About
Hi, I am Reiner.
I hold a M.Sc. degree in Computer Science from the Karlsruhe Institute of Technology (KIT) with a specialization in compiler construction and computer graphics. I am a doctoral researcher in the computer graphics group of Carsten Dachsbacher with a focus on real-time visualization. My core professional interests lie in optimizing compilers and manual optimization for GPUs—especially when targeting cross-platform deployment. More specifically, I am interested in:
- GPUs - scheduling, load balancing, performance analysis and optimization, portability, multi-GPU programming, shader compilers and drivers
- Compilers for GPUs and Deep Learning Accelerators - optimization and analysis, parallelization, profile guided optimization, memory allocation and packing, lossy/non-conservative optimizations for deep learning and computer graphics, static specialization, joint design of hardware and algorithms
- Computer Graphics and Visualization - real-time rendering, rendering on many devices, large-scale data processing including acceleration structures
Contact
You can reach me via hello@reinerdolp.com.
Publications
Reiner Dolp, Johannes Hanika, and Carsten Dachsbacher
Proceedings of the ACM on Computer Graphics and Interactive Techniques, Volume 7, Issue 1, I3D, 2024
We introduce a technique to apply a series of increasingly dilated filters or convolutions quickly. This technique implements a recursive permutation of GPU thread indices. Intuitively, this permutation sorts and groups threads by their accessed memory addresses to increase memory throughput.