Benchmark scenarios
Here is a list of the benchmark scenarios we would like to run:
-
A 10^16 eV proton shower 10^16, 30° zenith angle -
A 10^15 eV proton shower, 45° zenith angle, with 200 radio antennas -
A 10^15 eV proton shower, 85° zenith angle, with 200 radio antennas -
A run with 100 proton/gamma showers of 5 x 10^12 eV, 30° zenith angle, possibly cut off at 4,000 m asl. -
A run with 100,000 proton showers of 10^9 eV, varying zenith angles -
A run with 1,000 electron showers of 5 x 10^11 eV, 0.01° zenith angle, no magnetic field, with 99 Cherenkov telescopes -
A non-C7 use case like a shower transitioning from air to ice, compare in-ice part with PROPOSAL.
Things to take into account:
- Logging switched to „warn“ level
- All compiler optimizations switched on
- Energy cuts to be discussed and set same on C7 and C8
- Profiling as a follow-up step (using VTune probably)
- Also monitor memory usage?
- Evolve into validation tests? (Statistics over many showers! Regular tests not too long)
The goal is to check here which of the benchmarks we have already done and comment the run times, so we keep track of progress.