Funny how it's always something else. The 727 (an insanely well performing plane) takes to much, Skymaxx takes too much, it's a common mistake, it's normal that performance tanks when you run a flightsim... It's definitely not XP that is unable to make use of contemporary hardware 🙂

Except this all is not true when you run that other flightsim.

Isn't that a bit contradictory?

44 minutes ago, rka said:

Isn't that a bit contradictory?

Maybe you missed the last post of Janov?

2 minutes ago, Murmur said:

Maybe you missed the last post of Janov?

I was referring to several of the other comments on the previous page, maybe you missed them?

Edit: It is of course a good thing that Janov can see it now, did you notice when that happened?

Edited September 11, 20214 yr by rka

1 hour ago, rka said:

It is of course a good thing that Janov can see it now, did you notice when that happened?

I saw that for a long time and never argued against it - so there is no need for your (or your sidekicks) instigation at all.

No one ever doubted that "the other sim" looks better than XP11 - and now (after SU5/6) also outperforms XP11 in the "graphics per framerate" metric. There is no contest and no surprise, considering the age of both competitors.

16 hours ago, Randall said:

I believe that’s exactly what they’re planning for Xplane 12. 

So I am a little biased towards MSFS, I guess I will clarify that up front because I rarely post here in the X-Plane forums.

But since I also come from a software development background, I am very fascinated at how LR will respond to MSFS's latest changes.  For those who haven't kept up with MSFS, in Sim Update 5, there was a big boost in FPS for everyone because Asobo found a way to optimize MSFS further.  Now there was a lot of complaining when Sim Update 5 was released because some of the FPS gains were from culling the scenery which lead to pop-ins when you panned the camera around.  However, in the recent World Update 6, Asobo added a pre-cache terrain slider so people could eliminate the culling by setting the slider to Ultra, effectively giving users of MSFS a way to go back to caching scenery that wasn't being viewed, which was how it before Sim Update 5.  Suffice to say, even if you set the pre-cache terrain slider to Ultra, people are still getting higher FPS and a more smoother experience than before Sim Update 5 (because Asobo probably made some changes to the underlying architecture of MSFS).  So right now, the rendering is very smooth in MSFS and almost everyone is getting high FPS.

So it will be interesting to see how LR matches this with X-Plane 12.  As you may know, usually if a developer tries to raise the graphics in a game, the FPS will typically suffer as a result.  I am going to guess that X-Plane 12 is using older architecture and has likely accumulated a lot of "technical debt" over the years. Because of the technical debt that X-Plane has accumulated, this is the challenge for LR, if LR wants X-Plane 12 to match MSFS in both graphics and FPS.  Most likely, to match MSFS in both FPS and graphics, because of the technical debt accumulated, LR would probably need to do an overhaul of its architecture.  However, overhauling the whole X-Plane architecture is a huge project in itself, and costs money and/or time.  

I think one of the reasons MSFS had such a huge boost in FPS for Sim Update 5 is because it has a lot less technical debt, because the architecture is much newer.  If the architecture of MSFS is not refactored over the next 10 years, by 2030, MSFS will have accumulated a lot of technical debt too and if there is a new flight simulator on the block in 2030 with new architecture, MSFS will be in a similar situation as X-Plane today.  Having said that, what Asobo managed to do for Sim Update 5 to increase the FPS was very impressive, something that I never expected, especially because the PC version of MSFS is still on DirectX 11 and the X-Box version of MSFS is on a DirectX 12 wrapper I believe, so it's possible that when MSFS fully transitions to DirectX 12, there can be further optimizations with DirectX 12.

So I will be very interested to see how LR attempts to overcome this challenge of matching MSFS in both graphics and FPS for X-Plane 12.  Will LR be only able to close the gap in graphics with MSFS but the FPS will mostly be the same as X-Plane 11?  I think this is the more likely outcome.  But if LR is able to close the gap in graphics with MSFS and also raise the FPS at the same time, that will be the most unexpected outcome!  I guess we will see when X-Plane 12 is released!

I will say one thing - if you are really cheering for X-Plane 12 to match MSFS in both graphics and FPS, it's better to hope X-Plane 12 takes longer to come out.  The reason is, if it takes longer for X-Plane 12 to come out, that is a sign that LR is seriously redoing the underlying architecture of X-Plane.  The faster that X-Plane 12 comes out though, the less likely that LR decided to redo the underlying architecture of X-Plane.

(PS. Please note I'm not here to bash X-Plane in this writeup, rather, because I have a software development background, I am very fascinated with the technical challenge LR has and how they will overcome it - I guess we will see).

I can agree with your post, abrams_tank!

It is the same situation one faces when upgrading an old house. Renovating it is often more expensive than a complete rebuild.

Laminar does not have the resources (money/data/servers...) at hand that Microsoft can pour into MSFS, so it will not be able to match it face to face. It is an asymmetrical engagment, so to say.

XP12 will not try to match or outperform MSFS in it´s strongest trait - graphics. It will pull close in areas it can (lighting, weather, trees,...), but it can´t and won´t match the satellite and photogrammetry data.

The performance that MSFS is showing with SU6 is nothing short of top notch, I guess - and Asobo needs to be congratulated for pulling it off. Stunning visuals at smooth, fast framerate. It sounds as if they are nearing the limit of what is physically possible with todays hardware, so - short of reinventing the laws of physics - Laminar won´t be able to better that.

X-Plane´s strong points are also what is holding it back - one can call it technical debt, I would call it feature inertia. You can´t accelerate quickly if there are a lot of features and backward compatibility to consider with every change. So while an abundance of third-party aircraft and scenery is a "selling point" for X-Plane, it is at the same time often curbing radical development moves.

My guess is that the introduction of X-Plane 12 will be somewhat of an anticlimactic process. They will show some preview and progress videos, but there won´t be a big media hype and parade of youtubers screaming with glee like we saw a year ago. The real innovation will not be of the "OMG it´s such a pretty sunset" type - and for most regulars it will feel comfortably familiar and they can carry on with business as usual (just with a bit of a prettier view outside of the windows).

On 9/8/2021 at 4:10 PM, jcomm said:

I use X-Plane 11 now because as much as MFS can look great scenery wise, and even weather-depiction wise ( not consistency ... ) I still prefer the way I interact with the cockpits of my preferred XP addons, I still prefer by far the flight dynamics and ground physics in XP, I even still prefer ( graphics / visuals not taken into consideration ) XP's default weather over MFS's default weather...

You have so eloquently put into words how I feel.  Thank you so much for doing this.

3 hours ago, abrams_tank said:

So I am a little biased towards MSFS, I guess I will clarify that up front because I rarely post here in the X-Plane forums.

But since I also come from a software development background, I am very fascinated at how LR will respond to MSFS's latest changes. 

Also a software engineer and in aviation, I wrote something very similar here a couple days ago.  There are so many "core" things to contend with here it would also seem to require a complete engine overhaul for X-Plane and I don't know how it is possible to start fresh when you have tens of millions of lines of interdependent code built up over decades, and a small team / budget.

Think about all various plug-in subsystems, object models, rendering pipelines, calculations, texture formats, SDK, etc. a developer would say "we need to move this into how we'd create it in 2021".

While Vulkan can get you the low-level graphics portion of this, that is still dependent on how everything else is written and talks with everything else in core rendering loops.

MSFS had the luxury of throwing hundreds of millions of dollars and hundreds of software engineers to tackle this problem in exactly that way.

I'm rooting for them but I have no idea how this can come together.

Edited September 11, 20214 yr by Gulfstream

7 minutes ago, Gulfstream said:

Also a software engineer and in aviation, I wrote something very similar here a couple days ago.  There are so many "core" things to contend with here it would also seem to require a complete engine overhaul for X-Plane and I don't know how it is possible to start fresh when you have tens of millions of lines of interdependent code built up over decades, and a small team / budget.

Think about all various plug-in subsystems, object models, rendering pipelines, calculations, texture formats, SDK, etc. a developer would say "we need to move this into how we'd create it in 2021".

While Vulkan can get you the low-level graphics portion of this, that is still dependent on how everything else is written and talks with everything else in core rendering loops.

MSFS had the luxury of throwing hundreds of millions of dollars and hundreds of software engineers to tackle this problem in exactly that way.

I'm rooting for them but I have no idea how this can come together.

Yup, it's definitely the challenge for LR, for everything that you mentioned.  Let's see how the tackle it!  It will be exciting to see the result! Cheers.

Edited September 11, 20214 yr by abrams_tank

1 hour ago, Janov said:

My guess is that the introduction of X-Plane 12 will be somewhat of an anticlimactic process. They will show some preview and progress videos, but there won´t be a big media hype and parade of youtubers screaming with glee like we saw a year ago. The real innovation will not be of the "OMG it´s such a pretty sunset" type - and for most regulars it will feel comfortably familiar and they can carry on with business as usual (just with a bit of a prettier view outside of the windows).

I completely agree with this assessment on doing "everything they can" can only get them so far by a +1 increment in the version number due to all these tangled legacy pieces.

From a business standpoint, anything anti-climactic is going to be a tough blow, but I am sure they understand this and see it coming.

You can dive into the CalculateAtomosphericRefraction() stack, try to nudge in something like TruSky (if are even open to third-parties to lessen the burden), adding windscreen rain effects  ... but re-doing textures, auto-gen distribution logic, working with new aerodynamic math, determine why the simulator is now crashing to desktop on a specific Mac when running a particular joystick, etc

This really requires two things.  Top-tier engineers and enough of them to cover the "to do" list.

They have the first part, the second part is where it gets tricky.

For the record I'm someone who purchased MSFS months ago and I am still unable to run it on my gaming PC.  It runs every AAA game without a hitch and is even my primary development machine.  It's been rock solid since I bought it.

It crashes to desktop instantly with "Fly Now!".  The most recent comment from engineering (I am finally working with someone who can work with me as a senior software engineer directly) was "we made a lot of improvements in SU6, please patch to that and let us know".

It still crashes to desktop. 

How is this relevant to X-Plane?  It's an example of how absurd handling everything in these complex beasts can be, and then having to handle the sheer insanity of the related support requests.

Edited September 11, 20214 yr by Gulfstream

To proceed with the performance vs graphics discussion, just to understand what IMO the current bottleneck in Xplane is;

Vulkan, besides it is faster, is much more predictable.

If I'm a scenery author, I know the exact rules in which Xplane manages texture VRAM, unlike OpenGL/DX11 which has lots of non-controllable factors. Hence I can actually calculate almost the exact budget that I have.

If I'm a plugin author, there is a window in Xplane that shows EXCATLY how much time it takes to do it's job.

So it's up to scenery and plugin authors to optimize their work, and the tools are with out a single doubt available, frankly one of the best I've seen.

As for Xplane out of the box performance, If you actually test the settings and watch numbers, most of us are actuallt CPU bound while our GPUs set there on idle; This is due to old rendering pipelines, the clouds are billboard based which is not efficient these days compared to other options, the trees as well, the reflections etc..

Putting more work on the GPU with fancy shaders will also mitigate the need to use the CPU as much  just like the latest tree tech they showed.

Xplane now is like a house (rendering engine) that was renovated from the ground up including foudations overhaul but still using the same old ugly furniture (rendering pipelines and fancy shaders). 

So to describe Xplane as an old archticture is not even remotly true.

20 hours ago, fogboundturtle said:

RTX 3080Ti on Vulkan

That should definitely yield more than 28 FPS. How much are you getting at the same place without any third party add-ons whatsoever?

 

14 hours ago, alexcolka said:

I would say yes. For example the FF Boeings got new life with those sounds.

I've risked a membership and downloaded the sounds for the A321. Let's see...

I would never fly XP11 in a Vanilla state. no one does either. 

Good to see that in one year, people went from steadfastly stating: "Another nail in the coffin for XP!!!1!11!!1" (even before the competing sim was released!) to: "Let's see how XP12 will try to compete".

One year later, XP11 still has more average users than it had right before the temporary surge due to the pandemic. It's very likely that XP12, in time, will get more users than XP11 currently has (infact, in addition to current XP11 users, there will be new users who have now lost interest in XP11, but will likely buy XP12). Will XP12 have more users of MFS? Most likely not, but neverthless, time will once again vindicate the ridiculous "another nail in the coffin" statements made one year ago.