Putting the cart before the horse

[Jimm 437]

Those of you out there, who aren't familiar with this idiom, typically refers to something that is done contrary to a conventional order or relationship.  When XP11 was released, Ben Supnik stated that they would now have the time available to address performance, as well as other glaring issues with XP11, that has been present in the Betas, but never fully resolved.  In a couple of recent posts on the XP Developer Blog, Ben "teased" the public with narratives about VR.  Now I have no personal problem with that, as it is another feature that many are wanting in XP, but as the title of this post suggests, wouldn't it be more prudent to put all of the development team in full-time status to make sure that the more important aspects of the sim are nailed down before tackling VR?

After XP11 was released, I took the time to fire up the sim to do  a little more testing.  While I still found the performance to be stable and the framerates somewhat on par from the previous Betas, I did discover that the default weather still needed some work.  I know I]m part of a small group who insisted on using the default weather in the Betas, and rightfully so, as it was Beta and the weather needed testing, just like everything else.  It wasn't until near the end of the Beta phase, that the cloud rendering took a complete 180, and started to become a good competitor to the other 3rd party devs out there.  I reported my findings and issues (framerates and fog/haze layer), with hopes that this too would be looked at further to be optimized or worked on a little more.  This past weekend, I took the Aerobask Panthera out for a spin, and while I can't say enough good things about the aircraft, I can say that general XP performance has diminished a bit more than before, and perhaps the aircraft has something to do with it, I tend to believe it is more about the simulator.  I never changed my settings from what they were before, as those settings appeared to provide me with good stability, all-around good performance from my cpu and gpu, and above average frames.  I should note though that my recent testing wasn't as in-depth as it could have been, with flights in a default aircraft, some flights with and without weather, to contribute more data for analysis, but considering this greater drop in performance now, i have concern on whether LR is still on track to address the issues we all have reported or they are going to call it good and move on.  If the default weather stays the way it is, then I may have to shop around for better weather generation, and this is where you all come in.

I'd love to hear about how XP11 is performing on your systems, with or without 3rd party weather, what aircraft you are using (default or payware), and maybe get some insight into what, if any, could be the problem on my end.  With that, let me know if you continue to file bug reports to LR, or even if you've had the honor of speaking with Ben directly.

As for LR, rolling out new features isn't a bad idea, but I do think it wise to ensure the sim is performing the way it should be, before tackling new things.

[alpha12125 213]

well the current status is that chris is doing the basic vr integration, ben is doing performance/bugfix backlog, phillip is doing g1000/fmc/gns and austin is doing atc.

Personally i feel this is a great mix of fixing stuff / adding new stuff. It is one new thing (vr) and 3 "old things". Performancewise i run 6700k@4.4, 32gigs, ssd and 1080  at around 3-40ish frames at la guardia with mesh v3 and pretty much full settings. For me that performance is enough to satisfy all my needs for now. Apart from bragging rights, another 20fps would not help me alot.

I think weather is not ideal in default, but its ok. The question will always be, how much effort is enough? And who does it help? No matter how much work you put into it, you will never get as16 as default and people will still use xenviro or smp. For me the current way they are doing things is fine and alot more oriented to what we want then ever before

[Murmur 3,941]

I think VR implementation and performance improvements are not disjointed. In his blog, Ben hinted that a big part of VR implementation will include improving performance in terms of FPS and smoothness, and his work will be focused on this for the near future. I assume further improving the performance of clouds will be part of this as well. So at the end of the day this should benefit even non VR users.

 

[rototom 198]

A G1000? Yes please, then Carenado can swap the one in the SR22 because I can't get anything to work and it freezes constantly.

[Dumont 9]

2 hours ago, Jimm said:

Those of you out there, who aren't familiar with this idiom, typically refers to something that is done contrary to a conventional order or relationship.  When XP11 was released, Ben Supnik stated that they would now have the time available to address performance, as well as other glaring issues with XP11, that has been present in the Betas, but never fully resolved.  In a couple of recent posts on the XP Developer Blog, Ben "teased" the public with narratives about VR.  Now I have no personal problem with that, as it is another feature that many are wanting in XP, but as the title of this post suggests, wouldn't it be more prudent to put all of the development team in full-time status to make sure that the more important aspects of the sim are nailed down before tackling VR?

After XP11 was released, I took the time to fire up the sim to do  a little more testing.  While I still found the performance to be stable and the framerates somewhat on par from the previous Betas, I did discover that the default weather still needed some work.  I know I]m part of a small group who insisted on using the default weather in the Betas, and rightfully so, as it was Beta and the weather needed testing, just like everything else.  It wasn't until near the end of the Beta phase, that the cloud rendering took a complete 180, and started to become a good competitor to the other 3rd party devs out there.  I reported my findings and issues (framerates and fog/haze layer), with hopes that this too would be looked at further to be optimized or worked on a little more.  This past weekend, I took the Aerobask Panthera out for a spin, and while I can't say enough good things about the aircraft, I can say that general XP performance has diminished a bit more than before, and perhaps the aircraft has something to do with it, I tend to believe it is more about the simulator.  I never changed my settings from what they were before, as those settings appeared to provide me with good stability, all-around good performance from my cpu and gpu, and above average frames.  I should note though that my recent testing wasn't as in-depth as it could have been, with flights in a default aircraft, some flights with and without weather, to contribute more data for analysis, but considering this greater drop in performance now, i have concern on whether LR is still on track to address the issues we all have reported or they are going to call it good and move on.  If the default weather stays the way it is, then I may have to shop around for better weather generation, and this is where you all come in.

I'd love to hear about how XP11 is performing on your systems, with or without 3rd party weather, what aircraft you are using (default or payware), and maybe get some insight into what, if any, could be the problem on my end.  With that, let me know if you continue to file bug reports to LR, or even if you've had the honor of speaking with Ben directly.

As for LR, rolling out new features isn't a bad idea, but I do think it wise to ensure the sim is performing the way it should be, before tackling new things.

I agree with Jimm.
I use translator to post and often car comes in front of horse.
It was talked about performance issues, weather, well I bought X-Plane 11 so I want something good from what I bought, if I can improve it by spending more is another story.
At the moment I only use default aircraft, alias I use X-Plane 11 practically 100% default, except for X-RAAS, I already sent Bug Report on problems of 747-400, it's been over 30 days, so I was very surprised , Reported problems have not been fixed, they are now in larger numbers.
So we have a system that needs to be finished or get corrections of the basic but before that we speak in implementations, at least exotic for the situation, this is putting car in front of the horse.

 

João Alfredo

[propsfan 52]

3 hours ago, Jimm said:

  Now I have no personal problem with that, as it is another feature that many are wanting in XP, but as the title of this post suggests, wouldn't it be more prudent to put all of the development team in full-time status to make sure that the more important aspects of the sim are nailed down before tackling VR?

As for LR, rolling out new features isn't a bad idea, but I do think it wise to ensure the sim is performing the way it should be, before tackling new things.

When you code you own sim, you'll also have the luxury of deciding what's prudent and wise. Until then you can do what everyone else does and wait patiently for what you want. Anyone who's been around XP for a time knows this is your only realistically available option.

You have the best cpu/gpu combination for XP so you should be getting excellent performance. You'll have to adjust your expectations I'm afraid.

[Paraffin 2,197]

Just agreeing with the others above, that work on VR looks to be distributed in a way that isn't delaying other performance improvements. I don't use VR myself, but I do think it's important that X-Plane have native VR support, and soon, just to maintain market position and reputation as a top-line flight sim. Like it or not, VR is here to stay, at least for niche applications like flight sims. 

I think we also need to be realistic about just how much future improvement we can expect for us flat screen users. For what it's worth, I'm getting a very comfortable average of 40-45 fps most places I fly, and never below 30-35 fps over dense urban areas. That's on a single 1920x1200 monitor, with the hardware specs listed in my sig below. I'm running graphics at High(HDR) and Maximum world objects, low reflections and no shadows on world objects.

I'd need a stronger GPU if I was running higher resolution, but I'm happy with what I'm getting here. Frankly, I consider this to be amazing performance considering how great everything looks in XP11, compared to XP10 and earlier versions. Although of course if higher performance is possible I'd welcome it, because then I could turn ground shadows on. Meanwhile, it looks like I can delay thinking about a 1080 series card for a while longer.

One more thing: When doing frame rate tests, the aircraft model can make a BIG difference. Some are more CPU/GPU hungry than others. For example, the Carenado PC-12 eats around 5-10 fps more on my system than something basic like the default C172, probably due to the more complex cockpit displays and plugins loaded in the background. So it's important to use something like the C172 to see what your baseline frame rate is, and then judge each aircraft's performance in relation to that. 

Even in a 64-bit sim, you can't expect to run a complex aircraft with a cockpit full of bells and whistles without some hit on the frame rate. That's one thing that doesn't change when comparing X-Plane to FSX/P3D or other flight sims.

[CarlosF 1,235]

Don't forget Vulkan integration is a reality and that itself will improve performance considerably, a more efficient use of CPU/GPU resources is a welcome change.

 

[Mountain Man 142]

I don't think they've put the cart before the horse.  I think the cart is exactly where it should be.  XP11 was in fine shape for a "gold" candidate -- stable, solid performance, no show-stopping bugs -- and now that the features are locked down, they can proceed to the next phase of development.

I suppose you could argue that they should have locked the features and done the final spit and polish during beta, but that's really not how software development is done any more.  This seems to be the "new normal" across the industry and has been for many years.

[Slai 10]

No dev team is scheduled to clear out accumulated tech debt unless it can be tied to a commercial deliverable.  VR is a perfect opportunity to do so and you can sense in Ben's blog that he is excited to do so.  

I am enthused about this imitative because VR has a way of amplifying the sense of realism such that if there are issues, it causes people to throw up.  My hope is once the sim is good for VR, it will mean that the "smoothness" (Ben's word and much more important than FPS) fixes will give that sense of realism, air and not flying on rails for both VR and regular monitors.  Only Laminar can bring this to the sim.  After that, provide a solid API and leave most of the rest to third parties.