The Si3 project, financed by ERC advanced grant (grant agreement 101142294) and in part for my PhD expenses by the Swedish Research Council (grant number 2022-03688_VR), aims to develop an innovative 3D monolithic silicon detector for nuclear medical imaging such as Positron Emission Tomography or Single Photon Emission Tomography. The operational principle of the device is going to be similar to Compton Cameras but bearing substantial differences. Firstly, it will not need a distinction between scatterer and absorber, and it will also be able to reconstruct events that scatter more than twice or that do not deposit the full energy in the detector. Thanks to the sequencing algorithms used, it will be possible to rejecting the gamma rays that scattered in the patient further enhancing the image quality. Thanks to these improvements and the small pixel size (in the order of 100 µm) the Si3 detector will be able to achieve increased spatial resolution, and a theoretical efficiency increase up to 10^6 compared to standard state of the art detectors. We will use Allpix squared with Geant4 to accurately simulate gamma ray interactions with the detector, as well as electron tracks, charge propagation and collection by the electrodes. Given the complexity and the desired precision, this workload will require heavy computational cost, both on the GPU but especially on the CPU, when wanting to test different parameters for the geometries of the detector, source or phantoms. With Dardel it will be possible to use thread-parallel Allpix runs and large job arrays, shortening iteration cycles, and substantially reducing the time required to optimize the parameters and gather enough data to then test the reconstruction algorithms. Without Dardel, i would have to rely on my laptop, which would greatly hinder the progress of the project, especially now when we are moving from smaller investigations to whole system level simulations and characterizations, requiring millions of incident photons for each set of parameters to gather enough statistic.