No, its the same as in the modifier:
1 Like
As 3DArtGuy just replied, the two choices in 4.2 are just the current booleans. I am working on the āmagical new oneā but sadly I was wrong about the 500 times faster. The measurement I made was of the code that just registered the operation; the actual operation happens only on export. It will be faster than the current float solver for large meshes, but probably only a low multiple of times faster at best. Sorry. I still think it will be worth it, because it should be much more reliable than the current float solver.
26 Likes
Well, I think everyone would warmly welcome even a 1% improvement and more reliability. Theyāre net positives
1 Like
Especially for animations, more reliability very nice.
ālow multiples of times fasterā sounds like an amazing improvement to me. Thank you for continuing to work on it 
2 Likes
You can link a collection instanceā¦
You can add a geometry node group to default cube that references the collection instance object.
You can then find and replace materials on the collection instance.
You can delete faces from the collection instance.
Why donāt they just let us put modifiers on the collection instance object?
2 Likes
this work is important* however again - you would get a larger performance speed up from saving structures that donāt change from frame to frame and re-using them, right?
(like bvhtree in local space for a mesh)
like bvhtree in raycast (and potentially Boolean right?)
kdtree for nearest neighbor type stuff*
( sample surface / sample nearest / index of nearest in mesh etc )
There are all sorts of possibilities for caching things related to he boolean calculations. I may eventually get to some of them but right now my mind is just on getting the basic functionality done. (And then there are other modeling things that Iāve been putting off, like Bevel V2.)
10 Likes
@BluePrintRandom:
Caching things is usually easy. Knowing if and when a cache gets invalid not so much.
greetings, Kologe
1 Like
manipulating some attributes should tag a mesh as ādirtyā and need re-evaluation.
else keep the one you have - if the tree/structure can be in local space you can store it in the āoriginatorsā data ā grab that if it exists or build one if its missing or dirty.
(for objects on the scene its a little more clear cut than āvirtualā objects created / used inside a geometry node setup)
Someone comes in with arguably promising captures of a multiphysics solver that already has some kind of integration with Blender, and instead of working proactively with the developers, all they manage to do is outline the shortcomings of geonodes as a justification for doing the sims in their own node system (instead of improving geonodes to handle it) and fail to realize theyāre talking to one of the principal developers of geonodes (all the while claiming they had a ābig discussionā with them). It barely started and itās already looking bad. As a user I am exceedingly frustrated by this kind of exchange.
11 Likes
You need to mention a comment praising him and asking Blender institute to give him money from 10 minute old account with same English as OP, who is totally different person of course.
11 Likes
haha, i mean, i would be excited for it but its the second system i saw in two weeks. I wonder why but the new fluid system (forgot the name) seems legit nice and already available to buy.
I wish he would work with themā¦ or anyone wanted to implement a āuniversalā solverā¦
EDIT: The insta of that guy for example uses AI generated Thumbnails and the videos arent looking even close to them from the quality of simulationā¦
I can understand him wanting to earn some money as well as him trying to capitalize on the lack of modern physics solvers in Blender. I also understand that devtalk is not the platform for self marketing. That being said, Iād gladly pay for an addon, even if non geo nodes, if that means me being able to do some softbody or fluid stuff which is better than what we have currently.
While Iād prefer a polished geo/sim nodes solution, somehow this comes to mind as well -
9 Likes
Upbge already has a pretty robust physics integration and is open source/blender fork.
Youle did the legwork mostly and Loki.
Its a object level sim, however geometry nodes work in game, you can even advance the timeline in game using bpy to run simulation nodes, and we can use bpy in game to write fcurve animations.
We now have matrix nodes at least (which is a step towards solver nodes in asset form because of how they are a key building block to creating them).
Blender VFX should be supercharged once the devs. have resolved all issues preventing them from shipping group node assets for things like particles and many other effects (I for one can at least do particles, but I have never been able to wrap my head around matrices).
Judging purely from the resultsā¦those physics simulations were out-of-this-world spectacular. (Obviously canāt judge the speed of simulation or user interface)
And Iām all for the developer getting paid appropriately to make it a reality.
I understand the BFās philosophical point-of-view but I tend to think there was also is some envy and turf-protection going on w/some of those comments in that dev thread.
I canāt help but think of Brecht and his development of Cyclesā¦on his own time. He ultimately was invited to finish out the work as a paid employee. I wonder what would have happened if he got such a hostile reaction?
Sometimes external add-ons change the game and raise the whole value, excitement and usefulness of the DCC. I am reminded what x-particles has done for c4d.
6 Likes
Comments like this really go a long way towards exemplifying how friendly and welcoming the Blender community is to new ideas and contributors 
yāall canāt complain about the small number of devs with one breath and devote yourself to chasing interested new ones away with the other
2 Likes
Yea Iām not sure itās warranted either, itās probably some other guy who happens to know the developer of āTangraā from other projects.
Anyway, I agree that this sort of development (or any sort really) should be eligible for a grant, provided this is the route the developer wants to take and heās a good fit for the project. People absolutely need to get paid, and contributing from the goodness of your heart shouldnāt be a requirement. I hope he chooses to adhere to the foss philosophy and considers a Blender integration rather than selling his product on the side. This would be an immense benefit for everyone, instead of for a few.
3 Likes
I 100% agree with this, but I think itās important to look at how people get hired at Blender. For developers, almost everyone was contributing in some way before they got a grant or hired. And this is not a coincidence. For the foundation, itās too risky to hire someone just from an āapplicationā.
Having contributed before gives everyone a good idea about how someone works, what skill they have, how they deal with review, etc. etc.
9 Likes