stutter every couple seconds

[CaptainLHKLM]

On 2/4/2024 at 10:47 AM, Ben Russell said:

It's both.

The bug/problem is most prevalent when your hardware is under heavy system load.

 

Hi Ben,

Is there anything we can do about the gizmo part, though? Whether I run XP on high or low settings, my machine (which I still consider good, with an i7 10700KF, RTX3080 and 32GB RAM) still produces the rhythmic stutter every couple seconds. I'm coincidentally on a brand new Windows install too with little to no clutter software, but I had the same issue in my previous install too. I'm not asking because I want to be a pain, but because I've never been able to fly the challenger and properly enjoy it. I'm more than happy to provide any logs or files from my part!

Best wishes,

Daniel

EDIT: Just to illustrate, I attached a screenshot with an output graph of the FPS. In this scenario, I left the aircraft stationary and did not move the camera.

Edited February 22 by CaptainLHKLM

[Pils]

Let’s not make assumptions here. How did you determine this was a Gizmo issue exactly?

[Ben Russell]

1 hour ago, Pils said:

Let’s not make assumptions here. How did you determine this was a Gizmo issue exactly?

Gizmo is very likely to be the cause. 

Early in the release we did some deep digging into the issue. 

There is code inside gizmo that calls sleep(0)... (the reasons for which is beyond the scope of this reply) 

Windows, at its discretion, can sleep for upto 16ms/whatever it needs to service the rest of the OS/apps. This can show up as dropped frames or massive perf chugs. 

There is a command line option for xplane to leave some cores alone so the os can service itself on those. It helps but the real solution is to fix gizmo.. 

 

I have been working on a gizmo update that is based on the three? Year old version that X-Aviation ships. 

It contains bug fixes for the most prominent issues (including this one) but is mostly an LTS update. It will also have Apple silicon support. 

I hope we can bring it to you all as an update soon. 

Edited February 23 by Ben Russell

[CaptainLHKLM]

3 hours ago, Ben Russell said:

Gizmo is very likely to be the cause. 

Early in the release we did some deep digging into the issue. 

There is code inside gizmo that calls sleep(0)... (the reasons for which is beyond the scope of this reply) 

Windows, at its discretion, can sleep for upto 16ms/whatever it needs to service the rest of the OS/apps. This can show up as dropped frames or massive perf chugs. 

There is a command line option for xplane to leave some cores alone so the os can service itself on those. It helps but the real solution is to fix gizmo.. 

 

I have been working on a gizmo update that is based on the three? Year old version that X-Aviation ships. 

It contains bug fixes for the most prominent issues (including this one) but is mostly an LTS update. It will also have Apple silicon support. 

I hope we can bring it to you all as an update soon. 

Great to hear, thank you!

[CaptainLHKLM]

On 2/23/2024 at 5:59 AM, Ben Russell said:

Gizmo is very likely to be the cause. 

Early in the release we did some deep digging into the issue. 

There is code inside gizmo that calls sleep(0)... (the reasons for which is beyond the scope of this reply) 

Windows, at its discretion, can sleep for upto 16ms/whatever it needs to service the rest of the OS/apps. This can show up as dropped frames or massive perf chugs. 

There is a command line option for xplane to leave some cores alone so the os can service itself on those. It helps but the real solution is to fix gizmo.. 

 

I have been working on a gizmo update that is based on the three? Year old version that X-Aviation ships. 

It contains bug fixes for the most prominent issues (including this one) but is mostly an LTS update. It will also have Apple silicon support. 

I hope we can bring it to you all as an update soon. 

Hi Ben,

Do you have an update regarding the gizmo update by any chance? Or do you have a fix for the stuttering issue I could use in the meantime?

Best wishes,

Daniel

EDIT: I found a newer version of Gizmo on your website, and installed it. However, it seems that the FPS drops are still present when observing through the X-Plane data output graph and the new built-in Gizmo graph.

 

Edited June 5 by CaptainLHKLM

[Pils]

3 hours ago, CaptainLHKLM said:

EDIT: I found a newer version of Gizmo on your website, and installed it. However, it seems that the FPS drops are still present when observing through the X-Plane data output graph and the new built-in Gizmo graph.

As Ben hinted at, make sure you start X-Plane with --num_workers=11. Also, look at the frame profiler, it'll help point at a particular part of the frame rendering that's being slow.

[CaptainLHKLM]

14 hours ago, Pils said:

As Ben hinted at, make sure you start X-Plane with --num_workers=11. Also, look at the frame profiler, it'll help point at a particular part of the frame rendering that's being slow.

Thanks! I'll add this next time I launch X-Plane. In the meantime, I did the following: uninstall the old Challenger, remove Gizmo and all related files, install the new Gizmo plugin, install the Challenger beta and it seems  (I'm only on my first flight) that this has also resolved this issue for me. Now, the spikes are only slightly upward and so short in duration that they're unnoticeable to me.

EDIT: They're still noticeable on landing (the stutters going down again, instead of up on the FPS graph), but it's still wayy better than it used to be. I'll give the --num_workers=14 (or should it be 11? I followed the step-by-step that was posted in 2022 and it said the value should be the number of logical processors minus 2) a try now, and might do a clean gizmo install too, if that doesn't help. Thanks for your help again Pils!

Edited June 6 by CaptainLHKLM

[bpcw0001]

Can the new Gizmo plugin also be installed into the 1.7 Challenger release?
Would this even be of any benefit?

[Pils]

15 hours ago, CaptainLHKLM said:

11

I recommend 11 on 8-core HT/SMT processors as there’s resources to spare. The original advice was more for processors with 8 or less threads.

Edited June 7 by Pils

[CaptainLHKLM]

On 6/7/2024 at 2:07 AM, Pils said:

I recommend 11 on 8-core HT/SMT processors as there’s resources to spare. The original advice was more for processors with 8 or less threads.

Ok, thanks!

[Pils]

Please post Log.txt

[bpcw0001]

On 6/6/2024 at 2:08 PM, bpcw0001 said:

Can the new Gizmo plugin also be installed into the 1.7 Challenger release?
Would this even be of any benefit?

I tried that, but Gizmo requires a serial number.
So I suppose just dropping a new Gizmo version as replacement for the existing one does not work?

[Pils]

2 hours ago, bpcw0001 said:

I tried that, but Gizmo requires a serial number.
So I suppose just dropping a new Gizmo version as replacement for the existing one does not work?

Define “work”.

[Ben Russell]

15 hours ago, bpcw0001 said:

I tried that, but Gizmo requires a serial number.
So I suppose just dropping a new Gizmo version as replacement for the existing one does not work?

Gizmo-LTS is fully functional without a serial. 

You can download and try it free of charge. 

If it provides you benefit/improvement and you want to get rid of the nag dialog you can buy a serial number. 

[bpcw0001]

6 hours ago, Ben Russell said:

Gizmo-LTS is fully functional without a serial. 

You can download and try it free of charge. 

If it provides you benefit/improvement and you want to get rid of the nag dialog you can buy a serial number. 

Ok - thanks for the info!

[bpcw0001]

On 6/11/2024 at 3:47 PM, Pils said:

Define “work”.

in my case, "not work" would mean what I observed:

- using Gizmo-LTS requires a serial
- when waiting out the Gizmo nag screen, the Challenger asks for activation again

-> so Gizmo-LTs it is not a drop-in replacement for the version that ships with the Challenger in the sense that you just replace it and everything just works as before, only better.

Edited June 13 by bpcw0001