PC V2.6 EXP - updated b14

Will they do something with the online gameplay on the servers? Since you and alpha 16 haven't done anything good for the servers, thereby losing online, because many people are already tired of the usual 7 days to die, and you don't give modders any way.

If not for the modders all that modder-friendly XML probably wouldn't even exist.

And if you are speaking about console-players, it is almost certain that Sony and/or Microsoft didn't allow it. For TFP the limitation on console did only add unneccesary workload and nothing to gain
 
@faatal

Found something that can be optimized with minimal changes which has quite an impact on traversal in the world:

Chunk mesh uploads from same-frame blocking to next-frame deferred (changes in VoxelMesh.cs -> copyToMesh() and MeshDataManager.cs -> ApplyAndDispose()).

Previously MeshDataManager.Add() was called with sameFrameUpload: true, meaning background tasks would kick off copying vertex data, and then LateUpdate would hit ApplyAndDispose() which called m_statesAllComplete.Wait(), stalling the main thread until every copy task finished. With several chunks loading at once this was eating multiple milliseconds per frame in LateUpdate.
Now with sameFrameUpload: false, the copy tasks get a full frame to run. Frame N kicks them off, frame N+1 calls ApplyAndDispose and by then they're already done, no waiting. Chunks appear one frame later than before, but there's no main thread stall.
 
@faatal
  • Added tougher biome spawn zombies to the snow and wasteland biomes.

I have really enjoyed this, it made those biomes much more challenging. Any chance the wandering hordes could be updated with something similar? even late-game in the wasteland it still consists of a handful of regular zombies. They are not a threat at all :)
 
@faatal

Another easy one, AnimatorCullingMode is set to AlwaysAnimate, so even when not visible they still animate.

Also the tick slicing thing can do more harm than good.

TickEntities the code calculated tickEntitySliceCount based on frame time averaging and spread entity updates across multiple frames. When FPS dropped, fewer entities got updated per frame, but the stale entities accumulated work, causing worse FPS on subsequent frames, a death spiral
 
@faatal

Found something that can be optimized with minimal changes which has quite an impact on traversal in the world:

Chunk mesh uploads from same-frame blocking to next-frame deferred (changes in VoxelMesh.cs -> copyToMesh() and MeshDataManager.cs -> ApplyAndDispose()).

Previously MeshDataManager.Add() was called with sameFrameUpload: true, meaning background tasks would kick off copying vertex data, and then LateUpdate would hit ApplyAndDispose() which called m_statesAllComplete.Wait(), stalling the main thread until every copy task finished. With several chunks loading at once this was eating multiple milliseconds per frame in LateUpdate.
Now with sameFrameUpload: false, the copy tasks get a full frame to run. Frame N kicks them off, frame N+1 calls ApplyAndDispose and by then they're already done, no waiting. Chunks appear one frame later than before, but there's no main thread stall.
I asked the programmer who made that system. There may have been some issue with making it false.
 
I asked the programmer who made that system. There may have been some issue with making it false.
I tested this with my decompiled/recompiled binary and had no issues, what I suspect is that at some point this was used to prevent entities from falling through the world but these days entities don't even update when the chunk isn't loaded yet but who knows, hopefully there is nothing to it, with that change traversal is butter smooth for me with no hitching.
 
Is there any news yet on when the stable version for PS5 will be released, and whether the performance issues with version 2.6 will be fixed?


I really want to keep playing, but right now it's practically unplayable because of the stuttering.
 
Is there any news yet on when the stable version for PS5 will be released, and whether the performance issues with version 2.6 will be fixed?


I really want to keep playing, but right now it's practically unplayable because of the stuttering.
they still have it in experimental... when you see PC 2.6 stable announced, then they'll submit it to Sony & Microsoft for their usual sluggish certification system. even if they did it today you'd probably not see it for weeks cause of certification
 
Hey TFP my dudes, can we get a Server IP block ability on the search function? All I get are fake Chinese servers. So many that I can't even find local things I know exist because the search reaches its limit on finds on pure bs chinese servers. Region is set to africa but these pigs don't care, here to hack PCs prolly. Noticed they all use the same server IP. Can the game not be made to recognize server locale? Maybe have it in the server code to report its region.
 
they still have it in experimental... when you see PC 2.6 stable announced, then they'll submit it to Sony & Microsoft for their usual sluggish certification system. even if they did it today you'd probably not see it for weeks cause of certification
That's not entirely accurate. Sony generally gets the updates out very quickly... usually within only a few hours of Steam having them. Microsoft is the one who is slow, and you can see anything from a few hours to a few days for most updates, with a week or two only happening sometimes. But both do have their process of approving the updates and the speed that it happens will depend on different factors. So you might get a long delay from Sony on an update even if they are normally fast and a short delay from Microsoft even though they are normally longer.
 
@faatal

Another easy one, AnimatorCullingMode is set to AlwaysAnimate, so even when not visible they still animate.

Also the tick slicing thing can do more harm than good.

TickEntities the code calculated tickEntitySliceCount based on frame time averaging and spread entity updates across multiple frames. When FPS dropped, fewer entities got updated per frame, but the stale entities accumulated work, causing worse FPS on subsequent frames, a death spiral
I believe AnimatorCullingMode not animating can break things in some cases. Not looking for any new bugs right now.

You should not be able to death spiral past all of them updating at the same time as it used to do. I will review it when I get a chance.
Post automatically merged:

I tested this with my decompiled/recompiled binary and had no issues, what I suspect is that at some point this was used to prevent entities from falling through the world but these days entities don't even update when the chunk isn't loaded yet but who knows, hopefully there is nothing to it, with that change traversal is butter smooth for me with no hitching.
He said you would see no mesh for a frame, which is why the delayed version is not used. We may look at it when time permits to keep the previous mesh around during the delay.
 
Is there any news yet on when the stable version for PS5 will be released, and whether the performance issues with version 2.6 will be fixed?


I really want to keep playing, but right now it's practically unplayable because of the stuttering.
It has been submitted, but I believe they said MS was being slow with the approval for an unknown reason. Any day now I would guess.

One of our console devs did some benchmarks and performance seems as good as 2.4.
 
they still have it in experimental... when you see PC 2.6 stable announced, then they'll submit it to Sony & Microsoft for their usual sluggish certification system. even if they did it today you'd probably not see it for weeks cause of certification
That isn't how TFP operates. They don't announce stable for PC and THEN submit it to Sony and Microsoft. They wait until certification comes from both console companies before announcing stable for everyone. Then they release the stable for everyone at the same time. When the announcement comes there might be a few hours delay for one or both of the consoles but it should happen same day for sure.
 
That isn't how TFP operates. They don't announce stable for PC and THEN submit it to Sony and Microsoft. They wait until certification comes from both console companies before announcing stable for everyone. Then they release the stable for everyone at the same time. When the announcement comes there might be a few hours delay for one or both of the consoles but it should happen same day for sure.
Has that changed? We've seen Microsoft not release the update for days or even more than a week, not just for XBox but also for XBox Store players on PC. I'd have thought that was due to the delay with certification, but maybe it's some other reason? Either way, it has been delayed a lot for XBox in the past.
 
Has that changed? We've seen Microsoft not release the update for days or even more than a week, not just for XBox but also for XBox Store players on PC. I'd have thought that was due to the delay with certification, but maybe it's some other reason? Either way, it has been delayed a lot for XBox in the past.
Maybe there have been delays beyond a day. I don't follow the console side of it closely. But the delays were NOT because TFP submitted for certification after they announced stable. Perhaps they anticipated the approvals coming and then it didn't happen but the submission for approval had already happened.

As faatal stated, the submissions for 2.6 have already been made and they have not announced stable yet.
 
Back
Top