Search

This paper explores advanced techniques in highlevel synthesis (HLS) utilizing metamodel structures. Metamodels act as models of hardware models, generating internal hardware models based on parameter inputs and exploring the solution space to find optimal configurations. The focus is on enhancing HLS processes through metamodeling, enabling more efficient hardware design and optimization. Key contributions include a novel metamodel framework and a case study demonstrating its application in complex system designs. The proposed methods show significant improvements in synthesis efficiency and scalability, making them highly relevant for modern FPGA and ASIC design workflows.

The validation of the measurements quality after on-site diagnostic system installation is necessary in order to provide reliable data and output results. This topic is often neglected or not discussed in detail regarding measurement systems. In the paper recently installed system for soft X-ray measurements is described in introduction. The system is based on multichannel GEM detector and the data is collected and sent in special format to PC unit for further postprocessing. The unique feature of the system is the ability to compute final data based on raw data only. The raw data is selected upon algorithms by FPGA units. The FPGAs are connected to the analog frontend of the system and able to register all of the signals and collect the useful data. The interface used for data streaming is PCIe Gen2 x4 for each FPGA, therefore high throughput of the system is ensured. The paper then discusses the properties of the installation environment of the system and basic functionality mode. New features are described, both in theoretical and practical approach. New modes correspond to the data quality monitoring features implemented for the system, that provide extra information to the postprocessing stage and final algorithms. In the article is described also additional mode to perform hardware simulation of signals in a tokamak-like environment using FPGAs. The summary describes the implemented features of the data quality monitoring features and additional modes of the system.

The paper presents improvements of the developed system for hot plasma radiation measurement in the soft Xray range based on a Gas Electron Multiplier (GEM) detector. Scope of work consists of a new solution for handling hardware time-synchronization with tokamak systems needed for better synchronization with other diagnostics and measurement quality. The paper describes the support of new modes of triggering on PC-side. There are communication and data path overview in the system. The new API is described, which provide separate channels for data and control and is more robust than the earlier solution. Work concentrates on stability and usability improvements of the implemented device providing better usage for end-user.

A novel approach to a trigger mode in the Gas Electron Multiplier (GEM) detector readout system is presented. The system is already installed at WEST tokamak. The article briefly describes the architecture of the GEM detector and the measurement system. Currently the system can work in two trigger modes: Global Trigger and Local Trigger. All trigger processing blocks are parts of the Charge Signal Sequencer module which is responsible for transferring data to the PC. Therefore, the article presents structure of the Sequencer with details about basic blocks, theirs functionality and output data configuration. The Sequencer with the trigger algorithms is implemented in an FPGA chip from Xilinx. Global Trigger, which is a default mode for the system, is not efficient and has limitations due to storing much data without any information. Local trigger which is under tests, removes data redundancy and is constructed to send only valid data, but the rest of the software, especially on the PC side, is still under development. Therefore authors propose the trigger mode which combines functionality of two existing modes. The proposed trigger, called Zero Suppression Trigger, is compatible with the existing interfaces of the PC software, but is also capable to verify and filter incoming signals and transfer only recognized events. The results of the implementation and simulation are presented.