GTX 1080 Performance in Assetto Corsa

  • Thread starter RReed43
  • 21 comments
  • 17,359 views
404
United States
Aubrey, Texas
I bought a GTX 1080 FE from Nvidia, it arrived today and I installed it in my sim racing PC replacing the GTX 970 OC that I had been using previously. I run a i7 4820 3.7 ghz liquid cooled and over clocked to 4.6 ghz. My monitor is 4k at 60hz. AC is the only racing game that I have been able to run at full resolution and maintain 60 fps. I do this by having AA off and medium settings on most of the quality sliders.

I purchased the 1080 primarily so I would be able to have the highest quality possible when my Oculus Rift arrives later this summer.

The 1080 installed without issue beyond the normal USB/AC controller issues that seem to occur when you change out critical hardware. It is a longer card and much heavier than the 970 it replaced.

I benchmarked the 970 immediately before I replaced it with the 1080.

AC VERSION: 1.6.3 (x64)
FPS: AVG=98 MIN=51 MAX=128 VARIANCE=1 CPU=53%
POINTS: 14426

LOADING TIME: 17s
GPU: NVIDIA GeForce GTX 970 (3840x2160)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1

I did no OC or tuning on the 1080, simply installed it and ran the benchmark again.

This is the 1080 results.

AC VERSION: 1.6.3 (x64)
POINTS: 19586
FPS: AVG=133 MIN=67 MAX=166 VARIANCE=6 CPU=68%

LOADING TIME: 16s
GPU: NVIDIA GeForce GTX 1080 (3840x2160
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1

The score increased by about 40%, I expected more of an increase but this is still a pretty big jump. The CPU was being more heavily utilized moving from 53% to 68% but still has headroom. The avg frames moved from 98 to 133. More importantly the minimum moved from 51 to 67 producing smoother running noticeable at the crowded corners.

Since 4k at 60 fps is about equal to the Rift at 90 fps I think this card will be a worthwhile investment. I will do some overclocking and see what the benchmark looks like with that.
 
I bought a GTX 1080 FE from Nvidia, it arrived today and I installed it in my sim racing PC replacing the GTX 970 OC that I had been using previously. I run a i7 4820 3.7 ghz liquid cooled and over clocked to 4.6 ghz. My monitor is 4k at 60hz. AC is the only racing game that I have been able to run at full resolution and maintain 60 fps. I do this by having AA off and medium settings on most of the quality sliders.

I purchased the 1080 primarily so I would be able to have the highest quality possible when my Oculus Rift arrives later this summer.

The 1080 installed without issue beyond the normal USB/AC controller issues that seem to occur when you change out critical hardware. It is a longer card and much heavier than the 970 it replaced.

I benchmarked the 970 immediately before I replaced it with the 1080.

AC VERSION: 1.6.3 (x64)
FPS: AVG=98 MIN=51 MAX=128 VARIANCE=1 CPU=53%
POINTS: 14426

LOADING TIME: 17s
GPU: NVIDIA GeForce GTX 970 (3840x2160)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1

I did no OC or tuning on the 1080, simply installed it and ran the benchmark again.

This is the 1080 results.

AC VERSION: 1.6.3 (x64)
POINTS: 19586
FPS: AVG=133 MIN=67 MAX=166 VARIANCE=6 CPU=68%

LOADING TIME: 16s
GPU: NVIDIA GeForce GTX 1080 (3840x2160
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1

The score increased by about 40%, I expected more of an increase but this is still a pretty big jump. The CPU was being more heavily utilized moving from 53% to 68% but still has headroom. The avg frames moved from 98 to 133. More importantly the minimum moved from 51 to 67 producing smoother running noticeable at the crowded corners.

Since 4k at 60 fps is about equal to the Rift at 90 fps I think this card will be a worthwhile investment. I will do some overclocking and see what the benchmark looks like with that.
Now I want the 1080 or 1070 or the new readon Polaris
 
Thanks for sharing :cheers:

Even though I ran the below benchmarks at 2K with my Nvidia GTX 980 Ti OC, this might be useful info for some. Once my 4K monitor is in place, i'll share the benchmark results. :gtpflag:

Intel Core i7-6700K DDR4/ NVIDIA® GeForce GTX980Ti 6GB GDDR5 / 16 GB RAM DDR4 2133MHz / 2x M.2 PCI-e SSD RAID 0 512 GB + 2TB SATA III

Maxed out (Assetto Corsa v1.5)
Assetto Corsa - Nvidia GTX 980 Ti Benchmark - Maxed out by M-Bimmer, on Flickr


High (Assetto Corsa v1.5)
Assetto Corsa - Nvidia GTX 980 Ti Benchmark - High by M-Bimmer, on Flickr
 
Last edited:
Thank's for sharing, one thing I wonder though, can you really notice a difference between this game at 100+ fps as opposed to around 60 fps especially considering you concentrate so much on the screen all the time. I only ask because a few people have told me about how once you adjust to 120+ fps playing at around 60 seems really choppy. Personally I can't see why it would make that much difference with a 60hz monitor but you never know.
 
On a 60hz monitor you can't see faster that 60 fps. The issue is the minimum fps. Every time the feed rate drops below 60 fps then you will see a judder.

In my bench marking above you will notice that minimum frame rate went from 51 on the GTX 970 to 67 on the GTX 1080. This produced a noticeably smoother display. I have began to crank up the quality elements on AC; bench marking to keep the minimum above 60.

I think people have variable sensitivity to the fps but the reason the Rift and Vive are adamant about have feeds above 90 fps is that the likely hood of motion sickness is greatly increased if the frame rate is less than 90. I expect your friends are telling the truth about the smoothness of 120 hz displays in comparison to 60 hz displays. Unfortunately it takes a LOT of computer firepower to upgrade a high resolution visual at 90 fps and above.
 
I've mentioned it on this forum previously, but it might be worth mentioning again - Kunos apparently coded so that processes (physics, force feedback, etc.) are calculated at multiples of 50 rather than 60. My guess is that maybe they started work based on netKar Pro code, with it's likely PAL-centric timings, and never converted over to the more modern standard of 60/120Hz. So depending on what is supported by an individual's display, 50/75/100Hz may be slower than what is possible, but should in theory be a smoother and more aligned experience.
 
I only ask because a few people have told me about how once you adjust to 120+ fps playing at around 60 seems really choppy. Personally I can't see why it would make that much difference with a 60hz monitor but you never know.
By choppy they are probably only talking about motion blur, so "choppy" is not the best word really.

On a perfect screen with no motion blur, stutter or other artifacts I would argue that nobody would be able to see the difference between 60 fps and 120 fps when studying the only the motion itself, because both are as smooth as it gets.

But in the real world you can certainly see the difference visually on a 120 Hz screen because of the inherent motion blur on LCD screens, which is more noticeable on a 60 Hz screen. And also you can often feel the difference when using a controller because of less input lag at 120 Hz.

So to answer your question: Yes, you can (well, at least I can) definitely tell the difference between AC running at 60 Hz on a 60 monitor compared to 100 hz on a 120 monitor. The difference is in the reduced motion blur.
But, if you are used to playing on a 60 Hz screen and think it's perfectly fine, then I honestly wouldn't bother with it. Because it isn't THAT big of a deal, and your eyes are used to it and you probably pay no attention to it at this point anyway. :)
 
In regard to the performance of the GTX 1080 with Assetto Corsa as measured with the AC Benchmark, I have done further work.

From my previous experiments with the 1080 and attempts to over clock it I had found that the AC Benchmark did not respond to a 10% OC (the most I could achieve) on the GTX 1080. I surmised that this result indicated that the 1080 was maxing my CPU on core #1 which does the rendering for the display. The other cores are used to do the physics on the automobiles themselves.

I was running a i7 4820 OC to 4.7 GHz. The 4820 is a top processor from two generations back. I decided to up grade my CPU and replaced the motherboard, the CPU and the ram. I went with the i7 6700 4.0 GHZ which is a top of the line current generation CPU.

I overclocked the I7 to 4.6 GHz. I then benchmarked the new system on the AC Benchmark. Here are the results.

**************************************************************
AC VERSION: 1.6.3 (x64)
POINTS: 27571
FPS: AVG=188 MIN=63 MAX=232 VARIANCE=12 CPU=66%

LOADING TIME: 13s
GPU: NVIDIA GeForce GTX 1080 (3840x2160)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:2048 BLUR:0
WORLD DETAIL: 5 SMOKE:3
PP: QLT:3 HDR:0 FXAA:1 GLR:3 DOF:3 RAYS:1 HEAT:1
***************************************************************

Here for comparison are the results from when I started with the i7 4820 and GTX 970.


AC VERSION: 1.6.3 (x64)FPS: AVG=98 MIN=51 MAX=128 VARIANCE=1 CPU=53%
POINTS: 14426

LOADING TIME: 17s
GPU: NVIDIA GeForce GTX 970 (3840x2160)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1
**************************************************************

Here are the results for the I7 4820 with the GTX 1080


AC VERSION: 1.6.3 (x64)
POINTS: 19586
FPS: AVG=133 MIN=67 MAX=166 VARIANCE=6 CPU=68%

LOADING TIME: 16s
GPU: NVIDIA GeForce GTX 1080 (3840x2160
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1
*******************************************************************

The i7 6700 with the GTX 1080 is able to produce a score of about double my starting point. This is an incredible amount of performance on a 4K monitor. It does point out that the 1080 is creating a bottleneck on CPU's even for relatively high performance CPU's like the i7 4820.
 
Thank's for sharing, one thing I wonder though, can you really notice a difference between this game at 100+ fps as opposed to around 60 fps especially considering you concentrate so much on the screen all the time. I only ask because a few people have told me about how once you adjust to 120+ fps playing at around 60 seems really choppy. Personally I can't see why it would make that much difference with a 60hz monitor but you never know.

I have a 144hz monitor and did some testing recently to see if I could see/feel a difference with various refresh rates/fps. After all that I decided that 90 to 100 fps/hz is my new minimum preferred frame rate. Using some of the tech demos I can see screen tearing all the way up to 144hz, but that's under worst possible case scenarios. When playing AC I never feel screen tearing with vsync off, and I'm normally running hz at 144 and fps is averaging around 100. 60 fps is fine, but 100+ fps feels better/smoother to me.
 
I purchased the the 1080 a couple of weeks ago. Along with the Acer X34A. An ultrawide 34" monitor, with the ability to overclock to 100hz.

I've only tested with Raceroom and Iracing so far. In short the difference is amazing. Easily runs well over 70fps @ 1440p. My last graphics card was the 970 SSC.

I'll give assetto corsa a try today. I have no doubt it'll perform and look even better!
 
5cb9ac7998bce163e9905231db75d1c5.jpg


Well sh**
 
5cb9ac7998bce163e9905231db75d1c5.jpg


Well sh**

I have a 750 ti and can consistently run 80+ fps, it's all about what you turn down. I run most things off or towards the low end except for AA. I also turned on msaa in nvidia inspector and I enjoy AC plenty without all the eye candy.
 
AC VERSION: 1.6.3 (x64)
POINTS: 8941
FPS: AVG=61 MIN=41 MAX=94 VARIANCE=0 CPU=40%

LOADING TIME: 43s
GPU: NVIDIA GeForce GTX 680 (5760x1080)
OS-Version: 6.2.9200 () 0x300-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:8X SHDW:512 BLUR:0
WORLD DETAIL: 4 SMOKE:3
PP: QLT:4 HDR:1 FXAA:1 GLR:4 DOF:4 RAYS:1 HEAT:1


FPS drops are happening only at the start.
Later on it goes 60fps and up.

AA and Shadows are the most expensive settings.
 


AC VERSION: 1.6.3 (x64)
POINTS: 12349
FPS: AVG=84 MIN=5 MAX=128 VARIANCE=5 CPU=77%

LOADING TIME: 14s
GPU: NVIDIA GeForce GTX 750 Ti (1920x822)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 4
FULLSCREEN: ON
AA:4X AF:4X SHDW:512 BLUR:0
WORLD DETAIL: 0 SMOKE:0
PP: QLT:0 HDR:0 FXAA:0 GLR:0 DOF:0 RAYS:0 HEAT:0

The 5fps was at the very beginning. Not sure why it does that.

I originally started running 1920x822 as a test to simulate an ultrawide aspect as I'm contemplating a monitor upgrade. Since I'm exclusively running in cockpit view I felt no need to go back to 1920x1080. I did rerun the test at 1080p and the results were almost identical.
 
my benchmark is...

AC VERSION: 1.6.3 (x64)
POINTS: 7941
FPS: AVG=54 MIN=37 MAX=64 VARIANCE=1 CPU=67%

LOADING TIME: 34s
GPU: NVIDIA GeForce GTX 960 (1680x1050)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 4
FULLSCREEN: ON
AA:4X AF:4X SHDW:2048 BLUR:0
WORLD DETAIL: 3 SMOKE:2
PP: QLT:3 HDR:1 FXAA:1 GLR:3 DOF:3 RAYS:1 HEAT:1

how i could improve further the frames per second ?
 
my benchmark is...

AC VERSION: 1.6.3 (x64)
POINTS: 7941
FPS: AVG=54 MIN=37 MAX=64 VARIANCE=1 CPU=67%

LOADING TIME: 34s
GPU: NVIDIA GeForce GTX 960 (1680x1050)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 4
FULLSCREEN: ON
AA:4X AF:4X SHDW:2048 BLUR:0
WORLD DETAIL: 3 SMOKE:2
PP: QLT:3 HDR:1 FXAA:1 GLR:3 DOF:3 RAYS:1 HEAT:1

how i could improve further the frames per second ?

I would turn just about everything all the way down/off except the AA settings. That would give you a good idea of what fps your system is capable of.
 
I would turn just about everything all the way down/off except the AA settings. That would give you a good idea of what fps your system is capable of.
I will try it...
But soon i will get a new CPU (i7 6700k) and i will get more RAM (now i have 8 GB but i will get 8 more) and i hope framerate will increase more... ;)
 
In a perfect world the benchmark output would include the processor model number, and would also live in a database somewhere. I'm pretty sure my performance is cpu bound, but would like more proof before I make the investment of upgrading to an i7-6700K, which I'm guessing may help me more than a gpu update.
 
Thank's for sharing, one thing I wonder though, can you really notice a difference between this game at 100+ fps as opposed to around 60 fps especially considering you concentrate so much on the screen all the time. I only ask because a few people have told me about how once you adjust to 120+ fps playing at around 60 seems really choppy. Personally I can't see why it would make that much difference with a 60hz monitor but you never know.
I just wanted to add that when I'm league racing online with 26+ drivers in the session my fps drops from 80s into the high 50s-low 60s, and I do notice judder. Sounds like it's time for a cpu update.
 
In regard to the performance of the GTX 1080 with Assetto Corsa as measured with the AC Benchmark, I have done further work.

From my previous experiments with the 1080 and attempts to over clock it I had found that the AC Benchmark did not respond to a 10% OC (the most I could achieve) on the GTX 1080. I surmised that this result indicated that the 1080 was maxing my CPU on core #1 which does the rendering for the display. The other cores are used to do the physics on the automobiles themselves.

I was running a i7 4820 OC to 4.7 GHz. The 4820 is a top processor from two generations back. I decided to up grade my CPU and replaced the motherboard, the CPU and the ram. I went with the i7 6700 4.0 GHZ which is a top of the line current generation CPU.

I overclocked the I7 to 4.6 GHz. I then benchmarked the new system on the AC Benchmark. Here are the results.

**************************************************************
AC VERSION: 1.6.3 (x64)
POINTS: 27571
FPS: AVG=188 MIN=63 MAX=232 VARIANCE=12 CPU=66%

LOADING TIME: 13s
GPU: NVIDIA GeForce GTX 1080 (3840x2160)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:2048 BLUR:0
WORLD DETAIL: 5 SMOKE:3
PP: QLT:3 HDR:0 FXAA:1 GLR:3 DOF:3 RAYS:1 HEAT:1
***************************************************************

Here for comparison are the results from when I started with the i7 4820 and GTX 970.


AC VERSION: 1.6.3 (x64)FPS: AVG=98 MIN=51 MAX=128 VARIANCE=1 CPU=53%
POINTS: 14426

LOADING TIME: 17s
GPU: NVIDIA GeForce GTX 970 (3840x2160)
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1
**************************************************************

Here are the results for the I7 4820 with the GTX 1080


AC VERSION: 1.6.3 (x64)
POINTS: 19586
FPS: AVG=133 MIN=67 MAX=166 VARIANCE=6 CPU=68%

LOADING TIME: 16s
GPU: NVIDIA GeForce GTX 1080 (3840x2160
OS-Version: 6.2.9200 () 0x100-0x1
CPU CORES: 8
FULLSCREEN: ON
AA:1X AF:2X SHDW:1024 BLUR:0
WORLD DETAIL: 4 SMOKE:2
PP: QLT:5 HDR:0 FXAA:1 GLR:5 DOF:5 RAYS:1 HEAT:1
*******************************************************************

The i7 6700 with the GTX 1080 is able to produce a score of about double my starting point. This is an incredible amount of performance on a 4K monitor. It does point out that the 1080 is creating a bottleneck on CPU's even for relatively high performance CPU's like the i7 4820.


Hello, I just came across this excellent benchmark testing. Great work! :cheers: How well would you predict that your i7-6700/GTX 1080 would perform when running the maximum number of cars at.....Road America (mod) with 48 cars for instance? Do you feel that the i7-6800k might handle a higher number of cars better than the i7-6700k? Again, great work here! :bowdown:
 
Back