6 Commits

Author SHA1 Message Date
Syoyo Fujita
4de26ada7a Add Mandelbulb procedural mesh generation to benchmarks
Implement mathematically accurate Mandelbulb fractal geometry generation
with configurable level-of-detail parameter. Includes:

- MandelbulbGenerator class with distance estimation algorithm
- Surface extraction using marching cubes approach
- Normal calculation for proper lighting
- Vertex deduplication for optimization
- Integration with existing ubench benchmark framework
- Multiple LOD benchmarks (16, 32, 64 resolution)
- USD GeomMesh creation and USDA file export benchmarks

Performance results show ~17ms for LOD 16 generating ~335 vertices.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-12 22:12:12 +09:00
Syoyo Fujita
2066386e4a Add microbenchmark framework 'criterion' 2025-02-10 06:09:06 +09:00
Syoyo Fujita
5f27971086 Refactor codes. 2022-08-06 23:13:34 +09:00
Syoyo Fujita
490d2c94d5 Implement some of Tydra render data structure. 2022-07-31 22:38:14 +09:00
Syoyo Fujita
795a4509d9 Use tinyusdz-customized linb::any. 2022-07-21 22:20:13 +09:00
Syoyo Fujita
0967b7561b Add benchmark(to measure the performance of C++ data structure). 2022-07-21 19:13:53 +09:00