90 FPS, 45 FPS, 23 FPS...I don't see the difference

[Dojo]

Good question.

3dmark?

 

No. I have no interest in synthetic benchmarks, other than fun reads on Tom's/Anand's.

 

I'm re-baselining my machine at the moment, using some handpicked "scenarios" in DCS. I'll then re-run them once I've switched cards later today.

 

In the training mission we most frequently play in DCS with my group (a heavy Nevada mission), the largest bottleneck for the Oculus is the VRAM required. Our scenario requires about 9.1GB VRAM on average for the Oculus, under a standard config with 1.7 pixel density.

 

The dropped frames can directly be tied to that. (i.e. an A-10C empty Nevada mission only uses ~5.5GB of VRAM, and is solid at 45fps on my machine, regardless of where I fly). I imagine the 1080Ti, at a minimum, will solve that easily with its 11GB, preventing the machine from leveraging with the much slower system RAM in this scenario.

 

The Caucasus are a completely different consideration, due to the terrain engine differences. It doesn't play as nicely with new shiny hardware as Nevada does.

[hansangb]

I've always suspected the VRAM being a bottleneck (and why I went with Titan X over 980 ti (back in the days). It's also why 1080ti is appealing to me because it's 11GB (1GB short, and smaller L2 cache) but with a faster clock.

[fitness88]

No. I have no interest in synthetic benchmarks, other than fun reads on Tom's/Anand's.

 

I'm re-baselining my machine at the moment, using some handpicked "scenarios" in DCS. I'll then re-run them once I've switched cards later today.

 

In the training mission we most frequently play in DCS with my group (a heavy Nevada mission), the largest bottleneck for the Oculus is the VRAM required. Our scenario requires about 9.1GB VRAM on average for the Oculus, under a standard config with 1.7 pixel density.

 

The dropped frames can directly be tied to that. (i.e. an A-10C empty Nevada mission only uses ~5.5GB of VRAM, and is solid at 45fps on my machine, regardless of where I fly). I imagine the 1080Ti, at a minimum, will solve that easily with its 11GB, preventing the machine from leveraging with the much slower system RAM in this scenario.

 

The Caucasus are a completely different consideration, due to the terrain engine differences. It doesn't play as nicely with new shiny hardware as Nevada does.

 

 

 

I like practical scenarios where I can actually see how/where the new hardware benefits actual game play. Since I don't have Nevada, I use the open water of the Black Sea as a baseline, I pick Sochi and various areas to do more comparisons. Ultimately I do max test with multiple planes and snow/rain.

[Dojo]

I like practical scenarios where I can actually see how/where the new hardware benefits actual game play. Since I don't have Nevada, I use the open water of the Black Sea as a baseline, I pick Sochi and various areas to do more comparisons. Ultimately I do max test with multiple planes and snow/rain.

 

Indeed. Finding common stress points in your actual play is the best measure.

[Dojo]

I've pulled this thread off track enough, so just some final notes, and I'll create a separate thread when ready:

 

Disabling ASW had a significantly positive effect on my performance, including FPS. Noticeable decrease in judder in 2.0.5. I've been recording observations sporadically through out the day; I'll publish in the new thread once I have the new card (hopefully tomorrow, as weather here in the North east delayed delivery) and can complete my "real world" tests.

 

I'm sure it helps many machines, and many games, but as of this post, ASW is currently no good on my rig for DCS.

[Cibit]

I like practical scenarios where I can actually see how/where the new hardware benefits actual game play. Since I don't have Nevada, I use the open water of the Black Sea as a baseline, I pick Sochi and various areas to do more comparisons. Ultimately I do max test with multiple planes and snow/rain.

 

Indeed. Finding common stress points in your actual play is the best measure.

 

I have found for testing that a track works well. I have a quick flight in a huey record the track and use it in conjunction with fraps MinMaxAvg to see how different settings impact performance. The thing about a track is reducing variables in a scenario to a minimum

[fitness88]

I've pulled this thread off track enough, so just some final notes, and I'll create a separate thread when ready:

 

Disabling ASW had a significantly positive effect on my performance, including FPS. Noticeable decrease in judder in 2.0.5. I've been recording observations sporadically through out the day; I'll publish in the new thread once I have the new card (hopefully tomorrow, as weather here in the North east delayed delivery) and can complete my "real world" tests.

 

I'm sure it helps many machines, and many games, but as of this post, ASW is currently no good on my rig for DCS.

 

 

 

BoxxMann's Tips -

• CTRL+KP1 - Disable ASW and USE ATW
  
• CTRL+KP2 - Force apps to 45hz, DISABLE ASW
  
• CTRL+KP3 - Force apps to 45hz, ENABLE ASW
  
• CTRL+KP4 - Enable ASW to operate automagically
  

I'd like to see a break down of these four options...what does what?

[Dojo]

I have found for testing that a track works well. I have a quick flight in a huey record the track and use it in conjunction with fraps MinMaxAvg to see how different settings impact performance. The thing about a track is reducing variables in a scenario to a minimum

 

I like the idea, but I've had horrible luck with tracks. I fly, take off, and land, and in the track, I crash shortly after takeoff. Last time I tested was with the Viggen in beta. But you're right, it's worth a re-try for my testing style.

 

I'd like to see a break down of these four options...what does what?

 

Those key combos weren't working for me today, so I had to download the SDK, and use the debug tool to shut it off.

 

Here's an explanation directly from Oculus on the core concept: https://developer.oculus.com/blog/asynchronous-spacewarp/

[Cibit]

The link to the tray tool I posted earlier in the thread allows for the option of disabling ASW or other settings

[fitness88]

I like the idea, but I've had horrible luck with tracks. I fly, take off, and land, and in the track, I crash shortly after takeoff. Last time I tested was with the Viggen in beta. But you're right, it's worth a re-try for my testing style.

 

 

 

Those key combos weren't working for me today, so I had to download the SDK, and use the debug tool to shut it off.

 

Here's an explanation directly from Oculus on the core concept: https://developer.oculus.com/blog/asynchronous-spacewarp/

 

 

 

Thanks, I took a quick look...requires a long look.