Here is full comparision of when disabled vs enabled
Here is full comparision of when disabled vs enabled
* Double-precision BVH build and traversal
* Validation (in speedtest) of shadow ray queries
* Faster GPU BVH4 traversal (+ ~5%)
* TLAS preparations
* BVH for list of AABBs
* ...More.
Link: github.com/jbikker/tiny...
* Double-precision BVH build and traversal
* Validation (in speedtest) of shadow ray queries
* Faster GPU BVH4 traversal (+ ~5%)
* TLAS preparations
* BVH for list of AABBs
* ...More.
Link: github.com/jbikker/tiny...
New in this version:
Ultra-fast CPU BVH builds using the threaded, binning SAH builder (BVH::BuildAVX).
TinyBVH now builds a BVH for a 220k triangle model in 7ms, twice as fast as before.
Dev build: github.com/jbikker/tiny...
New in this version:
Ultra-fast CPU BVH builds using the threaded, binning SAH builder (BVH::BuildAVX).
TinyBVH now builds a BVH for a 220k triangle model in 7ms, twice as fast as before.
Dev build: github.com/jbikker/tiny...
* New C++17 threading system
* TinyOCL now uses OpenCL 3.1.1
* Fix: watertight tri test for shadows
* Added BVH::BuildAABB
EA SEED (Gigi), SideFX and Remedy are using TinyBVH! Check it for yourself: github.com/jbikker/tiny...
* New C++17 threading system
* TinyOCL now uses OpenCL 3.1.1
* Fix: watertight tri test for shadows
* Added BVH::BuildAABB
EA SEED (Gigi), SideFX and Remedy are using TinyBVH! Check it for yourself: github.com/jbikker/tiny...
github.com/jbikker/tiny...
This release has been prepared in collaboration with @wjakob.bsky.social. As of 1.9.0, TinyBVH switched from 'single-header' to 'header-only', because the tiny thing was getting unwieldy...
Overview of what's new:
(1/3)
github.com/jbikker/tiny...
This release has been prepared in collaboration with @wjakob.bsky.social. As of 1.9.0, TinyBVH switched from 'single-header' to 'header-only', because the tiny thing was getting unwieldy...
Overview of what's new:
(1/3)
BVH bvh;
bvh.Build( myVerts, triCount );
bvh.Intersect( ray );
Link to repo: github.com/jbikker/tiny...
(1/3)
BVH bvh;
bvh.Build( myVerts, triCount );
bvh.Intersect( ray );
Link to repo: github.com/jbikker/tiny...
(1/3)
📉 <2GB of BVH memory, vs. up to 80GB for conventional animated BVHs.
⚡ 500M+ animated triangles ray traced per frame at 60+ FPS.
See the breakdown ⬇️
gpuopen.com/learn/how-te...
📉 <2GB of BVH memory, vs. up to 80GB for conventional animated BVHs.
⚡ 500M+ animated triangles ray traced per frame at 60+ FPS.
See the breakdown ⬇️
gpuopen.com/learn/how-te...
* Improved speed of FullSweep (thanks Julian)
* Reverted JobSystem to basic threading because of instabilities
* Wave_tracer now uses TinyBVH!
* Fixed speedtest of GPU BVH
Because of the threading issue: Please upgrade to 1.6.6 asap.
github.com/jbikker/tiny...
* Improved speed of FullSweep (thanks Julian)
* Reverted JobSystem to basic threading because of instabilities
* Wave_tracer now uses TinyBVH!
* Fixed speedtest of GPU BVH
Because of the threading issue: Please upgrade to 1.6.6 asap.
github.com/jbikker/tiny...
github.com/jbikker/tiny...
Fast GPU BVH traversal using a 4-wide BVH.
Code is in traverse.cl and included in speedtests.
One more to go: CWBVH!
github.com/jbikker/tiny...
Fast GPU BVH traversal using a 4-wide BVH.
Code is in traverse.cl and included in speedtests.
One more to go: CWBVH!
Reposts appreciated :)
zeux.io/2025/03/31/m...
Reposts appreciated :)
zeux.io/2025/03/31/m...
to
I know, I'll test against bounding spheres around the triangles
to
I know, I could nest the bounding spheres and discard whole regions at once
to
This would be better with bounding boxes
to
I've reinvented BVH, haven't I?
to
I know, I'll test against bounding spheres around the triangles
to
I know, I could nest the bounding spheres and discard whole regions at once
to
This would be better with bounding boxes
to
I've reinvented BVH, haven't I?
Spatial splits significantly reduce the cost of ray queries in a BVH (by about 25%).
An SBVH can still be optimized further with the BVH-post-optimizer.
github.com/jbikker/tiny...
Spatial splits significantly reduce the cost of ray queries in a BVH (by about 25%).
An SBVH can still be optimized further with the BVH-post-optimizer.
github.com/jbikker/tiny...