Performance Chart For RTX2080TI / I74790K CPU

[dburne]

This really has me puzzled, SSAA in game never worked for me in VR and pretty sure ED stated it would not work in VR. I will have to go in and try setting it again to see what happens.

[Tinkickef]

This really has me puzzled, SSAA in game never worked for me in VR and pretty sure ED stated it would not work in VR. I will have to go in and try setting it again to see what happens.

 

Now I have some framerates (thanks knock knock) I have abandoned SSAA. It appears I am immune to shockingly low frames. From what I just witnessed, I must be accustomed to mid 20s with my old 980TI.

Albeit with the caveat that I usually do moonlit missions, not daytime.

 

I am doing 12 noon daytime testing because it will be more useful to most users and pushes the card harder.

 

So in the interests of this test I have abandoned SSAA, enabled 2x MSAA and dropped PD to 2.1. I kept trees full, reduced ground clutter by 25%. I cannot really comment on how it looks by comparison, as I always think daytime looks washed out and shimmery.

 

 

So here we have the Hornet over the city at low level at mach 1.2 and in a shallow turn to load the card. 44FPS. I can live with that, still not seeing ASW artifacts or stutters.

 

However I suspect I have encountered some thermal throttling on the card as turning around to fly back over the city, framerates drop to mid 30s and dont recover for a few seconds. I will also have to look at fan curves as well as the Nvidia easy overclock scan tool. clearly I have a lot more work to do.

 

 

%5Bimg%5Dhttps%3A//i.imgur.com/pzIdNwM.jpg[/img]

Edited September 29, 2018 by Tinkickef

[Tj1376]

Performance Chart For RTX2080TI / I74790K CPU

 

Can you repost all your pics? They are broken links for me. I am using Tapatalk so let me know if forum users can see them and I’ll log on when I get back home.

 

I’m a big 4790k shouldn’t be getting a 2080ti guy when running 1080p monitors (I still believe that 1080p triple monitors can easily be managed by a 1080.) I’m really interested in your results with VR - and specifically interested in if increasing PD significantly increased your cockpit visibility. Are gauges clear to read? What about plane spotting? Is it night and day difference with visuals now? I read the below from your post- but I couldn’t tell if you thought it was a great improvement or if it was maybe slightly better with just solid frame rates.

 

Really interesting stuff. Thanks for posting.

 

TJ

 

Currently have PD at 2.3 and SSAA at 2.0. Resolution of cockpit slightly better, can read the time and distance to waypoint text on the top right of the Hornet MFCD without leaning forward.

 

Medium range much sharper. Buildings not blurred at 10,000ft. Runway lights seen at 20, 000ft on night mission. Individual streetlights apparent at 8000ft.

Runway lights on approach seen approx 2 - 3x distance further out.

 

 

 

Sent from my iPhone using Tapatalk

Edited September 29, 2018 by Tj1376

[kevman]

Kudos to the OP for doing the tests - cant help but think the DCS should have some decent benchmarking function. Its difficult to interpret what the actual improvement of the 2080 ti is

[Tinkickef]

Can you repost all your pics? They are broken links for me. I am using Tapatalk so let me know if forum users can see them and I’ll log on when I get back home.

 

I’m a big 4790k shouldn’t be getting a 2080ti guy when running 1080p monitors (I still believe that 1080p triple monitors can easily be managed by a 1080.) I’m really interested in your results with VR - and specifically interested in if increasing PD significantly increased your cockpit visibility. Are gauges clear to read? What about plane spotting? Is it night and day difference with visuals now? I read the below from your post- but I couldn’t tell if you thought it was a great improvement or if it was maybe slightly better with just solid frame rates.

 

Really interesting stuff. Thanks for posting.

 

TJ

 

Currently have PD at 2.3 and SSAA at 2.0. Resolution of cockpit slightly better, can read the time and distance to waypoint text on the top right of the Hornet MFCD without leaning forward.

 

Medium range much sharper. Buildings not blurred at 10,000ft. Runway lights seen at 20, 000ft on night mission. Individual streetlights apparent at 8000ft.

Runway lights on approach seen approx 2 - 3x distance further out.

 

 

 

Sent from my iPhone using Tapatalk

 

 

https://i.imgur.com/rKyfdDs.jpg

 

https://i.imgur.com/jzuNGy4.jpg

 

https://i.imgur.com/pzIdNwM.jpg

[Tinkickef]

Another test. VRAM.

 

With Oculus ASW off. Raises another set of questions. No matter what I do re left ctrl and num, My upper framerate remains capped at 45fps in game, as if Oculus will not allow higher, unless the DCS framecounter is just counting one eye in which case the rates in reality are double what it is showing. ASW is off as now the frames are fluctuating between a low of maybe 28 - 30fps and 45fps.

 

MSAA off, PD 2.0, SSAA off. Otherwise all other settings same as before.

 

MSAA still knocks hell out of framerates. 2x = 5fps or more depending on what your altitude is. 4x = 10 - 15fps.

 

graph is of Hornet cruising through downtown Dubai at around 400kts and a couple of hundred feet maybe.

 

FPS = 30.

 

CPU = 23%

 

GPU = 60% with VRAM at 8GB.

 

I'm thinking this is a memory issue holding things back. Maybe my DDR3 1866 is holding things up a little. Or it may be Rift runtime computations that are slowing things up.

 

Anyone any ideas?

 

Also noticed that DCS hangs onto MSAA settings if you try to change them on the fly. You have to shut the game down and restart before the new setting makes an effect on the framerate.

 

Had enough testing for one day. Its been a rollercoaster of excitement and disappointment, albeit I know I am making the card work hard in the worst of cases; speed, large detailed city, textures on high, low level, high PD, high draw distance, High trees and high ground clutter. The framerates do not match the smoothness I am seeing. If the Dubai graph is true, then my 980TI in the same circumstances was drawing 5 - 10fps and that does not seem true either.

 

Cant make head or tail of it.

 

 

%5Bimg%5Dhttps%3A//i.imgur.com/QH3hZfI.jpg[/img]

Edited September 29, 2018 by Tinkickef

[dburne]

I personally think you would be better off lowering that PD to 1.5.

Going much above that really seems to have diminishing returns, maybe the 2080 will be different in that regard but that has certainly been the case with the 10 series cards.

[Tinkickef]

What I don't understand is that neither the CPU or the card are anyway near being maxed out. The CPU is clearly delivering all the data the card needs and the card is putting it out with ease.

There is another bottleneck somewhere. I did suspect RAM, but lookat the VRAM. no peaks or troughs whatsoever and no hourglass, jitter or stutter.

Could Rift computing / display engine be slowing it all down?

 

 

EDIT :doh: Tried same mission again, switched off ASW and this time got a jitter fest at 30fps. Looks like I was suffering from fat finger syndrome.

Edited September 29, 2018 by Tinkickef

[etherbattx]

What I don't understand is that neither the CPU or the card are anyway near being maxed out.

there is always latency and interruptions in a system. which prevents either component from being able to run at full speed.

[Nascar]

There is another bottleneck somewhere.

 

Maybe test external view. I bet that Res cockpit display 1024 every frame is hurting you. I don't notice much difference on the lower 512 setting.

 

Chimney smoke is a killer for me but that shouldn't matter on PG.

 

Start out on the VR preset and turn things up one at a time.

[Knock-Knock]

I bet that Res cockpit display 1024 every frame is hurting you. I don't notice much difference on the lower 512 setting.

 

That only applies to MFD viewports, like Targeting pod, Maverick. Makes no difference on any dials, HUD or just regular MFD pages.

[Tinkickef]

After a few beers, well more beer than is good for me, I have come to a final conclusion.

 

It is a conclusion that is already well known and simply that DCS is so poorly optimised for VR that no hardware can make up for it. No doubt ED will have plans in the pipeline to address this sometime in the future.

 

So to cut to the chase.

 

If you want to run high settings and far higher PD than before, then look closely at 1080TI benchmarks and compare them to my results before committing to 2080TI. You will be relying on ASW pretty much all the time anyway.

 

Current VR hardware cannot utilise the enormous power this card has and is something else to consider.

 

If you have a 1080TI and are considering the 2080TI, my advice would be to hold off for now.

 

If you have a 9 series card, are looking to upgrade and can still afford to eat after buying the 2080TI, then by all means buy it. I have no regrets and see no point in buying obsolete tech, not knocking the 10 series, fine cards, but they are now obsolete. Until DLSS and ray tracing comes onstream, I see little point in buying the vanilla 2080.

 

What the 2080TI will give you is a stable as a rock experience with very very few jitters even at common chokepoints like turning onto the runway. Very smooth even at high settings.

 

I think that overall the card has coped admirably with the torture I have thrown at it. Remember all my testing was done in absolutely worst case scenarios. Flying my usual missions are impressive as Hell.

 

MSAA and SSAA. Frame killers and I cannot say they are worth it. Not sure if they are a placebo or not. SSAA, I may just turn back on to load the card.

 

Pixel Density. This has the most effect on image sharpness, I can see the difference between 1.8 and 2.3, even if it does suffer from diminishing returns and I will bump this back up. No point running the card at 60% chasing frame rates that are simply not there. DCS needs serious optimisation work.

You still need to trade image quality for frame rates, I think this bottleneck is within the DCS engine itself. Again individual choice, I prefer quality over quantity, but then for me, combat is secondary to flying.

 

I7 4790K OC to 4.48ghz. Impressive as Hell, it has taken everything I threw at it, chewed it up and spat it out.

 

 

Hope you found these tests to be as unbiased and as truthful as possible, my intention is to help others make a more informed decision re upgrading.

 

For myself, I am happy overall, maybe a little disappointed in some areas, but those areas are not the fault of the card. The 2080TI is not a panacea for all of ED's ailments.

 

Thanks for looking in.

Edited September 30, 2018 by Tinkickef

[Sydy]

Hi Tinkickef,

 

I just would like to say thank you for your posts. I have one question for you, although I know you don’t have a 2080. Do you think a 2080 maybe a better relation price/performance than a 2080TI for a non-VR user? I have a 1080 right now, pretty satisfied, but looking for an upgrade. At this moment 2080TI is rare on the market, but 2080 is plenty and affordable.

 

What do you think?

 

Tks,

 

Sydy

[Tinkickef]

Hi Tinkickef,

 

I just would like to say thank you for your posts. I have one question for you, although I know you don’t have a 2080. Do you think a 2080 maybe a better relation price/performance than a 2080TI for a non-VR user? I have a 1080 right now, pretty satisfied, but looking for an upgrade. At this moment 2080TI is rare on the market, but 2080 is plenty and affordable.

 

What do you think?

 

Tks,

 

Sydy

 

As with everything...It depends.

 

2080 has around the same performance as the 1080TI at a higher price point. It depends how much value you put on ray tracing and DLSS, neither of which is valid for DCS at this point in time.

If you plan on buying games like Tomb Raider with RT and DLSS, it is probably worth it.

[Fangio]

 

It is a conclusion that is already well known and simply that DCS is so poorly optimised for VR that no hardware can make up for it. No doubt ED will have plans in the pipeline to address this sometime in the future.

 

 

This. VR has been a barely supported afterthought by ED from day 1. Basically, they support it just enough for marketing to claim support but not enough to actually deliver a good experience to the average VR flyer. This continues and is why to won't see Vulcan until 2020 or later.

[Zoomer]

The 4790k is still one of the fastest cpu`s out there and will have no problem running two RTX 2080ti`s if needed at high resolution. It`s the software utilization that is the main issue. You should be able to run the 4790k at 4.6 ghz w/o issue. I have mine running at 4.6ghz @ 1.26v, max temps would be 68-71c @100% testing load or 48-50c with DCS running (approx 15-20% load- rift & Gtx 1070)). That`s with an AIO cooler. It`s a great CPU that will last many more years until it`s anywhere near maxed.

Edited September 30, 2018 by Zoomer

[some1]

What I don't understand is that neither the CPU or the card are anyway near being maxed out. The CPU is clearly delivering all the data the card needs and the card is putting it out with ease.

 

That's probably single core CPU performance. You show 23% CPU usage, which is about two CPU cores fully loaded, and that sounds about right for DCS.

 

Also even without ASW, if you get frame rates close to 45, then you'll see 45 fps. It's just how the older smoothing algorithm (ATW) works.

[Zoomer]

After a few beers, well more beer than is good for me, I have come to a final conclusion.

 

It is a conclusion that is already well known and simply that DCS is so poorly optimised for VR that no hardware can make up for it. No doubt ED will have plans in the pipeline to address this sometime in the future.

 

So to cut to the chase.

 

If you want to run high settings and far higher PD than before, then look closely at 1080TI benchmarks and compare them to my results before committing to 2080TI. You will be relying on ASW pretty much all the time anyway.

 

Current VR hardware cannot utilise the enormous power this card has and is something else to consider.

 

If you have a 1080TI and are considering the 2080TI, my advice would be to hold off for now.

 

If you have a 9 series card, are looking to upgrade and can still afford to eat after buying the 2080TI, then by all means buy it. I have no regrets and see no point in buying obsolete tech, not knocking the 10 series, fine cards, but they are now obsolete. Until DLSS and ray tracing comes onstream, I see little point in buying the vanilla 2080.

 

What the 2080TI will give you is a stable as a rock experience with very very few jitters even at common chokepoints like turning onto the runway. Very smooth even at high settings.

 

I think that overall the card has coped admirably with the torture I have thrown at it. Remember all my testing was done in absolutely worst case scenarios. Flying my usual missions are impressive as Hell.

 

MSAA and SSAA. Frame killers and I cannot say they are worth it. Not sure if they are a placebo or not. SSAA, I may just turn back on to load the card.

 

Pixel Density. This has the most effect on image sharpness, I can see the difference between 1.8 and 2.3, even if it does suffer from diminishing returns and I will bump this back up. No point running the card at 60% chasing frame rates that are simply not there. DCS needs serious optimisation work.

You still need to trade image quality for frame rates, I think this bottleneck is within the DCS engine itself. Again individual choice, I prefer quality over quantity, but then for me, combat is secondary to flying.

 

I7 4790K OC to 4.48ghz. Impressive as Hell, it has taken everything I threw at it, chewed it up and spat it out.

 

 

Hope you found these tests to be as unbiased and as truthful as possible, my intention is to help others make a more informed decision re upgrading.

 

For myself, I am happy overall, maybe a little disappointed in some areas, but those areas are not the fault of the card. The 2080TI is not a panacea for all of ED's ailments.

 

Thanks for looking in.

 

Do you fly IL2 BOS, I`d be interested to see the performance figures from same. BOS in my view is better optimised for VR than DCS and is an exceptional experience in VR, even with the 1070. I am looking at the 1080ti or 2080 as an upgrade but leaning towards the 1080ti due to it`s 11gb vram and cuda core count. It looks like a CV2 could be two years away at which point the 2080ti will be superseded by the next gen gpu`s so I am limiting my gpu investment until we see what the new HMD`s bring to the party. A deal on a 1080ti (new or 2nd hand) will keep me near the top in performance for a limited spend. The 2080ti is just too much money imo for what it can do and the HMD`s we have atm, but best of luck and enjoy yours.

Edited September 30, 2018 by Zoomer

[Tinkickef]

Do you fly IL2 BOS, I`d be interested to see the performance figures from same. BOS in my view is better optimised for VR than DCS and is an exceptional experience in VR, even with the 1070. I am looking at the 1080ti or 2080 as an upgrade but leaning towards the 1080ti due to it`s 11gb vram and cuda core count. It looks like a CV2 could be two years away at which point the 2080ti will be superseded by the next gen gpu`s so I am limiting my gpu investment until we see what the new HMD`s bring to the party. A deal on a 1080ti (new or 2nd hand) will keep me near the top in performance for a limited spend. The 2080ti is just too much money imo for what it can do and the HMD`s we have atm, but best of luck and enjoy yours.

 

I do not have Il2 BoS. I do have 1946 and CLoD, although I have not used either for 2 years or more.

My honest advice between 1080ti TI and 2080 is go for the 1080TI for the reasons you stated.

 

The 2080TI is an astonishing card, but is badly let down by the DCS engine. I finally settled on a PD setting of 2.3 for my every day use!!! and find this fine.

 

When DCS finally goes over to Vulcan and Gen 2 headsets come on stream, that's when the 2080TI will shine.

[Supmua]

Hoping to test a 2080Ti with my Vive Pro soon. I’m curious to see what kind of FPS I would get with my i7 7700k. I find that optimal PD for the Pro is 1.4-1.5 for visuals.

 

Your FPS report is helpful because we need objective measurement not subjective feel. I for one can’t tolerate FPS below 45 in VR, and need to adjust graphic settings to maintain that number.

 

 

Sent from my iPad using Tapatalk

[Tinkickef]

Hoping to test a 2080Ti with my Vive Pro soon. I’m curious to see what kind of FPS I would get with my i7 7700k. I find that optimal PD for the Pro is 1.4-1.5 for visuals.

 

Your FPS report is helpful because we need objective measurement not subjective feel. I for one can’t tolerate FPS below 45 in VR, and need to adjust graphic settings to maintain that number.

 

 

Sent from my iPad using Tapatalk

 

It will be interesting to hear your results with Vive Pro.

Raw resolution via higher pixel count should be much easier to maintain for the same image quality than the software solutions we currently have.

Running 2.3PD for a very tiny difference over 1.8 PD is an horrendously expensive way of doing it.

[Supmua]

My ASUS Dual 2080Ti came in today. I haven't done any serious testing but did OC GPU clock speed to 2000 MHZ, haven't tried to push memory clock speed yet. I found that I could go a little higher with other games, but got BSD with DCS every time I pushed past 2000.

 

I have an Intel 7700K CPU (OC'ed to 4.7GHz). With the Vive Pro and in-game PD setting 1.4, I can get steady and smooth 45+ FPS running mostly high except visibility is medium, shadow is flat. In areas that typically tank my FPS especially khasab Airport in early AM I would get mostly steady 45 FPS (dips to 41 upon touchdown for 1-2 sec). When I fly with just clouds in the background the FPS goes up to 90.

 

Image quality is quite excellent on the Vive Pro at PD 1.4 which minimizes shimmering artifacts, DCS really shines in VR when you have water at high setting with some shadows. This was a huge turn off for me before with my old GPU (1080) having to compromise setting parameters to get acceptable frame rate. It's still not as good as 2D large screen but it's good enough for me now. The only thing that is missing is the sense of speed when flying low that you can feel when playing in 2D, which might have something to do with 45 vs 100+ FPS.

 

I think that I can push the settings even further with better CPU and higher clock RAM, maybe water-cooling this GPU. Hoping to build my next PC by the end of the year with Intel 9900K and some 3600Mhz RAM.

Edited October 4, 2018 by Supmua

[Tinkickef]

My ASUS Dual 2080Ti came in today. I haven't done any serious testing but did OC GPU clock speed to 2000 MHZ, haven't tried to push memory clock speed yet. I found that I could go a little higher with other games, but got BSD with DCS every time I pushed past 2000.

 

I have an Intel 7700K CPU (OC'ed to 4.7GHz). With the Vive Pro and in-game PD setting 1.4, I can get steady and smooth 45+ FPS running mostly high except visibility is medium, shadow is flat. In areas that typically tank my FPS especially khasab Airport in early AM I would get mostly steady 45 FPS (dips to 41 upon touchdown for 1-2 sec). When I fly with just clouds in the background the FPS goes up to 90.

 

Image quality is quite excellent on the Vive Pro at PD 1.4 which minimizes shimmering artifacts, DCS really shines in VR when you have water at high setting with some shadows. This was a huge turn off for me before with my old GPU (1080) having to compromise setting parameters to get acceptable frame rate. It's still not as good as 2D large screen but it's good enough for me now. The only thing that is missing is the sense of speed when flying low that you can feel when playing in 2D, which might have something to do with 45 vs 100+ FPS.

 

I think that I can push the settings even further with better CPU and higher clock RAM, maybe water-cooling this GPU. Hoping to build my next PC by the end of the year with Intel 9900K and some 3600Mhz RAM.

 

I downloaded Evga Precision X1 very late in the testing to see if it would bump up frames (nope still locked to a max of 45fps by Oculus despite asw off)by doing a little overclocking on my Asus Dual. It incorporates the Nvidia scan tool for one button overclocking.

I got a score of 70 after all the adjusting and testing (whatever the 70 means) with a target temp of 84 degrees and 110% power.

It hit temp before power target even with fan curve set more aggressively, so definitely a case there for fitting a waterblock.

 

Interesting to note that during the scan tests, the results differ wildly. I have seen a score of 30 and a score of 90, so I guess you run it repeatedly until you get the score you want and save that setting. I found it important to reset the card to default before the next scan if I didn't want the results to tank.

 

Don't really want, or have the confidence to manually OC the card at this moment in time. Maybe in two or three years when it starts to struggle....

 

Can't wait for Vulcan, the promised VR optimizations and Gen two headsets. Should make life easier.

Edited October 4, 2018 by Tinkickef

[hannibal]

test in caucausus, not in persian gulf?

[Tinkickef]

test in caucausus, not in persian gulf?

 

Look in on the Aurelious Augustus thread. He is a pro and doing a far better job at testing than my amateur attempt.

 

Also, my tests were in response to specific requests for test data on how the 4790k in particular, handled the 2080TI when I revealed I had one on pre order, so everything was conducted in the worst case, CPU intensive scenarios I could think of, including default quick mission generator AI friends and enemies.

I concentrated far more on CPU performance than card performance (excellent card performance in relation to my old 980TI being a given factor), although once Knock Knock showed me how to include frame rates, I did so.

 

Nice to see many of AA's conclusions are the same as mine though.

 

AA's thread is much more professional, generalised, card /headset orientated, and of more worth to the average VR user. Plus he is doing his tests on Caucasus.

 

Time to bow out and let the Pros handle it from now on.

 

:)

Edited October 6, 2018 by Tinkickef