1 hour ago, Bjoern said:

If framerate determined market share, Aerofly FS2 would have like 80% of it. But it doesn't.

It is _one_ of market share determinants, and with the probable rise of VR usage and the more competitive landscape with MFS possibly having a new and performing graphic engine, performance could become more and more important of a determinant.

9 minutes ago, Murmur said:

It is _one_ of market share determinants, and with the probable rise of VR usage and the more competitive landscape with MFS possibly having a new and performing graphic engine, performance could become more and more important of a determinant.

A very high frame rate is important for VR, yes, but I think of that as a compartmentalized feature that doesn't matter to the vast majority of the flight simulator market. We now have almost a year's data on the live XP survey of users who have ever flown in VR in 2019, and the percentage is still hovering around just 2% of users. The numbers will be higher in something like AFS2, but that sim makes tremendous compromises compared to FSX, P3D, or XP to get frame rates that high.

I'm also of the opinion that frame rate "matters" more to those who don't have sufficient hardware to run a flight sim with reasonable eye candy at a reasonable frame rate, say 35 to 40 fps. Some of these folks are looking for a magic bullet from the sim developers, so they don't have to spend more money on hardware.

If you own a state-of-the-art gaming PC with a fast CPU, tons of RAM, and one of the high-end Nvidia RTX series cards, you're probably not going to care that much if Laminar "only" manages to get another 10 fps improvement in the next major update. 

Regarding the new MSFS, I'm not ready to assume that it's a "new and performing graphic engine" until I and everyone else get their hands on it. There is a tremendous amount of hype for this sim generated by MS marketing savvy, and we have no idea how it will actually perform.

Let's see what the actual frame rates are, at medium eye candy settings, before assuming that XP has some major catching up to do. I think people may be in for a surprise. There is no free lunch and no magic programming tricks, when it comes to balancing frame rate with eye candy.

Actually there is a lot of free lunch, because XP is almost always CPU limited, and that is (by Ben himself admission) because CPU has still a lot of load that could be transferred to GPU, and that is what Vulkan will be for, and also because it can be further optimized for multicores.

I don't know how hyped MFS is, I know that for example having those clouds with the performance reported by early testers has never been seen in a flight sim.

From what I understood, the FPS increase is is just from swapping from OpenGL to Vulkan (very simplistically) to get the image looking exactly the same. Ben has noted there are new  ways of doing things and optimisations they can do once stable. So more FPS may come out of the Vulkan switch yet. 

If it's a 10fps increase for free then this is a welcome win in my book. There is not so much difference going from 30-40fps, but 15 to 25 or 20 to 30 is quite significant for a smooth flight. The other important factor that Vulkan will provide is hopefully far less stuttering (Especially when panning the view). Since I use and AMD card, hopefully I'll see the biggest improvement, although I'm not expecting it will have AF2 performance.

The other important thing to remember is that this gives LR a better foundation to start adding new next-gen features and improvements to the sim which we've all been waiting for.  

Of course the real metric is not fps but % increase. In the last presentation Philipp showed graphs (for nVidia) with performance going from 60 to 70 at medium settings, and from 28 to 35 at extreme settings. Certainly better than nothing but hopefully there's still headroom for improvement. Even considering that XP needs new visual effects (weather and not only) to remain competitive, and performance headroom would be useful for that.

You realize that 28 to 35 is a 25% increase? And this "certainly better than nothing"? 🤣

from the tweet charts AMD is a little better than Novideo with Vulkan.   

 

Edited October 23, 20196 yr by HumptyDumpty

16 hours ago, Murmur said:

Actually there is a lot of free lunch, because XP is almost always CPU limited, and that is (by Ben himself admission) because CPU has still a lot of load that could be transferred to GPU, and that is what Vulkan will be for, and also because it can be further optimized for multicores.

Microsoft and the ASOBO devs have said that they're building the new sim as an evolution of FSX, and not something completely new from the ground up. They may be in a similar situation of having to tweak legacy code to move more of the processing onto the GPU. We just don't know yet.

Anyway, I'm not ready to assume yet that the MSFS is some high performance miracle of programming compared to X-Plane, until more people get their hands on it, and we can discuss actual frame rate comparisons.

8 hours ago, Janov said:

You realize that 28 to 35 is a 25% increase? And this "certainly better than nothing"? 🤣

Yes, better than nothing but still possibly not enough given the probable competitive landscape. 🤣

Also Ben himself suggested 60 fps as a target for 2D (single monitor). Given the shown graphs, I can assume those 60 fps as a target are referred to max settings (or very close to). So still a way to go.

I shouldn't need to remind that XP is also the sim that currently receives the most complaints about performance in VR, compared to others.

Edited October 23, 20196 yr by Murmur

Just out of curiosity, just learned from a recent interview with Matt Wagner ( ED  / DCS ) that DCS is also moving to Vulkan + Metal in it's next graphics engine step.

OTOH, one aspect that really puzzles me is how performant and graphically gorgeous simulators like IL-2 Great Battles can be, still relying on good-old DirectX.

Then there's War Thunder, with great graphics and huge performance, and it's own graphics engine - some say rather complex for programmers, but excepcionally portable since WT runs muti-platform like no other PC-based flight simulator I have used before.

I really don't know what Vulkan ** really ** is, have to Google more for it, but I just wish this civil sims could perform at least 80% as well as WT, and even IL-2, do.

And please don't throw that falacy of "the amount of stuff" they do and the combat sims don't, because if we go that track then it's even more of a shame for the civil sims - well AEFS being a notable exception of course!

Edited October 24, 20196 yr by jcomm

21 minutes ago, jcomm said:

Then there's War Thunder, with great graphics and huge performance, and it's own graphics engine - some say rather complex for programmers, but excepcionally portable since WT runs muti-platform like no other PC-based flight simulator I have used before.

Isn´t it time to UNINSTALL that game soon? 🤣

14 minutes ago, Janov said:

Isn´t it time to UNINSTALL that game soon? 🤣

Was a second about to, after watching Austin's latest video and thinking - THIS IS IT ! - but wife called for dinner, and I didn't have time to hit the uninstall button 😕

Eheheh, joke apart, WT is here to stay - it's a great fun to play 🙂

Edited October 24, 20196 yr by jcomm

On 10/23/2019 at 5:12 PM, HumptyDumpty said:

from the tweet charts AMD is a little better than Novideo with Vulkan.   

 

Not much of a surprise.
 

Quote

Vulkan is derived from and built upon components of AMD's Mantle API, which was donated by AMD to Khronos with the intent of giving Khronos a foundation on which to begin developing a low-level API that they could standardize across the industry.

https://en.wikipedia.org/wiki/Vulkan_(API)

Edited October 24, 20196 yr by Bjoern