- Posts: 163
- Joined: Tue Apr 26, 2016 12:20 pm
I have a scene with a massive polygon object. I need it so dense because I'm using vertex offset to simulate a kind of transparency wipe.
That massive object is also a emitter region mesh for two massive particle systems and also pKiller region mesh for third emitter.
Then my PC got "lazy" Uh! I can hardly open the scene! It takes about 20sec just to assimilate the FBX geometry. But once it opens it goes reasonably fast. Perfect.
Now I want to add the last particle emmiter, but my system says it's too much. As soon as I create and link it, Fusion starts going wrong. Nodes in red, long waiting times for every single operation... OK, I need to optimize before going on.
Two questions:
Of course, I have only one instance of the FBXmesh loaded. Then it breaks in five branches, one for the renderer, another to a second renderer with different UV's, and three more branches to the particle nodes.
I assume Fusion is opening one single matrix in memory to hold those vertices. Not five....
If Fusion opens a new matrix for each use of the imported mesh, then that's my 1st problem. I'd import a low-poly version of the mesh for particles region branches, I don't need it to be so precise there.
Second question.
I'm tweaking Memory Swap file configuration and not very sure what's my best setting.
I traditinally config swap file size to be twice the physical RAM, and hold it in my fastest disk. But on this new PC it's not so easy.
- I have 64GB RAM and
- Only +-100GB available in fast M.2. SSD. (I have Fusion and Davinci cache in that disk too.)
- 256GB available in an auxiliar, not so fast, SSD.
- Two more SSD in RAID 0 for project files.
If I open the swap en the fast disk, it must be quite small, about 64GB, same size than RAM instead of double.
If I open it in the bigger disk It can be as big as I want, but that SSD is much slower than the M.2.
I can open multiple swap files, one per disk, I can also let Windows decide everything.
I'll do a lot of tests these days, but that's painfully tedious and slow... any advice will be appreciated.
That massive object is also a emitter region mesh for two massive particle systems and also pKiller region mesh for third emitter.
Then my PC got "lazy" Uh! I can hardly open the scene! It takes about 20sec just to assimilate the FBX geometry. But once it opens it goes reasonably fast. Perfect.
Now I want to add the last particle emmiter, but my system says it's too much. As soon as I create and link it, Fusion starts going wrong. Nodes in red, long waiting times for every single operation... OK, I need to optimize before going on.
Two questions:
Of course, I have only one instance of the FBXmesh loaded. Then it breaks in five branches, one for the renderer, another to a second renderer with different UV's, and three more branches to the particle nodes.
I assume Fusion is opening one single matrix in memory to hold those vertices. Not five....
If Fusion opens a new matrix for each use of the imported mesh, then that's my 1st problem. I'd import a low-poly version of the mesh for particles region branches, I don't need it to be so precise there.
Second question.
I'm tweaking Memory Swap file configuration and not very sure what's my best setting.
I traditinally config swap file size to be twice the physical RAM, and hold it in my fastest disk. But on this new PC it's not so easy.
- I have 64GB RAM and
- Only +-100GB available in fast M.2. SSD. (I have Fusion and Davinci cache in that disk too.)
- 256GB available in an auxiliar, not so fast, SSD.
- Two more SSD in RAID 0 for project files.
If I open the swap en the fast disk, it must be quite small, about 64GB, same size than RAM instead of double.
If I open it in the bigger disk It can be as big as I want, but that SSD is much slower than the M.2.
I can open multiple swap files, one per disk, I can also let Windows decide everything.
I'll do a lot of tests these days, but that's painfully tedious and slow... any advice will be appreciated.