Kernel Tuner
Kernel Tuner greatly simplifies the development of highly-optimized and auto-tuned CUDA, OpenCL, and C code, supporting many advanced use-cases and optimization strategies that speed up the auto-tuning process.
Observing processes that are inaccessible to optical telescopes
A new era in the exploration of the Universe has begun. The distant Universe has so far only been accessible to us via photons of different wavelengths. But with the recent discovery of high-energetic extraterrestrial neutrinos new and unique opportunities arise to observe processes that are inaccessible to optical telescopes.
The KM3NeT neutrino telescope, currently under construction In the Mediterranean Sea, enables researchers to fully exploit this new opportunity. The Mediterranean Sea appears to be an ideal place for this future installation: it provides water of excellent optical properties at the right depth and excellent shore-based infrastructure for marine operations and on-shore data processing. Light created by particles in neutrino interactions in the Earth or in water is collected with a large 3-dimensional grid of sensitive photo detectors. All information from the detectors is sent to shore for further filtering and processing.
The current hardware and time constraints do not allow the most detailed processing of the information come from the detectors. This means that not all interesting event signatures will be stored. Current algorithms focus on the selection of specific signatures in the hit correlations which are expected from high energetic interactions. If we can enhance the strict filtering to a full online event reconstruction, so that we can retrieve accurate information on the potential neutrino candidates in real time, the data could be recorded more efficiently.
An accurate online reconstruction of the detailed event properties can also provide interesting neutrino triggers or feedback of external triggers to other astronomical facilities like optical or gamma ray observatories. At these observatories these signals can immediately be followed up to not miss out on interesting events in the Universe.
A database-integrated KM3NeT solution for automated processing
Interpretable large scale deep generative models for Dark Matter searches
Large scale statistical data analysis in particle physics
The intelligent Dark Matter survey
Modern big data front and backends in the hunt for Dark Matter
The evolution of embedded star clusters
Kernel Tuner greatly simplifies the development of highly-optimized and auto-tuned CUDA, OpenCL, and C code, supporting many advanced use-cases and optimization strategies that speed up the auto-tuning process.