Finding out whether your FPS are CPU or graphics card limited

[dud]

Do just take the "unreal is cpu intensive" unmeasured blah-blah statements.

Do this:

• get framerate in a reproducible way that you like
• go into the BIOS and downclock your CPU by 1/3rd, repeat benchmark
• download an overclocking utility for your graphics card and use it to downclock your GPU by 1/3rd, repeat benchmark
• for added fidelity, do same for:
• RAM
• graphic card RAM

I don't have time right now but I'll post mine later.

We need a reference thread about hardware-to-fps anyway.

 

[metaDp]

Aaaaaaaaaannnnnnd 5 likes now .

Alone on track:
Same time/weather as the quick race, doing 2 laps.

Epic graphic settings:
CPU Stock: 50/95/110
CPU @4,3: 82/92/110 *Nice minimum FPS boost.

Medium graphic settings:
CPU stock: 110/134/167
CPU @4,3: 120/145/183


I think I will work on a dedicated thread this week .

 

[T_rac3]

Time to upgrade to a Ryzen

I feel for you

From my part I'm done with the upgrade for ... some time. I'm OK with AC and rFactor2. As ACC, from now on it is Kuno's job to do some optimization. I give him my support, now it is his turn to give me some support (optimization)

 

[Deleted member 387850]

I finally did a bit of benchmarking this evening by measuring the min/avg/max and came to the conclusion that it doesn't make a blind bit of difference what I set the graphics settings at - in a 19 car AI race my Ryzen 5 1600 is maxed out and my FPS barely changes whether I set everything to Epic or Low. If I look on the bright side, I guess there's no point rushing out to buy myself a 1080 Ti

If I jump into the Special Events quick race I'm averaging about 68fps with everything set to either Epic or High. Not bad I guess considering I'm only running on a single 1080p at 60Hz, and my 1060 6GB is cruising whilst it waits for the CPU.

One thing I am experiencing (or noticing more!) is some pretty horrid screen tearing, so I've had to enable Fast Sync in NVCP as I just can't bare it any longer. It seems to cause some occasional micro stutter, but I can live with it better than constant tearing through each bend!

 

[2stains]

I finally did a bit of benchmarking this evening by measuring the min/avg/max and came to the conclusion that it doesn't make a blind bit of difference what I set the graphics settings at - in a 19 car AI race my Ryzen 5 1600 is maxed out and my FPS barely changes whether I set everything to Epic or Low. If I look on the bright side, I guess there's no point rushing out to buy myself a 1080 Ti

If I jump into the Special Events quick race I'm averaging about 68fps with everything set to either Epic or High. Not bad I guess considering I'm only running on a single 1080p at 60Hz, and my 1060 6GB is cruising whilst it waits for the CPU.

One thing I am experiencing (or noticing more!) is some pretty horrid screen tearing, so I've had to enable Fast Sync in NVCP as I just can't bare it any longer. It seems to cause some occasional micro stutter, but I can live with it better than constant tearing through each bend!

I thought fast sync was only advisable when you are doubling the monitor refresh rate , why not use vysnc ? Latancy ?
So Ryzen 5 is big Time bottleneck at 1080p with 1060 !

 

[Opus131]

The 1600 isn't the best Ryzen out there but it's possible the game is just horribly optimized.

An alternative to getting a new CPU, especially if you have loads of DDR3 RAM, is getting a used intel CPU, something like an i7 4790K. Intel CPUs still have more powerful single core performance which is an advantage for most games still, while Ryzen is more of an all purpose CPU and is more future proof since it fairs a lot better in DX12 games. It's also cheaper if you buy new. Where i live the 2600x is 200 euros i see no reason to bother settling for anything less.

 

[2stains]

I went 7700k to pair with my 1080Ti and hopefully be fine for my new 3080Ti later next year !

 

[Mitja Bonca]

Well, if you were able to get your hands on a 4770K (or even 4790K) for say €100 then it might be a simple and attractive upgrade - no need to change anything else. However, a quick bit of googling suggests the prices for those i7s are still stupidly high. Weird! I'd been thinking the prices would have dropped way down by now.

i7 4770K has same specs in some even slower than i5 4670K according to http://cpu.userbenchmark.com website.
4790K is 10 to 20% faster and with the same 1150 socket, its a smart think to change with.
I own 4670K to, and the only reasonable change atm is only 4790K, all else would need a new motherboard as well (and new windows install too).

 

[Muzzy]

Do you guys think it's worth getting a K processor (e.g.: i7-4790K) even if the motherboard does not support overclocking?

I'm considering changing my i5-4790 to a i7-4790k.

 

[RasmusP]

Do you guys think it's worth getting a K processor (e.g.: i7-4790K) even if the motherboard does not support overclocking?

I'm considering changing my i5-4790 to a i7-4790k.

No. You'll get a slightly higher boost clock but that's really not much...
If you pay below 50€ for it then you could do it but otherwise... not really!

 

[Jempy]

Below 50€, I'm guessing it will be difficult to find one ( certainly still of good quality ).
So better .... keep the money ( I guess +- 100€ for a correct one ) .... and spare for a change within some months ( or even maybe 1 year ).

That's why I decided finally .... too much for such a change and such a result ... I prefer sparing.

 

[RasmusP]

Below 50€, I'm guessing it will be difficult to find one ( certainly still of good quality ).
So better .... keep the money ( I guess +- 100€ for a correct one ) .... and spare for a change within some months ( or even maybe 1 year ).

That's why I decided finally .... too much for such a change and such a result ... I prefer sparing.

I agree!
To make it clear: I meant within 50€ for selling his current one and getting the other one!

 

[Jesper Van der Woude]

I have been following this thread since the beginning and I notice that alot of people are talking about lower fps due to a CPU bottleneck.
With my setup I don't have particularly low fps with 19 AI. I can keep a steady 55-60 with epic/high settings. EVGA GTX970 SC and i5-4690(non k) and DDR3 2x4GB 1600mhz on a ASrock H87 pro 4.
Monitor is a 1080p 32 inch 60hz Samsung TV.
But the thing that bothers me is that I experience stuttering in this scenario. But I don't see anything happening to the FPS and RAM usage is at 6300 mb. And later on, when the field has devided, I still experience this.
Does anyone else relate to this and does anyone have an explanation for this?

 

[RasmusP]

I have been following this thread since the beginning and I notice that alot of people are talking about lower fps due to a CPU bottleneck.
With my setup I don't have particularly low fps with 19 AI. I can keep a steady 55-60 with epic/high settings. EVGA GTX970 SC and i5-4690(non k) and DDR3 2x4GB 1600mhz on a ASrock H87 pro 4.
Monitor is a 1080p 32 inch 60hz Samsung TV.
But the thing that bothers me is that I experience stuttering in this scenario. But I don't see anything happening to the FPS and RAM usage is at 6300 mb. And later on, when the field has devided, I still experience this.
Does anyone else relate to this and does anyone have an explanation for this?

Vsync on or off? With it on you'll see a stutter every time you drop below 60 as that means the same frame is shown a second time.
It works like that:
0.9s = 60 fps
0.1s = 30 fps (image shown 2x)

Averaged over a second it will be 58 fps for example. (Didn't calculate it )

To your non existent CPU bottleneck: did you try 19 AI in dark storm? Or sunset? Should give the CPU something to do with all the reflections and shadows

 

[Jesper Van der Woude]

Vsync on or off? With it on you'll see a stutter every time you drop below 60 as that means the same frame is shown a second time.
It works like that:
0.9s = 60 fps
0.1s = 30 fps (image shown 2x)

Averaged over a second it will be 58 fps for example. (Didn't calculate it )

To your non existent CPU bottleneck: did you try 19 AI in dark storm? Or sunset? Should give the CPU something to do with all the reflections and shadows

I always have Vsync off.
I didn't try anything other than sunset with 19 AI, because that already gave me stuttering.
But could this stuttering be related to my CPU?

 

[2stains]

Do you guys think it's worth getting a K processor (e.g.: i7-4790K) even if the motherboard does not support overclocking?

I'm considering changing my i5-4790 to a i7-4790k.

I have a 7700k that will do 4.9 all day but i choose to run it at 4.5 just because atm im so GPU bound with my 1080Ti in VR and 4k that those extra Mhz are only creating heat and not gaining me any performance !
So having a K CPU will give you a bit in benchmarks but in reality it will depend on resolution and other specs , but not really a big deal if you don't get the k !

 

[Deleted member 197115]

Do we know if ACC is single rendering thread bottlenecked like AC.
Hard to gauge based on the current unoptimized state of the game and lack of in game rendering stats monitoring.

 

[RaceWasGood]

Wasn't sure where to post this, so i'll leave it here.

EDIT: Fullscreen is disabled just for the sake of getting a screenshot with snipping tool. Normally enabled

Here's the settings i've found works quite good on a 4K monitor @60Hz. The setup is: i7 6700K @4400MHz + factory overclocked STRIX-GTX1080 in OC mode. I've also set the UE4 sharpening filter to r.Tonemapper.Sharpen=1. With these settings i'm able to maintain a steady 60fps (v-sync enabled) in all weather/lighting conditions i've tried so far and i'm quite pleased with how everything looks right now.

 

[Amir Vodokotlic]

What tool is everyone using for min/average/high FPS tests.

Fraps?

 

[Case_]

Afterburner, more than likely. Fraps is ancient.

 

[RasmusP]

What tool is everyone using for min/average/high FPS tests.

Fraps?

Fraps is only awesome when you do frametime analysis. But you need to put the log with "frame to frame" pacing into excel and draw a graph.

Afterburner with it's graph is way easier and gives you a good overview!