arXiv subjects

Sebastian Dittmeier

Publications and source records attributed to Sebastian Dittmeier.

11 recordsLinked to original sources

arXiv2026

PQuantML: A Tool for End-to-End Hardware-aware Model Compression

PQuantML is a new open-source, hardware-aware neural network model compression library tailored to end-to-end workflows. Motivated by the need to deploy performant models to environments with strict latency constraints, PQuantML simplifies training of compressed models by providing a unified interface to apply pruning and quantization, either jointly or individually. The library implements multiple pruning methods with different granularities, as well as fixed-point quantization with support for High-Granularity Quantization. We evaluate PQuantML on representative tasks such as the jet substructure classification, so-called jet tagging, an on-edge problem related to real-time LHC data processing. Using various pruning methods with fixed-point quantization, PQuantML achieves substantial parameter and bit-width reductions while maintaining accuracy. The resulting compression is further compared against existing tools, such as QKeras and HGQ.

cs.LG↗

arXiv2025

AIE4ML: An End-to-End Framework for Compiling Neural Networks for the Next Generation of AMD AI Engines

Efficient AI inference on AMD's Versal AI Engine (AIE) is challenging due to tightly coupled VLIW execution, explicit datapaths, and local memory management. Prior work focused on first-generation AIE kernel optimizations, without tackling full neural network execution across the 2D array. In this work, we present AIE4ML, the first comprehensive framework for converting AI models automatically into optimized firmware targeting the AIE-ML generation devices, also with forward compatibility for the newer AIE-MLv2 architecture. At the single-kernel level, we attain performance close to the architectural peak. At the graph and system levels, we provide a structured parallelization method that can scale across the 2D AIE-ML fabric and exploit its dedicated memory tiles to stay entirely on-chip throughout the model execution. As a demonstration, we designed a generalized and highly efficient linear-layer implementation with intrinsic support for fused bias addition and ReLU activation. Also, as our framework necessitates the generation of multi-layer implementations, our approach systematically derives deterministic, compact, and topology-optimized placements tailored to the physical 2D grid of the device through a novel graph placement and search algorithm. Finally, the framework seamlessly accepts quantized models imported from high-level tools such as hls4ml or PyTorch while preserving bit-exactness. In layer scaling benchmarks, we achieve up to 98.6% efficiency relative to the single-kernel baseline, utilizing 296 of 304 AIE tiles (97.4%) of the device with entirely on-chip data movement. With evaluations across real-world model topologies, we demonstrate that AIE4ML delivers GPU-class throughput under microsecond latency constraints, making it a practical companion for ultra-low-latency environments such as trigger systems in particle physics experiments.

cs.LG↗

arXiv2025

TelePix2: Full scale fast region of interest trigger and timing for the EUDET-style telescopes at the DESY II Test Beam Facility

With increasing demands by future and current upgrades of particle physics experiments on rate capabilities and time resolution, the requirements on test beams are also increasing. The current infrastructure at the DESY II test beam facility includes particle tracking telescopes with long integration times, no additional timing but excellent spatial resolution. This results in readouts with multiple particles per trigger, causing ambiguities in tracking and assigning particles to triggers. Also, it is likely not to trigger on particles that pass through a small device under test, leading to inefficient data taking. These issues can be solved by adding TelePix2 as a timing and flexible region of interest trigger layer. TelePix2 is a full scale HV-CMOS chip based on the successful small scale prototype TelePix. The DAQ system and the sensors performance featuring efficiencies above 99 % and a time resolution of 3.844(2) ns are presented. The integration into EUDAQ2 and the AIDA-TLU to seamlessly work in the test beam environment as well as into the analysis chain is described. First successful use cases are highlighted to conclude that TelePix2 is a well-suited timing and trigger layer for test beams

hep-ex↗

arXiv2022

TelePix -- A fast region of interest trigger and timing layer for the EUDET Telescopes

Test beam facilities are essential to study the response of novel detectors to particles. At the DESY II Test Beam facility, users can test their detectors with an electron beam with a momentum from 1-6 GeV. To track the beam particles, EUDET-style telescopes are provided in each beam area. They provide excellent spatial resolution, but the time resolution is limited by the rolling shutter architecture to a precision of approximately 230 $\mu$s. Since the demand on particle rates -- and hence track multiplicities -- is increasing timing is becoming more relevant. DESY foresees several upgrades of the telescopes. TelePix is an upgrade project to provide track timestamping with a precision of better than 5 ns and a configurable region of interest to trigger the telescope readout. Small scale prototypes have been characterised in laboratory and test beam measurements. Laboratory tests with an injection corresponding to 2300 electrons show a S/N of above 20. Test beam characterization shows efficiencies of above 99% over a threshold range of more than 100 mV and time resolutions of 2.4 ns at low noise rates.

physics.ins-det↗

arXiv2021

The Mu3e experiment: Toward the construction of an HV-MAPS vertex detector

The Mu3e experiment searches for the lepton flavor violating decay $\mu^+~\rightarrow~e^+~e^+~e^-$ with an ultimate aimed sensitivity of 1 event in $10^{16}$ decays. This goal can only be achieved by reducing the material budget per tracking layer to $X/X_0 \approx 0.1 \%$. High-Voltage Monolithic Active Pixel Sensors (HV-MAPS) which are thinned to $50\ \mu m$ serve as sensors. Gaseous helium is chosen as coolant. Results of recent studies related to the sensor prototypes, the helium cooling, and module prototyping are presented. The recent chip submission MuPix10 has proven its functionality regarding efficiency and time resolution. The helium cooling system for the inner tracker could be verified using a full-scale prototype. A complete prototype equipped with MuPix10 chips will be tested inside the Mu3e magnet in summer 2021.

physics.ins-det↗

arXiv2020

MuPix10: First Results from the Final Design

Many years of research and development of High Voltage Monolithic Active Pixel Sensors (HVMAPS) have culminated in the final design for the Mu3e pixel sensor. MuPix10 is a fully monolithic sensor with an active pixel matrix size of $20\times20\,\mathrm{mm}^2$ produced in the $180\,\mathrm{nm}$ HV-CMOS process at TSI Semiconductors. The pixel size is $80\times80\,\mathrm{\mu m}^2$. Hits are read out using a column-drain architecture and sent over up to four serial links with up to $1.6\,\left.\mathrm{Gbit}\middle/\mathrm{s}\right.$ each. By means of DC/DC converters and exclusive usage of on-chip biasing, MuPix10 is fully operable with a minimal set of electrical connections. This is an integral requirement by the Mu3e experiment since it enables the construction of ultra-thin pixel modules with $0.1\,$% of a radiation length per layer. First results from lab characterisation and testbeam campaigns are presented.

physics.ins-det↗

arXiv2020

The Mu3e Data Acquisition

The Mu3e experiment aims to find or exclude the lepton flavour violating decay $\mu^+\to e^+e^-e^+$ with a sensitivity of one in 10$^{16}$ muon decays. The first phase of the experiment is currently under construction at the Paul Scherrer Institute (PSI, Switzerland), where beams with up to 10$^8$ muons per second are available. The detector will consist of an ultra-thin pixel tracker made from High-Voltage Monolithic Active Pixel Sensors (HV-MAPS), complemented by scintillating tiles and fibres for precise timing measurements. The experiment produces about 100 Gbit/s of zero-suppressed data which are transported to a filter farm using a network of FPGAs and fast optical links. On the filter farm, tracks and three-particle vertices are reconstructed using highly parallel algorithms running on graphics processing units, leading to a reduction of the data to 100 Mbyte/s for mass storage and offline analysis. The paper introduces the system design and hardware implementation of the Mu3e data acquisition and filter farm.

physics.ins-det↗

arXiv2018

Effects of Proton Irradiation on 60 GHz CMOS Transceiver Chip for Multi-Gbps Communication in High-Energy Physics Experiments

This paper presents the experimental results of $17~MeV$ proton irradiation on a $60~GHz$ low power, half-duplex transceiver (TRX) chip implemented in $65~nm$ CMOS technology. It supports short range point-to-point data rate up to $6~Gbps$ by employing on-off keying (OOK). To investigate the irradiation hardness for high energy physics applications, two TRX chips were irradiated with total ionizing doses (TID) of $74~kGy$ and $42~kGy$ and fluence of $1.4~\times$10$^{14}~ N_{eq}/cm^2$ and $0.8~\times$10$^{14}~N_{eq}/cm^2$ for RX and TX modes, respectively. The chips were characterized by pre- and post-irradiation analogue voltage measurements on different circuit blocks as well as through the analysis of wireless transmission parameters like bit error rate (BER), eye diagram, jitter etc. Post-irradiation measurements have shown certain reduction in performance but both TRX chips have been found operational through over the air measurements at $5~Gbps$. Moreover, very small shift in the carrier frequency was observed after the irradiation.

physics.ins-det↗

arXiv2018

Efficiency and timing performance of the MuPix7 high-voltage monolithic active pixel sensor

The MuPix7 is a prototype high voltage monolithic active pixel sensor with 103 times 80 um2 pixels thinned to 64 um and incorporating the complete read-out circuitry including a 1.25 Gbit/s differential data link. Using data taken at the DESY electron test beam, we demonstrate an efficiency of 99.3% and a time resolution of 14 ns. The efficiency and time resolution are studied with sub-pixel resolution and reproduced in simulations.

physics.ins-det↗

arXiv2016

The MuPix System-on-Chip for the Mu3e Experiment

Mu3e is a novel experiment searching for charged lepton flavor violation in the rare decay $μ^+ \rightarrow e^+e^-e^+$. Decay vertex position, decay time and particle momenta have to be precisely measured in order to reject both accidental and physics background. A silicon pixel tracker based on $50\,μ$m thin high voltage monolithic active pixel sensors (HV-MAPS) in a 1 T solenoidal magnetic field provides precise vertex and momentum information. The MuPix chip combines pixel sensor cells with integrated analog electronics and a periphery with a complete digital readout. The MuPix7 is the first HV-MAPS prototype implementing all functionalities of the final sensor including a readout state machine and high speed serialization with 1.25 Gbit/s data output, allowing for a streaming readout in parallel to the data taking. The observed efficiency of the MuPix7 chip including the full readout system is $\geq99\%$ in a high rate test beam.

physics.ins-det↗

arXiv2016

Feasibility studies for a wireless 60 GHz tracking detector readout

The amount of data produced by highly granular silicon tracking detectors in high energy physics experiments poses a major challenge to readout systems. At high collision rates, e.g. at LHC experiments, only a small fraction of data can be read out with currently used technologies. To cope with the requirements of future or upgraded experiments new data transfer techniques are required which offer high data rates at low power and low material budget. Wireless technologies operating in the 60 GHz band or at higher frequencies offer high data rates and are thus a promising upcoming alternative to conventional data transmission via electrical cables or optical fibers. Using wireless technology, the amount of cables and connectors in detectors can be significantly reduced. Tracking detectors profit most from a reduced material budget as fewer secondary particle interactions (multiple Coulomb scattering, energy loss, etc.) improve the tracking performance in general. We present feasibility studies regarding the integration of the wireless technology at 60 GHz into a silicon tracking detector. Spare silicon strip modules of the ATLAS experiment are measured to be opaque in the 60 GHz range. The reduction of cross talk between links is studied. An estimate of the maximum achievable link density is given. It is shown that wireless links can be placed as close as 2 cm next to each other for a layer distance of 10 cm by exploiting one or several of the following measures: highly directive antennas, absorbers, linear polarization and frequency channeling. Combining these measures, a data rate area density of up to 11 Tb/(s $\cdot$ m$^2$) seems feasible. In addition, two types of silicon sensors are tested under mm-wave irradiation . No deterioration of the performance of both prototypes is observed.

physics.ins-det↗