Fallout 4 Optimization and Performance Systems Explained

[Vinokopimi]

 

 

If you mean that you generated precombineds/previs multiple times for the same cell as different plugins touched it, and want to merge the generated results, that won't work at all. Since you mention merging XPRI and XCRI, I assume that is your intent.

Omg yeah that's what I was doing and I spent a week already fixing things :dry:

 

 

 

If you want to generate previs/precombineds for one cell that is touched by several mods, you would needs to flag them all as ESM (just the flag using xEdit, don't change the file extension), load them as masters in the CK, then generate previs and precombineds and save the new plugin and use that as your patch.

 

I've tried this route on CK but i can only generate precombine for the commonwealth world, there is no option to generate previs for the whole commonwealth.

 

Thank you for the VC and timestamp info that would really help me.

 

What I'm trying to accomplish is make precombines and previs for a merge patch containing mods such as beantown interiors project, boston natural surroundings and other mods that had added something to the commonwealth world. So that I'll have a near perfect fallout (spent a lot of time tweaking and fiddling trying to fix my load order to make it perfect, compared to my actual play time).

 

What I figured was, have a merge file of all these records then use CK and not the cli to generate the precombines. My belief is that CK would be more stable this way and will be forced to regenerate everything and retain everything? :pinch:

 

What would be the best approach to attain my goal? Should I just merge BT_interiors and BNS then generate precombines and previs on cli?

Edited May 26, 2020 by crimsonlourence

[SMB92]

The best way to deal with a "Merged patch" approach is to generate with all those mods loaded, at once. I started a project doing something like this but it's far more work/upkeep than I care to commit. You will need to keep on top of mod updates, and be prepared to edit and regenerate locations that cause problems for previs (loose example, trees in front of windows might cause a noticeable occlusion pop in/out). You will get far more consistent results this way, but it's a lot to maintain as said. From there you can consider using that xEdit script that optimises the generation.

 

RE version control, you may want to look into that for CK, but it might just be easier to keep manual backups of the plugin and generations. In fact I seem to have a lot of drama with previs gens with VC enabled (seems to hang a lot, probably due to network setup), not that I've bothered to solve it though.

[ArdaxT]

So, after reading almost all of this thread, I've got a few questions:

 

1. Has anyone managed to get a command-line workflow for precombine/previs building working end to end?

2. Is there an xEdit script for detecting (or better yet, fixing) that matswap issue w/ regenerate precombines?

3. Is there a good tutorial for getting the "-CheckinPlugin" bit to work? I've seen some tutorials for getting version control up and running, but when I do that I run into two issues: First, precombine generation doesn't work properly (I get an empty PrecombineObjects.esp); second, CheckinPlugin doesn't seem to modify my .esp (it's master-flagged). I worked around this with some xEdit scripting.

4. Do I still need to do the hex edits on the CreationKit.exe? My fallout crashed when I tried to teleport to a cell with newly generated precomb/previs.

 

Apologies if anything's already been answered (here or elsewhere), I've been pouring through a lot of information in a short amount of time, so some of it might have leaked out.

[ArdaxT]

So, after a bunch of faffing about, I figured out:

 

1. Not yet, but it's better. Once I spend some time on #3 it'll be pretty close to completely automated.

2. Edit: Figured out the answer. zilav posted it earlier in thread.

3. Never figured this out either, but I've built some xEdit scripts that do the same thing. (Okay, it's mostly "cribbed stuff from other scripts". I've got almost 20 years of experience with Delphi, but I haven't touched it in the past 5 years and don't know the xEdit API very well yet.)

4. Yes, the hex edits are necessary.

 

Command line generation is a little crashy for some reason. Like it can't handle the number of records in the file if it gets too large. When it works though, it's much more reliable than creating things in the CK GUI.

[VIitS]

Nukem is wonderful and awesome: https://github.com/Nukem9/Fallout4Test/releases/tag/0.1

First (early, limited) release of CK Fixes for FO4. Includes the commandline fix, among other things. Just remember to be respectful and don't pester him for features/updates.

 

Requires the latest VS2019 x64 redist, and if you have any issues there is no support. It is very much in the alpha stage, and Nukem barely touches FO4 so it is a much, much lower priority than any of his Skyrim stuff. Feel free to open an issue if you think there is a legitimate bug (rather than a PICNIC error) or you have a request for a feature/CK bug fix, but don't expect quick responses.

[ArdaxT]

That's an amazing thing that Nukem is doing with the CK. Wow!

 

So I figured out why the CK seemed to be "crashing" for me on certain files. It seems I actually have to go into the EditorWarnings.txt file and really go through it pretty carefully. I've now got my CK attempting to generate precombines for more or less every cell possible: interior and exterior, Commonwealth and DLC. If the CK thinks there's anything wrong with the plugin you pass it it won't even get started.

 

I'm pretty sure my script to build the plugin is adding more cells than strictly necessary. I need to figure out better ways to filter out cells that don't need precombines regenerated for them with xEdit scripts.

[Ares8]

I seem to be having a problem with regenerating the precombined and previs data of the mod "immersive and extended Lexington".

I generate precombineds for every cell the mod touches, which are 28 in a 9x9 radius around -9,9. Then I generate precombined previs for all loaded cells centered around -9,9.

 

Everything seems to be fine for the areas that were edited themselves, but areas that hadn't had their precombineds regenerated, but that were still in previs range due to the 3x3 range, seem to be having issues with flickering objects now. Deleting the affected cells in fo4edit doesn't change this problem.

 

I was wondering what could be causing this?

Edited June 30, 2020 by Ares8

[SMB92]

Be careful you didn't pick up that hidden occlusion box in Monsignor Plaza. It'll break the generation in this fashion. You'll need to find it and disable it. It's in the top of a building that had a cracked roof. Don't have a pic handy sorry.

 

I'm not familiar with that mod so I won't comment beyond that.

[Ares8]

Be careful you didn't pick up that hidden occlusion box in Monsignor Plaza. It'll break the generation in this fashion. You'll need to find it and disable it. It's in the top of a building that had a cracked roof. Don't have a pic handy sorry.

 

I'm not familiar with that mod so I won't comment beyond that.

I forgot to mention, for those who don't know the mod, that the mod only affects the are around corvega. I don't know if there is an occlusion box there, but the main area I'm having trouble with is the cell with the covenant hq (compound), even though it is not affected by the mod itself, only it's previs is regenerated because it's in a 3x3 block with some affected cells.

[VIitS]

I don't think the CK will let you do it anymore, though it definitely did at one point, but is the RVIS entry different for your plugin than vanilla? That is the part that defines the center of the 3x3 previs grid, and if it is out of sync with the rest of the worldspace you will see tons of flickering issues at cell borders.