Fast GPU-based ray tracing in radial GRIN lenses

Area of Science:

• Optics and Photonics
• Computational Science

Background:

• Radial gradient-index (GRIN) lenses feature a refractive index varying perpendicular to the optical axis.
• Accurate ray tracing is essential for GRIN lens illumination and rendering analyses.

Purpose of the Study:

• To evaluate the performance of graphics processing unit (GPU) accelerated ray tracing for radial GRIN lenses.
• To compare GPU performance against central processing unit (CPU) analysis for different refractive index distribution models.

Main Methods:

• Implemented ray tracing algorithms for radial GRIN lenses using both power series expansion and cubic spline interpolation for refractive index distribution.
• Performed comparative analysis of computation times between GPU and CPU implementations.

Main Results:

• GPU-based ray tracing was, on average, 19 times faster than CPU-based analysis for both distribution types.
• Achieved average GPU effective performance of 90% for power series and 40% for spline interpolation.

Conclusions:

• GPU acceleration offers substantial speedups for radial GRIN lens ray tracing.
• The findings highlight the effectiveness of GPU-accelerated ray tracing for demanding optical analyses involving GRIN lenses.