Fomenting Mutiny

[Legacy_Zarathustra217]

               

About supermodels - please correct me if I'm wrong - but I'm pretty sure once you compile a model, it makes no difference in size whether it has the animations in it's own file or in a supermodel, since they are always compiled into the file.

 

It's still a good practice though for many things because it assures that fixes and additions in the supermodel automatically transfers to other models.

 

Anyway, it all sounds great, keep it up !

               
               

               
            

[Legacy_Rolo Kipp]

               

<writing a thousand words...>

 

Note: Sword spider supermodels Giant spider. Edit: Er, the Giant Spider is supermodel for the sword spider. Hell, just look at the file sizes! :-P :-)

 

Edit2: Thought... on some models (notably models without anims), the compiled version is larger (about 33% in this case). What is the load time performance hit of a compiled supermodel with a dozen uncompiled children models compared to the same compiled supermodel with compiled children?

 

Or put another way, is it worthwhile compiling animation-less child models for load speed vs larger files? A 33% reduction in a significant portion of the hak could well be worth more than a momentary loading (once per model, not instance) lag.

 

<...with a jpeg quill>

               
               

               
            

[Legacy_Zarathustra217]

               

Well, those file sizes suggests I'm wrong about supermodel animations being compiled into the file, since otherwise it doesn't make sense the spider giant model is so much bigger. Might be worth testing though.

 

Anyway, I don't know the exact performance hit, but I think it's fair to assume that there is one. And in my view, performance should almost always be prioritized higher than download size. Download size reduction by leaving those files uncompiled would also be negligible given that it'll be the small files you reduce in size.

               
               

               
            

[Legacy_Rolo Kipp]

               

<speaking pure...>

 

Well, those file sizes suggests I'm wrong about supermodel animations being compiled into the file, since otherwise it doesn't make sense the spider giant model is so much bigger. Might be worth testing though.

 

Anyway, I don't know the exact performance hit, but I think it's fair to assume that there is one. And in my view, performance should almost always be prioritized higher than download size. Download size reduction by leaving those files uncompiled would also be negligible given that it'll be the small files you reduce in size.

 Definitely worth testing (which I do not have the time for at the mo....:-P )

However, I'd think the performance impact exactly the opposite. Disk access should be far slower than model compiling on modern CPUs. Both operations are performed once, So does the time spent loading 33% more file balance, overbalance or underbalance the time spent compiling a model?

 

If they balance out, then the convenience and filesize of an uncompiled mdl would tip the scales again.

<...conjecture>

               
               

               
            

[Legacy_Pstemarie]

               

Well, I ran some benchmarks on my system over the last few days using a hak with 200 models in an uncompiled state vs. the same hak with the models compiled. Load time differences at those file counts were neglible; however, it was slightly slower loading the uncompiled models. The biggest hit came with animated models - creatures.

 

Thus, based on my limited findings, if you increase the file count size it stands to reason that the loadtime spearation between the two state will increase as well. Keep in mind that we're talking seconds here and in a basic module that had nothing in it except a starting area built from the microset.

               
               

               
            

[Legacy_Rolo Kipp]

               

<nods...>

 

Right.

And the biggest hit was the animated models, the supermodels (even if it's a super of one).

If the *supers* are compiled (thus reducing filesize by close to half) but the non-animated child models are *not* compiled (saving filesize, again)... well, that just really appeals to me :-)

 

<...like he understands>

               
               

               
            

[Legacy_Zarathustra217]

               

New versions of the cep2_core0 and cep2_core2 haks:

 

https://dl.dropboxus...2/cep2_core0.7z

 

https://dl.dropboxus...2/cep2_core2.7z

 

I've fixed a few issues and changed to using CM 3.52 (now also using "fine" snap rather than "binary", since some models relied on the larger precision).

 

Also, I received some fixed models from Stefan/BlackRider (dag_tnocliff1, dag_tnocliff2, nwn2merch3, nw2temple_e3) and a missing tail model from DoD (c_tailliz2) send by a person named David (don't know his forum or community name).

 

For the changed files alone (mainly for you, TAD):

 

https://dl.dropboxus...677912/core0.7z

 

https://dl.dropboxus...677912/core2.7z

 

               
               

               
            

[Legacy_Lazarus Magni]

               

I am a bit confused... how did we go from cep 2.4 to 2.6? Is 2.5 already out there? And if so where?

               
               

               
            

[Legacy_The Amethyst Dragon]

               

I am a bit confused... how did we go from cep 2.4 to 2.6? Is 2.5 already out there? And if so where?

2.5 was supposedly "in the works" by the previous "official CEP Team"...but nothing has been heard for nearly 2 years now (and it's been over 3 years since 2.4 was released).

 

When the name of a "CEP 3" was floated, many wanted to save that designation for a full restructuring of everything (hak names, adding/removing content, full modularity, and so on)...way beyond what I'm up for as a PW admin that's been building with the CEP since version 2.1.  I decided that going with the 2.60 version number would be a slight jump up from whatever the old CEP Team never finished/released, but not a completely new (and incompatible) content compilation. Once the next update comes out, I'm thinking that smaller, more manageable versions will follow that won't warrant a 1/10 of a point jump (2.61, 2.62, etc.).

               
               

               
            

[Legacy_Lazarus Magni]

               

That makes sense. Thanks for the clarification AD.

               
               

               
            

[Legacy_Vanya Mia]

               

... I've just spotted this when looking for up to date information for updating D20 Modern 2.0 and panicked.

 

I don't know how far you have got down the road with this as a project, or even if it will matter, but I had discussions with one point about compatibility with D20 Modern haks - in so much as trying to avoid the lines we had use for D20 so that updates were easier. The later versions of CEP 2.4 didn't manage to do that, and while a lot of what is in there wasn't a great loss to a modern setting. Is there any possibility of this for a new version?

               
               

               
            

[Legacy_The Amethyst Dragon]

               

... I've just spotted this when looking for up to date information for updating D20 Modern 2.0 and panicked.

 

I don't know how far you have got down the road with this as a project, or even if it will matter, but I had discussions with one point about compatibility with D20 Modern haks - in so much as trying to avoid the lines we had use for D20 so that updates were easier. The later versions of CEP 2.4 didn't manage to do that, and while a lot of what is in there wasn't a great loss to a modern setting. Is there any possibility of this for a new version?

it depends. Do you have a list of the various 2da lines you've used?

 

I won't be changing lines that were already in use for as of version 2.4 (so as to not break backwards compatibility with existing CEP-using projects), but anything I've worked on since then is not yet set in stone (as far as 2da lines go). I happened to have color-coded my working 2da file for 2.60, so I could easily see which ones are new/changed for this version.

 

If I can get a list of 2da lines used just for d20 Modern stuff, I can see what I can do about moving some things around.

 

If anyone is interested in seeing what lines I've altered so far, it's on the changelog on the CEP wiki (link in signature).

               
               

               
            

[Legacy_Vanya Mia]

               

I've just started updating to 2.4, beginning with the clothing hak and not added anything new yet for D20 except lines for the things that clashed with D20 existing ones and I actually felt were worth making available. I can supply you a spread sheet with all the D20 lines marked, and let me know what new ones you want to use, and I can avoid them.

 

The other option is to compare notes as we go? I'm starting to use the forum on the Neverwintervault.org site for developments, in fact that's how I learned this is happening, so we can probably arrange to communicate there?

               
               

               
            

[Legacy_The Amethyst Dragon]

               

I've just started updating to 2.4, beginning with the clothing hak and not added anything new yet for D20 except lines for the things that clashed with D20 existing ones and I actually felt were worth making available. I can supply you a spread sheet with all the D20 lines marked, and let me know what new ones you want to use, and I can avoid them.

 

The other option is to compare notes as we go? I'm starting to use the forum on the Neverwintervault.org site for developments, in fact that's how I learned this is happening, so we can probably arrange to communicate there?

New Vault, here, either way works, since I use both forums (and the CEP forum, and the CEP wiki, and email, and...).

 

If you can send me a list of lines you've already used in the various 2das, I can most likely avoid those lines (if they don't already have CEP 2.4 content there). Otherwise, what I've been doing is filling in "holes" that haven't been used in the CEP reserved ranges.  The only thing I've done outside of the reserved ranges is for portraits.2da, where the old range got filled up...I added 6000 lines of "user" padding, then marked another range as reserved for CEP.

               
               

               
            

[Legacy_Tarot Redhand]

               

Thanks for that AD. As I've tried to make my stuff compatible with the CEP (exception visualeffects.2da) I appreciate you not using the user slots.

 

TR