"vis. cache" = "env cache" ?lacilaci wrote: Mon Jun 17, 2019 7:35 pm Btw. I do hope by "closing pending stuff" you also mean fix for vis. cache bug and also make that env cache work on gpu![]()
(But yes, that is the idea)
"vis. cache" = "env cache" ?lacilaci wrote: Mon Jun 17, 2019 7:35 pm Btw. I do hope by "closing pending stuff" you also mean fix for vis. cache bug and also make that env cache work on gpu![]()
provisory wrote: Sat Jun 01, 2019 7:30 pm I have two ideas how to make PGI cache (more) unbiased. ...
1. Cyclically re-generating the cache
Dade wrote: Mon Jun 17, 2019 2:56 pm I added the support for periodic update of the PGI caustic cache to PATHCPU
There is a problem with erratic photon distribution from infinite light leading to ghost reflections in glossy surfaces, that could be helped with cyclic cache.Dade wrote: Mon Jun 17, 2019 6:05 pmThe same process would be trivial to do for indirect cache (i.e. it would require to change only one line of code) but it would be practically useless: indirect cache already works well as it is.Sharlybg wrote: Mon Jun 17, 2019 5:12 pm Huuuh do you mean it can be the same for Indirect cache ? so it turn as an online cache ?
Persistent (indirect) cache should work, persistent (caustic) cache should work too but it quite pointless as the saved cache will be used only for the very first cycle ... don't talk about "Cycles" ... it confuse meepilectrolytics wrote: Mon Jun 17, 2019 8:23 pm Cyclic cache prevents persistent cache, that has to be exposed properly in Blender.
This should already work (see above).epilectrolytics wrote: Mon Jun 17, 2019 8:23 pm It could get complicated like this:
Rendering an animation with persistent indirect cache plus a cyclic caustic cache.
This is normal, the update is done at least after X samples, not exactly at X samples (threads continue to asynchronous render while the need of an update is recognized, they a will always render a bit more before they can stop for the update, this will be even more noticeable with GPUs).epilectrolytics wrote: Tue Jun 18, 2019 8:32 am Also the updating happens one sample later as specified:
If I want cache updates at sample 8,16,24,32... I have to choose 7 as parameter.
Are you sure? In my tests it worked in Blender, although the console messages indeed do not appear.epilectrolytics wrote: Tue Jun 18, 2019 8:32 am Only it does not really work yet:
Cache update does not seem to happen as also indicated by console view which shows no entries during render.
New test scene works, the reason was this:B.Y.O.B. wrote: Tue Jun 18, 2019 9:29 am Are you sure? In my tests it worked in Blender, although the console messages indeed do not appear.
My first tests in LuxCoreUI showed when I enter periodic sample step of 7, the caustic pattern in the render view will change after 8,16,24... samples (corresponding with console print).Dade wrote: Tue Jun 18, 2019 8:49 amThis is normal, the update is done at least after X samples, not exactly at X samples (threads continue to asynchronous render while the need of an update is recognized, they a will always render a bit more before they can stop for the update, this will be even more noticeable with GPUs).epilectrolytics wrote: Tue Jun 18, 2019 8:32 am Also the updating happens one sample later as specified:
If I want cache updates at sample 8,16,24,32... I have to choose 7 as parameter.
Maybe I'm stupid again, but I don't get it to work.Dade wrote: Tue Jun 18, 2019 10:26 am Added the support to OpenCL for PhotonGI caustic cache update.
It crashes for me too, when I stop the render during the cache update. (Win 10, CPU, not OpenCL)B.Y.O.B. wrote: Mon Jun 17, 2019 6:36 pm By the way, I get a crash when I cancel the render while the periodic update is running.
OS is Linux.