sergeneren / volumetric-path-tracer Goto Github PK

I use the VPT alpha 1.03 pre-built x64 binaries and assets.
Win 10, 980 Ti, cuda 10.2, drive 442.59.

VPT threw a "ptxas fatal: memory allocation error" when tried with a RTX-2060. This is probably related to arch_30, sm_30 compilation of ptx files of gvdb library and/or the VPT library render kernel.

This issue will be moved to in_progress when a test environment with rtx series cards is set.

Writing image with key "S" causes if it's used after window size is changed. Maybe image buffer needs to be updated.

ImGui version 1.72 removed "IsMouseHoveringAnyWindow" function, so how can I built the vs.file? Thank you~~~

Is there any method or article refer to?

I'm trying to render the sample dragon_with_xform vdb file but got a black window
here is the log info. looks like bound volume is incorrect
thank you.

number of vdbs: 1

Grid: density float
Dense Grid:
Memory footprint: 3.355 MB
Dimensions of grid : 141 x 99 x 63
Number of voxels: 879,417
Bounding box of voxels: [-70, 1, -31] -> [70, 99, 31]
Memory layout: XYZ (dissimilar to VDB)
value count: 879417
max density: 1.000000
0.079540 0.074172 -0.050714 0.000000
-0.054848 0.093727 0.051058 0.000000
0.071170 -0.010663 0.096027 0.000000
-0.200000 0.200000 -0.200000 1.000000

Number of volumes: 1
Computing volume bounding boxes... done! Computation took 0.180896 ms.
Computing scene bounding box... done! Computation took 0.249856 ms.
Total pre-computation time for scene was 0.430752 ms.

Scene boundingbox:
pmin: 3.40282e+38, 3.40282e+38, 3.40282e+38
pmax: -3.40282e+38, -3.40282e+38, -3.40282e+38
Computing Morton codes... done! Computation took 0.059232 ms.
Sort volumes... done! Sorting took 0.20448 ms.
Building radix tree... done! Took 0.0656 ms.
Building BVH... done! Took 0.076704 ms.
Total BVH construction time was 0.406016 ms.
Root node bounds
pmin.x: 340282346638528859811704183484516925440.000000, pmin.y: 340282346638528859811704183484516925440.000000, pmin.z: 340282346638528859811704183484516925440.000000
pmax.x: -340282346638528859811704183484516925440.000000, pmax.y: -340282346638528859811704183484516925440.000000, pmax.z: -340282346638528859811704183484516925440.000000
num volumes for child 3-0 is 0 max extinction: 0.000000
num volumes for child 3-1 is 0 max extinction: 0.000000
num volumes for child 3-2 is 0 max extinction: 0.000000
num volumes for child 3-3 is 0 max extinction: 0.000000
num volumes for child 3-4 is 0 max extinction: 0.000000
num volumes for child 3-5 is 0 max extinction: 0.000000
num volumes for child 3-6 is 0 max extinction: 0.000000
num volumes for child 3-7 is 0 max extinction: 0.000000

The project can't build use vs2017.
Thanks.

Pre Report
Please make sure you have completed below checklist before submitting a report

• Build VPT with Debug Level "Log"
• Run VPT to a log file [e.g ".\vpt.exe fireball.ins >log.txt" ]
• Attach the log file here with a link

Describe the bug
VPT loads a field If it is present but has 0 voxels. Main loop crashes

To Reproduce
Steps to reproduce the behavior:

1. Load a vdb with a field having 0 active voxels
2. Start rendering loop

Expected behavior
VPT ignoring the field and returning no density, emission or color

Screenshots

Desktop (please complete the following information):

• GPU: GTX 680
• GeForce Game Ready Driver 441.66

Additional context
Add any other context about the problem here.

Win 10, cuda 10.1, vs 2017

GVDB CUDA ERROR:
Launch status: file not found
Kernel status: no error
Caller: VolumeGVDB::LoadFunction
Call: cuModuleLoad
Args: D:/PROJECTS/gvdb-voxels/_output/lib/cuda_gvdb_module.ptx
Error. Application will exit.

i don't compile gvdb-voxels myself, so this is because some hardcode the gvdb path?

ps, how to compile shared_cudpp which in the gvdb-voxels?
use the cmake, can't compile the shared _cudpp.

Thanks.