3080 reviews are out

Interesting, where did you read that? Seems very surprising.

But I remembered the numbers wrong. 125 devs in total, 50% working on the core.

Hmm, thanks for that. Interesting.

 

 

But I remembered the numbers wrong. 125 devs in total, 50% working on the core.

Is this ED only or are they counting the third party delvelopers in ?

Did you read some of the back and forth convo between the CEO and redditors. Yes, I have seen that thread and yes they consider themselves small. It is all in there.

Is it me, or did this thread slightly derail? :P

To lazy to read though all 11 pages, I admit, but are there any 3080 DCS user reviews out there yet?

(Preferrably "before and after GPU swap" results)

It seems there is only one person who has it and I did ask for some numbers, just gotta give him time to do his own testing. It is too bad that DCS wasn’t included in all these 3080 reviews from tech sites/channels. Would love to see detailed analysis like MSFS. I might do it if time permits since I have 3 PCs from different generations (small sample size though).

It will have to be standardized enough that is applicable to everybody.

-use of various standard presets (VR vs medium vs high) and PD rather than your own single custom one.

-use of common missions (carrier, ramp, runways, on the air) that everybody has access to and also fly at the various altitudes (high, low, etc.)

-use of reliable fps counter such as fpsvr to report frame rate, frame time of CPU and GPU rather than eyeballing it.

-use of different CPU and GPU in various combinations.

-hyper threading vs non-hyperthreading

-VR vs single flatscreen vs multimonitor

-overclocking (CPU, GPU, RAM) vs default

See why people are confused?

It’s easy to test what your CPU limited frame rate is. Go to a heavy mission which is easily repeatable like a ground start with lots of other aircraft and such. Drop resolution (since this is obviously driven by your GPU) down to a very low value like 720p

The frame rate you see will be what your CPU is limiting you to.

You have to keep this in mind when evaluating graphics cards and settings. Again, I could use a 3080 with my CPU and not see any change in frame rate if I use a test scenario like above.

My GPU frame times are 10-19 my CPU frame times are 7... I'm GPU bound...

My GPU frame times are 10-19 my CPU frame times are 7... I'm GPU bound...

I what situation though? It can vary a lot depending on how many objects are in the scene. A frame time of 7 means you’re getting 144fps from your CPU. Is that in a free flight?

You can be CPU limited in some circumstances and GPU in another. Really I prefer to adjust settings based upon the worst case or the refresh rate of the display since there’s obviously no benefit to going higher than that. Once I know the CPU limit in the worst case, I can turn up settings with that in mind, frankly with a 2080Ti there is no limit, they can all be maxed. Sure there are some cases where I drop below 60fps but it’s not due to the GPU

the only time i am not GPU bound, is when i am in VERY high unit count missions

It seems there is only one person who has it and I did ask for some numbers, just gotta give him time to do his own testing. It is too bad that DCS wasn’t included in all these 3080 reviews from tech sites/channels. Would love to see detailed analysis like MSFS. I might do it if time permits since I have 3 PCs from different generations (small sample size though).

 

It will have to be standardized enough that is applicable to everybody.

-use of various standard presets (VR vs medium vs high) and PD rather than your own single custom one.

-use of common missions (carrier, ramp, runways, on the air) that everybody has access to and also fly at the various altitudes (high, low, etc.)

-use of reliable fps counter such as fpsvr to report frame rate, frame time of CPU and GPU rather than eyeballing it.

-use of different CPU and GPU in various combinations.

-hyper threading vs non-hyperthreading

-VR vs single flatscreen vs multimonitor

-overclocking (CPU, GPU, RAM) vs default

I'll be doing the same with my 3090,if and when it ever arrives. I'm thinking we should come up with a standardised test using the free DCS assets (Frogfoot and map) so everyone can get a good idea of performance?

the only time i am not GPU bound, is when i am in VERY high unit count missions

So if your CPU is capable of 144fps why are you only getting 45?

https://forums.eagle.ru/showpost.php?p=4493295&postcount=3

Wouldn’t it be better to just turn down graphic settings and get rid of reprojection artifacts?

So if your CPU is capable of 144fps why are you only getting 45?

https://forums.eagle.ru/showpost.php?p=4493295&postcount=3

Wouldn’t it be better to just turn down graphic settings and get rid of reprojection artifacts?

You need to check your eyesight tested ... it says 45 not 144 ... I get almost no reprojection artifacts ...

is it possible you are trolling? or really not understanding ? BECAUSE I AM GPU LIMITED...

I'll be doing the same with my 3090,if and when it ever arrives. I'm thinking we should come up with a standardised test using the free DCS assets (Frogfoot and map) so everyone can get a good idea of performance?

Good luck getting it, although I see there are some leaked Benchmarks now floating about, and it doesn’t look to be the FPS increase we were hoping for between the 3080 and 3090. Now obviously we all take leaked benchmarks with a whole mine of salt, but it might just might be true, and if so it makes it significantly harder to justify the extra cost. At least for me anyway. Currently planning on the EVGA ftw3 3090

I'll be doing the same with my 3090,if and when it ever arrives. I'm thinking we should come up with a standardised test using the free DCS assets (Frogfoot and map) so everyone can get a good idea of performance?

Using free and readily available DCS assets sounds good. Throw in various VR headsets and GPUs into the mix. Besides the planned 3090 purchase, I currently have NVIDIA 2080Ti, 1080, 970. I do have a few older GPUs but let em RIP. Fixed preset graphic settings and normalized flying altitude should help eliminate biases as well.

You need to check your eyesight tested ... it says 45 not 144 ... I get almost no reprojection artifacts ...

 

is it possible you are trolling? or really not understanding ? BECAUSE I AM GPU LIMITED...

I don’t mean to come across as trolling. It’s a genuine question. If your CPU frame time is 7, that equates to a frame rate of 144fps. I know frame time is a relevant statistic but frame rate is also.

Are you assuming that your cap is coming from the GPU just on the basis of frame times?

Frame rate is functionally capped by the gpu, it's really straightforward.

Frame rate is functionally capped by the gpu, it's really straightforward.

For a 1080Ti pushing 2160x4320 45 FPS sounds about right. But your CPU performance seems to indicate that with a much more powerful GPU you could indeed hit 90fps no trouble.

Err Yes, I think I said that.

So if a 9700K can (theoretically) hit 144fps in DCS. And in VR. Why bother with Vulkan at all?

Not sure it does, but because you still end up being limited by the "single thread"... And multi thread perf can provide more overall headroom...

Good luck getting it, although I see there are some leaked Benchmarks now floating about, and it doesn’t look to be the FPS increase we were hoping for between the 3080 and 3090. Now obviously we all take leaked benchmarks with a whole mine of salt, but it might just might be true, and if so it makes it significantly harder to justify the extra cost. At least for me anyway. Currently planning on the EVGA ftw3 3090

The thing about leaked benchmarks is that, even if they’re real they still reflect pre-released software state which is likely unoptimized for the new GPUs. It takes both hardware and optimized software components to bring out the full potential of any new GPUs.

I remember back when the 2080 was just released it could barely touched the 1080Ti in performance, even saw lower fps in some games. With subsequent newer GeForce drivers together with developers optimizing their software to accommodate the new GPU architecture, now we can clearly see the 2080 pulling ahead. It will be the same story with the 30 series. The software side will take sometimes to catch up.

So if a 9700K can (theoretically) hit 144fps in DCS. And in VR. Why bother with Vulkan at all?

Not going to hit 144 in VR, there are two images being drawn simultaneously.

Typically VR framerates are in the neighborhood of half what is on monitor.

I assume we are talking VR here...

The thing about leaked benchmarks is that, even if they’re real they still reflect pre-released software state which is likely unoptimized for the new GPUs. It takes both hardware and optimized software components to bring out the full potential of any new GPUs.

 

I remember back when the 2080 was just released it could barely touched the 1080Ti in performance, even saw lower fps in some games. With subsequent newer GeForce drivers together with developers optimizing their software to accommodate the new GPU architecture, now we can clearly see the 2080 pulling ahead. It will be the same story with the 30 series. The software side will take sometimes to catch up.

Oh fully appreciate that, hence why I stated take with a mine of salt. Still extremely tempted if only for the 24gb VRAM.

But certainly from the 4 leaks I’ve seen there is certainly a trend of only around 10 to 15 extra FPS at the mo. Which IF true is gonna be hard to justify the extra cost.

Really hope we get some official stuff soon. I mean the 3080 will be a huge improvement over my 1080ti, but I am concerned about the 10gb VRAM. Not totally convinced it’s enough.

All I want is 75 to 90 FPS in VR so i can finally reach the recommended framerates of the Oculus. I just want to be able to fly without ground stutter driving me nuts.