarring glitches yet but performance clearly leaves a lot to be desired. That should be their focus.

Still trying to find it. It was something somewhat recent. Not games, just a game.

The Evil Within.

 

[GHG]

I think GameGPU should really stop being used as a benchmark, they are inconsistent and unreliable. Why is this the norm on Neogaf anyways?

They are usually the first ones to get benchmarks out and they usually do it the day a game gets released so that's a big factor.

While not fool proof, they usually give you a good idea of what you can expect performance wise from similar hardware.

 

[MaLDo]

He can too. It all depends on view range and CPU. If you're not saying what you're using then other observations are useless.

The bottom line is:
- This game is considerably CPU bound
- What matters is solely single thread CPU performance
- Core usage goes proportionally to view distance

So, the FPS you'll have on your screen will depend on your CPU single thread performance and view range. That's all that matters here. SLI, 980s or 970s are not a factor in this game.

Default view distance, I wrote my config a few posts above. I've read a lot of users saying they can't fix their framerate using lowest settings. That's why I talked about performance bug.

 

[Smash88]

their benches pretty much align with everyones performance and andys report on geforce.com

A lot of people on this very thread would disagree, but I would still like benchmarks from a more reputable site.

They are usually the first ones to get benchmarks out and they usually do it the day a game gets released so that's a big factor.

While not fool proof, they usually give you a good idea of what you can expect performance wise from similar hardware.

I see where you are coming from, as I mentioned above I would've still liked it from a more reputable site.

 

[Kezen]

They are usually the first ones to get benchmarks out and they usually do it the day a game gets released so that's a big factor.

While not fool proof, they usually give you a good idea of what you can expect performance wise from similar hardware.

Sometimes Gamegpu.ru do not use very good scenes for benchmarking. Before looking at the data, look at the portion of the game they used. In the case of Dying Light there is a missed opportunity. They should have used an intense parkour run with a lot of vistas.

PCgameshardware.de have benchmarked the game :
http://www.pcgameshardware.de/Dying-Light-PC-257307/Specials/Technik-Test-Benchmarks-1149149/

 

[Hooks]

i5 4690k @ 4.4ghz
GTX 780 windforce 3GB
8GB Ram @ 2133

Played at max settings @ 1920x1080 for 4.5 hours, No crashes, No frame rate issues. Runs butter smooth.

 

[Embearded]

There is another theory that fits the facts; amd CPUs have long since been eclipsed by Intel's. Your CPU is beaten by some i3 CPUs in games.
Also your GPU is 3 years old and this game looks like it benefits from > 8GB RAM.

Sure, i won't argue with that that, but have you seen the requirements? I am above minimum and close to recommended. Right now i am testing the game with everything low on 900p and still 20-25fps.
It's obvious that whatever they are doing, they are doing it wrong.

 

[Kezen]

i5 4690k @ 4.4ghz
GTX 780 windforce 3GB
8GB Ram @ 2133

Played at max settings @ 1920x1080 for 4.5 hours, No crashes, No frame rate issues. Runs butter smooth.

Butter smooth ? How is it possible.

 

[icecold1983]

A lot of people on this very thread would disagree, but I would still like benchmarks from a more reputable site.



I see where you are coming from, as I mentioned above I would've still liked it from a more reputable site.

thats because people are comparing their perf with some low draw distance to gamegpus max draw distance benchmarks. gamegpu aligns perfectly with everything andy wrote on geforce.com

Sometimes Gamegpu.ru do not use very good scenes for benchmarking. Before looking at the data, look at the portion of the game they used. In the case of Dying Light there is a missed opportunity. They should have used an intense parkour run with a lot of vistas.

PCgameshardware.de have benchmarked the game :
http://www.pcgameshardware.de/Dying-Light-PC-257307/Specials/Technik-Test-Benchmarks-1149149/

they did use an expansive vista view for their cpu test. also you were right about the evil within. that game is also a joke wrt threading. its pretty much a 1.5 core game.

 

[Saiyan-Rox]

Any idea on what settings I could use?

i5 3570k (overclocked to 4.2GHz)
12GB RAM
AMD 7950

 

[Faith]

Butter smooth ? How is it possible.

Maybe 30fps are butter smooth for some?

 

[Seanspeed]

Maybe 30fps are butter smooth for some?

Its a little annoying when people post their performance impressions in these threads without actually giving any numbers. Stuff like, "My framerate is fine/good" is very vague.

 

[aiat_gamer]

Maybe 30fps are butter smooth for some?

Yeah probably. It is obvious the game has highly variable frame rate regardless of the hardware and yet, people are running around claiming they are running it on Max setting "buttery smooth"...kids these days.

 

[owasog]

https://www.youtube.com/watch?v=VDqVv5xQ5eo

Totalbiscuit did his analysis of it.

Wow. One CPU core at 99%, and the rest at 50% or lower. Could this be the very thing DirectX12 is supposed to fix? It looks similar. Greater view distance = more draw calls?

 

[Kezen]

Its a little annoying when people post their performance impressions in these threads without actually giving any numbers. Stuff like, "My framerate is fine/good" is very vague.

Maybe he has not stepped outside just yet. No erratic performance indoors.
When you wander outside the framerate tanks hard below 60.

I will be highly impressed if consoles use PC's max distance settings and are 30fps locked.

Greater view distance = more draw calls?

I think so but I'm no expert.

Hopefully DX12 isn't an excuse to do even less optimization on PC.

With the CPU headroom available offered by a lighter API they will optimize even less. Basically they will make us brute force their code but that will work much better this time.

 

[icecold1983]

Wow. One CPU core at 99%, and the rest at 50% or lower. Could this be the very thing DirectX12 is supposed to fix? It looks similar. Greater view distance = more draw calls?

yes dx12 will help multithreading and driver/draw call bottlenecks, but this is still just lousy coding. you can do much better than this on dx11. dx12 also wont be some magic cure, in some ways its going to take even more develop work, and much more developer responsibility not to crash your system. i expect many devs to stick with current methods.

 

[Gormenghast]

Wow. One CPU core at 99%, and the rest at 50% or lower. Could this be the very thing DirectX12 is supposed to fix? It looks similar. Greater view distance = more draw calls?

Yeah, DX12 to the rescue of incompetent coders

Hopefully DX12 isn't an excuse to do even less optimization on PC.

 

[icecold1983]

Maybe he has not stepped outside just yet. No erratic performance indoors.
When you wander outside the framerate tanks hard below 60.

I will be highly impressed if consoles use PC's max distance settings and are 30fps locked.

no way consoles use 100% max draw distance. even if the console version had great threading i suspect their gpus would become too much of a bottleneck for the increased rendering load at 1080p. id expect somewhere around 40 to 50% and a mix of medium and high settongs.

 

[martino]

Maybe he has not stepped outside just yet. No erratic performance indoors.
When you wander outside the framerate tanks hard below 60.

I will be highly impressed if consoles use PC's max distance settings and are 30fps locked.

I bet more on frequent dips

Spoiler

maybe even under 25

whatever setting used.

 

[Kezen]

no way consoles use 100% max draw distance. even if the console version had great threading i suspect their gpus would become too much of a bottleneck for the increased rendering load at 1080p. id expect somewhere around 40 to 50% and a mix of medium and high settongs.

I would not rule that out just yet.
Dying Light seems really unoptimized on PC in its current state.
If they spent more time on consoles this will show, and consoles can do more with less CPU power.

It's not like they are magic or anything, it's just that the PC API is very thick. Of course no one discovers that fact in 2015.

 

[Jimmy Rustle]

i5 4670k running at stock 3.4GHz
Sli 970 at 1378MHz
16gb of RAM

1620p fluctuates between 44-60fps with vsync enabled in Nvidia CP (in-game off). Some stutter when running through the town.

Resolution: 2880x1620
Texture Quality: High
Shadow Map Size: High
Foliage Quality: High
View Distance: Around 30%
Ambient Occlusion: On
Nvidia HBAO+: On
Nvidia Depth of Field: Off
Antialiasing: On

Sacrificed View Distance and Field of View (-1) quite a bit. Didn't really mind though. I might reduce resolution to 1440p if the stutter annoys me more.

 

[icecold1983]

I would not rule that out just yet.
Dying Light seems really unoptimized on PC in its current state.
If they spent more time on consoles this will show, and consoles can do more with less CPU power.

It's not like they are magic or anything, it's just that the PC API is very thick. Of course no one discovers that fact in 2015.

looking forward to DFs analysis. altho nx gamer on youtube does good comparison videos as well.

https://www.youtube.com/watch?v=Vnm6UDD6_TI

heres a video of him discussing ps4 framerate. anyone who has played these similar scenes can give their subjective thoughts on how the settings seem to match up to pc version. im also curious to know if the ps4 version has a better post AA. seems like the pc version might use fxaa. does the pc version also try to pass off ghosting as mblur?

 

[Klyka]

The difference in FPS with a low FOV is staggering but i just can't play at that. It's impossible to see ANYTHING.

 

[TheLegendOfMart]

Try lowering View Distance down to the complete minimum?

Thanks I've tried that, makes a few fps difference at best.

Guess I am CPU bound.

 

[Vuze]

Fix for Chromatic Aberration inside if sb needs it, tbh I don't mind it that much in this game: http://www.reddit.com/r/dyinglight/comments/2turvm/pc_how_to_tweak_and_mod_dying_light/

 

[dsk1210]

Anyone tried this with G-sync, does it smooth it out all?

 

[TAJ]

Its a little annoying when people post their performance impressions in these threads without actually giving any numbers. Stuff like, "My framerate is fine/good" is very vague.

Sometimes very specific comments are just as bad. "I got the PS4 version of Shadow of Mordor early and it's totes 60fps."