DynDOLOD 3.00 Alpha 182

[sheson]

6 hours ago, Daemonjax said:

How would I go about fixing this in iNeeds - Extended.esp (there's a bunch of these, but for brevity only posting one):

<Warning: Base record type not STAT or MSTT for large reference iNeed - Extended.esp [REFR:0009840B] (places _SNSnowDriftM01 "Snow" [ACTI:0A001829] in GRUP Cell 

It changes the large Showdrifts into activators (using same nif) so that you can gather snow.  That's cool, I'd like to keep it... but it's not critical to the mod functioning.

I know what the problem is, as per DynDoLod manual...

The base record not being of the type STAT or MSTT. For example if the base record is changed by a plugin for a large reference, even if the plugin is an ESM. A warning message Base Record Type Not STAT or MSTT For Large Reference is written to the message log.

... but not sure what to do about it if I want to keep these activators functioning. 

Would I really need to removing the mod's changes to these recard while also placing an invisible activator at all those locations?  Or is there something easier I can do? 

Maybe the easiest thing to do (not 100% sure this would work) would be to copy all these worldspace edits as new records into a new esp (instead of overrides) modify the nif to be invisible (or just have no nif?  dunno), move the model origin up by a couple of z units, then delete all the original worldspace edits?  EDIT: I'm working on trying this now... It REALLY sucks that xEdit doesn't have a "Deep Copy as New Record" function -- I have to "copy as new record" each one individually.  EDIT2:  There are too many records to do this by hand, and the xEdit script "Mark new records from overrides" would be perfect if it actually worked (fails on line 21).

These changes trigger large reference bugs, e.g. texture flicker between the full model and LOD model just outside the loaded cells in the large reference grid area.

If those bugs are visible for a cell, the workaround is to move the original reference underground to -30000 in a ESM flagged plugin to hide them and add new reference that is a copy as new of the original reference but uses the new ACTI base record (plugin type does not matter).

Use the xEdit discord for help with xEdit, its scripting or to report bugs with xEdit.

[SirHarbinger]

Integer overflow error. This never occurred before. I was able to successfully run dyndolod 3.0 (alpha68) one time today, and then the subsequent runs will all show this error.

DynDOLOD_SSE_log

DynDOLOD_SSE_Debug_log

bugreport

 

Edit: there's no main difference between the first successful dyndolod run and the ones that are currently causing bugreports. The only thing I changed within dyndolod was to set level32=1 because I looked back at "A Clear Map of Skyrim" and saw related stuff about that option so I tried it for the second time. This caused dyndolod to have integer overflow error, so I reverted changes to level32=0, and then returned "Clear Map of Skyrim"'s options for non-level32, but dyndolod continues to find fault with that one scaffolding from JK's that was never a problem prior (was using alpha 65 prior to today's one-time alpha 68)

Edited February 12, 2022 by SirHarbinger
Additional Info

[sheson]

18 minutes ago, SirHarbinger said:

Integer overflow error. This never occurred before. I was able to successfully run dyndolod 3.0 (alpha68) one time today, and then the subsequent runs will all show this error.

DynDOLOD_SSE_log

DynDOLOD_SSE_Debug_log

bugreport

https://stepmodifications.org/forum/topic/15606-dyndolod-300-alpha-68/?do=findComment&comment=256894

[RuneEva]

1 hour ago, sheson said:

 

 

Delete all old logs and bugreport.txt, then run this test version of DynDOLODx64.exe. It should still fail but hopefully provide more information. Upload new bugreport.txt, log and debug log.

Delete all old logs and bugreport.txt, then run this test version of TexGenx64.exe. In case there are still problems upload new bugreport.txt, log and debug log.

Thanks, same error as expected but here are the new logs.

bugreport.txt DynDOLOD_SSE_Debug_log.7z DynDOLOD_SSE_log.txt

[sheson]

36 minutes ago, RuneEva said:

Thanks, same error as expected but here are the new logs.

bugreport.txt 104.86 kB · 1 download DynDOLOD_SSE_Debug_log.7z 989.5 kB · 1 download DynDOLOD_SSE_log.txt 214.48 kB · 1 download

Thanks. This test version should fix the integer overflow error while processing a child world. In case it doesn't, upload new bugreport and logs.

[RuneEva]

10 minutes ago, sheson said:

Thanks. This test version should fix the integer overflow error while processing a child world. In case it doesn't, upload new bugreport and logs.

 Getting this error when trying to launch this version.

[sheson]

11 minutes ago, RuneEva said:

 Getting this error when trying to launch this version.

Restart the PC. Delete the DynDOLDO folder. Install the latest version. Then replace the exe with the new test version. Do not rename the exe.

If problem persists upload any log file found in the log folder if there is one.

Edit: And re-download this test version instead.

[SirHarbinger]

1 hour ago, sheson said:

Delete all old logs and bugreport.txt, then run this test version of DynDOLODx64.exe. It should still fail but hopefully provide more information. Upload new bugreport.txt, log and debug log.

Worked fine this time with that download, thanks.

i wanted to test it even further and rerun it a second time, but it gave the "Integer Overflow" error.

second time logs : DynDOLOD_SSE_log , DynDOLOD_SSE_Debug_log , bugreport

Rerun it a third time and it's another "Integer Overflow"

third time logs: DynDOLOD_SSE_log , DynDOLOD_SSE_Debug_log , bugreport

To be fair, since it was able to run properly with the test version of Dyndolod and successfully completed it for that one run, I'm already considering it a success on my part. Maybe Dyndolod 3 isn't meant to be run after completing it once? But, then, I distinctly remember dyndolod 3 alpha 65 being able to run like 30minutes to 1hour of each run with no bugs whatsoever.

Apologies if the new two tests are redundant info for you.

[sheson]

7 minutes ago, SirHarbinger said:

Worked fine this time with that download, thanks.

i wanted to test it even further and rerun it a second time, but it gave the "Integer Overflow" error.

second time logs : DynDOLOD_SSE_log , DynDOLOD_SSE_Debug_log , bugreport

Rerun it a third time and it's another "Integer Overflow"

third time logs: DynDOLOD_SSE_log , DynDOLOD_SSE_Debug_log , bugreport

To be fair, since it was able to run properly with the test version of Dyndolod and successfully completed it for that one run, I'm already considering it a success on my part. Maybe Dyndolod 3 isn't meant to be run after completing it once? But, then, I distinctly remember dyndolod 3 alpha 65 being able to run like 30minutes to 1hour of each run with no bugs whatsoever.

Apologies if the new two tests are redundant info for you.

See https://stepmodifications.org/forum/topic/15606-dyndolod-300-alpha-68/?do=findComment&comment=256899

[RuneEva]

24 minutes ago, sheson said:

Restart the PC. Delete the DynDOLDO folder. Install the latest version. Then replace the exe with the new test version. Do not rename the exe.

If problem persists upload any log file found in the log folder if there is one.

Edit: And re-download this test version instead.

No integer overflow with this but I'm getting "Item Not Found" instead.

DynDOLOD_SSE_Debug_log.txt couldn't fit this time (93mb / 4.4 mb compressed) so here's the link : https://mega.nz/file/K4Vg2ZCZ#xT7F2inJC6V9KhvpwzTZFr_YM0umSnBrKmuZsxUvNpg

bugreport.txt DynDOLOD_SSE_log.txt

[MajorSollus]

sheson, per your request on Nexus, I am moving my inquiry to here.

I really like what has been done with LODGen and DynDOLOD, so thank you. I have spent the last week trying to figure out how to solve an issue. I am pretty good at solving my own issues, but this one eludes me. I have read manuals, watched vids, etc., and perhaps I am missing something or this is a bugger, IDK. The issue is:

I run a Tree DynDOLOD generation and it works great except for one issue I have noticed that goes back a few years and still seems to be present for some people. The top of Dragonsreach is chopped off when approaching close to Whiterun after generating a Tree Only LOD. I have tried multiple resolution attempts and have solved the issue by adjusting the IgnoreLargeReferences=1 and uLargeRefLODGridSize=5, generating LOD, then once completed setting uLargeRefLODGridSize back to 11 and launching game. This fixes Dragonsreach, but this introduces another problem. I am running Bent Pines II, which does not contain LODS, so when I generate Tree LODS with DynDOLOD by ignoring the Large References to fix Dragonsreach, it appears those huge trees are also considered Large References, and so the LODS are not created. I then decided to run a Full LOD process with DynDOLOD with all the appropriate variables in place, but then I end up having textures missing in some snow and some waterfall areas.

So, my solution was to try and generate the Bent Pines II Tree LODS separately with DynDOLOD, and the References set accordingly for Large. However, it seems that by having two DynDOLOD Outputs loaded, whichever DynDOLOD file is loaded after the previous version overwrites it completely. Now, I'm trying to generate the Bent Pines II LODS with SSELODGen. I have created the appropriate Billboard, but I am unable to figure out the Text File (Atlas) that is supposed to accompany the .dds Billboard. I have also created the appropriate file folder directory for the Bent Pines .esp., just can't seem to get the text file to go along with creating these LODS with SSELODGen. Note, I have tried splitting the Atlas with Bent Pines loaded and all required dependencies, but tree LODS are not loading into the split Atlas from Bent Pines II; only vanilla trees from Tamriel.

Do you have any advice on how to accomplish this final step for a Tree LOD for a single MOD? I am assuming the Atlas file is generated somewhere in DynDOLOD. I'm almost there. Thanks in advance.   

[insxsni]

Hi, I'd like to address an issue with the newest Alpha 68 update. The program runs out of memory. I have a 16 gb system which was enough to generate LODs on the previous Alpha 65 version, but on this recent update, it can't make it pa st Falskaar during the Occlusion generation. I attached a picture of all of the worlds it generates for. 

I'm trying to generate ultra trees as well as grass lod, and I was also trying to generate seasons. But simply, it runs out of memory. Like literally, the message is "out of memory" and it can't generate anymore LODs. The program seems to hit a "wall" because the memory usage just climbs up, and up, and up, until it reaches 99% and my computer is frozen and nonfunctional. Even after leaving the program alone, no chrome or anything using a substantial amount of memory, it can't do it. I'd love to attach a screenshot, but my screenshot program can't function because there is only 100 mb of memory left for my computer during LOD generation. I don't know if its a bug, or if DyndoLOD now requires 32 gb of memory to function. Either way, I don't want to have to buy two ram sticks just to generate LODs in an 11 year old game. Thanks!

P.S. If I sound grumpy or mean in this post at all, I apologize in advance. This issue has just been bothering me all day and I really don't know what the problem is. I've never had it before!

Edit: I'd also like to add that the memory destruction starts happening an hour into the generation process? The time I don't really have a problem with, but the fact that it happens so far into it just seems like a huge waste of time. On the bright side, it's time I use spent away from my computer because it is unfunctional with the program running! The power usage from the constant ram running at 100% is a little worrying though.

 

Edited February 12, 2022 by insxsni

[sheson]

15 minutes ago, insxsni said:

Hi, I'd like to address an issue with the newest Alpha 68 update. The program runs out of memory. I have a 16 gb system which was enough to generate LODs on the previous Alpha 65 version, but on this recent update, it can't make it pa st Falskaar during the Occlusion generation. I attached a picture of all of the worlds it generates for. 

I'm trying to generate ultra trees as well as grass lod, and I was also trying to generate seasons. But simply, it runs out of memory. Like literally, the message is "out of memory" and it can't generate anymore LODs. The program seems to hit a "wall" because the memory usage just climbs up, and up, and up, until it reaches 99% and my computer is frozen and nonfunctional. Even after leaving the program alone, no chrome or anything using a substantial amount of memory, it can't do it. I'd love to attach a screenshot, but my screenshot program can't function because there is only 100 mb of memory left for my computer during LOD generation. I don't know if its a bug, or if DyndoLOD now requires 32 gb of memory to function. Either way, I don't want to have to buy two ram sticks just to generate LODs in an 11 year old game. Thanks!

P.S. If I sound grumpy or mean in this post at all, I apologize in advance. This issue has just been bothering me all day and I really don't know what the problem is. I've never had it before!

Edit: I'd also like to add that the memory destruction starts happening an hour into the generation process? The time I don't really have a problem with, but the fact that it happens so far into it just seems like a huge waste of time. On the bright side, it's time I use spent away from my computer because it is unfunctional with the program running! The power usage from the constant ram running at 100% is a little worrying though.

Read the first post which log files and bugreport.txt to upload when making posts.

Read https://dyndolod.info/Help/Occlusion-Data#Out-of-Memory-while-Generating-Occlusion-Data

[sheson]

47 minutes ago, MajorSollus said:

sheson, per your request on Nexus, I am moving my inquiry to here.

I really like what has been done with LODGen and DynDOLOD, so thank you. I have spent the last week trying to figure out how to solve an issue. I am pretty good at solving my own issues, but this one eludes me. I have read manuals, watched vids, etc., and perhaps I am missing something or this is a bugger, IDK. The issue is:

I run a Tree DynDOLOD generation and it works great except for one issue I have noticed that goes back a few years and still seems to be present for some people. The top of Dragonsreach is chopped off when approaching close to Whiterun after generating a Tree Only LOD.

Do you have any advice on how to accomplish this final step for a Tree LOD for a single MOD? I am assuming the Atlas file is generated somewhere in DynDOLOD. I'm almost there. Thanks in advance.   

It is not possible to generate LOD for a single mod, because how LOD works. As explained at https://dyndolod.info/How-LOD-Works, "These different LOD types are basically optimized and pre-combined models or meshes and textures for a group of cells that represents the full content in the exterior cells."

Also, for example, the tree LOD atlas for a worldspace contains the billboard textures for all trees in a worldspace. Hence in order to have all require billboards added to it, tree LOD needs to be generate for the entire worldspace and because LOD only works and looks correctly for the load it was generated for, it really means that tree LOD needs to be generated for the current load order. This includes the sort of of plugins to not have engine bugs like stuck tree LOD.

See https://dyndolod.info/Updating

It is not possible for tree LOD only generation (e.g. object LOD generation is not checked) to change existing object LOD *.BTO meshes, which are what affect LOD for Dragonsreach. It sounds like you are probably generating tree LOD without loading the exiting DynDOLOD plugins but then overwrite the old plugins with the new DynDOLOD plugins that only contain basic tree LOD information.

If there is an existing DynDOLOD installation and only tree LOD should be generated for the current load order, start DynDOLOD with the existing DynDOLOD output so that all its plugins are loaded. Only check tree LOD in the advanced options. Generate it and then merge new output with new output, the updated plugins should replace the old plugins. But only if the old plugins were loaded at generation time.

However, even if not generating ultra tree LOD, a few trees will have LOD done in object LOD. If those cases are noticed to be wrong in the game, check object LOD to be generated as well.

If updating fails or has incomplete results, generate from scratch.

[MajorSollus]

9 minutes ago, sheson said:

It is not possible to generate LOD for a single mod, because . This is how LOD works. As explained at https://dyndolod.info/How-LOD-Works, "These different LOD types are basically optimized and pre-combined models or meshes and textures for a group of cells that represents the full content in the exterior cells."

Also, for example, the tree LOD atlas for a worldspace contains the billboard textures for all trees in a worldspace. Hence in order to have all require billboards added to it, tree LOD needs to be generate for the entire worldspace and because LOD only works and looks correctly for the load it was generated for, it really means that tree LOD needs to be generated for the current load order. This includes the sort of of plugins to not have engine bugs like stuck tree LOD.

See https://dyndolod.info/Updating

It is not possible for tree LOD only generation (e.g. object LOD generation is not checked) to change existing object LOD *.BTO meshes, which are what affect LOD for Dragonsreach.

If there is an existing DynDOLOD installation and only tree LOD should be generated for the current load order, start DynDOLOD with the existing DynDOLOD output so that all its plugins are loaded. Only check tree LOD in the advanced options. Generate it and then merge new output with new output, the updated plugins should replace the old plugins. But only if the old plugins were load at generation time.

However, even if not generate ultra tree LOD, a few trees will have LOD done in object LOD. If those cases are noticed in the game, check object LOD to be generated as well.

I will move forward with your advice. Thank you for the prompt reply.