5 minutes ago, MattNischan said:

I'm going to have to defer to the experts, I'm afraid. I may be mistaken, so best to wait for an answer on the official forums.

Ahh, cool, thanks Matt 👍.

52 minutes ago, turbomax said:

 due to better thread optimisation

Hold on, was multi-threading optimization in the cards for SU15? I only know of memory use optimizations and glass cockpit avionics optimizations using the Coherent/Javascript tech stack (unless you mean thread optimizations around the latter, I don't recall the details).

I know general multi-thread optimizations are planned for MSFS 2024 as Seb says here at the FSExpo presentation: https://youtu.be/VPhScg_FINE?t=718 (i.e. FDE/physics engine and CFD calculations, etc). That along with their push for thinner clients and off-loading more computing to the server-side is going to improve performance overall even more, and that's important to handle the increased/complex visual fidelity and physics coming in 2024.
 

Edited February 15, 20242 yr by lwt1971

5 minutes ago, lwt1971 said:

I know general multi-thread optimizations are planned for MSFS 2024

I thought this was an ongoing process, you may be right though.

10 minutes ago, lwt1971 said:

Hold on, was multi-threading optimization in the cards for SU15?

I don't think so. IIRC improved multithreading was introduced in SU10 w/ DX12 and has (hopefully) continued to improve gradually since then. 

The only flaw I've found so far in SU15 is the reapperance of the infamous DX12 broken ground textures. Which could be related to memory improvements, as was the case with SU10, if my memory is correct. 

Edited February 15, 20242 yr by Cpt_Piett

Allright, so transplanting the new ground handling to the RV-14 did not work at all. In fact the aircraft was all over the place on takeoff and landing. 

7 minutes ago, Cpt_Piett said:

Allright, so transplanting the new ground handling to the RV-14 did not work at all. In fact the aircraft was all over the place on takeoff and landing. 

 Ya definitely sounds like it needs different values for those parameters than what the C172 uses. Hopefully after MS/Asobo update the SDK docs the aircraft devs can get to work on tuning these parameters for their birds.

11 minutes ago, lwt1971 said:

 Ya definitely sounds like it needs different values for those parameters than what the C172 uses. Hopefully after MS/Asobo update the SDK docs the aircraft devs can get to work on tuning these parameters for their birds.

Having tested some more it could have very well been pilot error on the first try, as subsequent attempts seemed fine. However I think I'll stop experimenting with this for now and wait until release / 3rd party updates. 

4 hours ago, Dutch727 said:

The most interesting improvement for me in the SU15 beta is the ground handling.  I'm hoping takeoff behavior is improved.  

Behavior before SU15:  As soon as the tail comes up the airplane makes an abrupt turn into the wind and starts heading for the grass.  Only immediate strong rudder would keep me on the runway, but then it would veer in the opposite direction.  Although I could find a spot on the rudder to keep it on the runway it was extremely hard to prevent crazy zigzagging until I could lift off.  True for both aircraft.  This is after much practice and adjusting rudder sensitivity. 

After SU15:  As soon as the tail comes up the plane starts turning into the wind but no longer is it so fast or abrupt.  I can apply much less rudder to keep near the centerline.  It's now possible on the B18 and the DC3 to keep reasonably straight on the runway with a crosswind.  It's not easy, but it shouldn't be.  Previously, it was all I could do veering back and forth to keep it on the pavement.  The difference is very noticeable for me.

I'm interested in other's experiences. 

Hoo, boy do I hope this can end up fixing BlackBird Sims F4U Corsair!

4 hours ago, pmplayer said:

 When will SU15 officially be released for everyone?

March 12.

1 hour ago, Cpt_Piett said:

Allright, so transplanting the new ground handling to the RV-14 did not work at all. In fact the aircraft was all over the place on takeoff and landing. 

I had a couple of dodgy transplants too!   The COWS DA-42 did not take well to the transplanted lines - the nose was actually zig zagging left and right a little (not the nosewheel which was going straight; the actual nose), like some very large bearings had broken 😄

I put this down to this line / value:-

ground_new_contact_model_gear_flex = 0.005

....rather than just affecting up/down flex / movement on the nosewheel, I believe this value effects 360 degree movement around the nosewheel fixture to the plane.  So great care has to be taken with this value, or it will appear that the nosewheel is attached to the fuselage using nothing but a large rubber band! 😄

Edited February 15, 20242 yr by JYW

5 minutes ago, JYW said:

nosewheel is attached to the fuselage using nothing but a large rubber band! 😄

Lol! I reset all the aircraft to default after reading this 😄

I've been busy with family matters for couple weeks. Finally back to flying. In the new beta, fps is up enormously. Not just on long flights or heavy scenery stutters when panning the outside view. All the time.

I use a "standard" flight for testing fly out of default JFK towards Manhatten skyscrapers in Textron steam gauge 172. Low in dense scenery getting 95 to 108 fps (using Frame Generation). When exiting JFK over the terminal buildings was 78 fps. This is about a 1/3 increase in fps. Panning in outside view is completely smooth. Obviously in cockpit view also.

Using the Ultra global rendering preset , 200 both LODs, and Quality DLSS. 1440p on one monitor only.

But I'm one guy doing one test (so far). Could be anecdotal. But the best performance ever for me, and really fun.

22 hours ago, omarsmak30 said:

DX11 or DX12?

DX11

3 hours ago, turbomax said:

...how can it be smoother than smooth and more constant than constant frame times?...

Well friend, in my view, we still need better optimisation to help people who have less powerful machines, and future headroom for development, so the boundaries should always be pushed. 
Just wait until they introduce ray tracing (a chance for MSFS2024) - we will re-inspect our frame times then!  

I do agree though.  We need better multi-threading to free up the main thread.  At 4k ultra, my 4080 is showing 40% utilisation in general, but my 14900k is bouncing off the top between 90-100% utilisation on two threads.  The others cores are only utilised to about 25%.

I have hyperthreading off in the BIOS, as I get no more performance with it on, but just more heat.  In my view, MSFS doesn't need any more than 8 cores / threads at the moment. 
Concentration needs to be on unloading the main thread more - Asobo see this, and from the twitch session a while ago, that is precisely what they are looking at.   

Thanks, I'm going to go now and turn hyperthreading off. I have a 3DX cpu which doesn't run cool.