In recent years, a new tomographic inversion method based on the Maximum Likelihood (ML) approach has been adapted to JET bolometry. Apart from its accuracy and reliability, the key advantage is its ability to provide reliable estimates of the uncertainties in the reconstructions. The original algorithm was implemented and validated using the MATLAB software tool. This work presents the accelerated version of the algorithm implemented using a compatible ITER fast controller platform with the Ubuntu 18.04 or the ITER Codac Core System distributions (6.1.2). The algorithm has been implemented in C++ using the open-source libraries: ArrayFire, ALGLIB, and MATIO. These libraries simplify the management of specific hardware accelerators such as GPUs and increase performance. The speed-up factor obtained is approximately 10 times. The work presents the methodology followed, the results obtained, and the advantages and drawbacks of implementation.
7 June 2021 in IEEE Transactions on Nuclear Science
Image-based diagnostics are key for fusion experiments. The operating conditions at ITER and the future machines require changing the role of such systems from monitoring and archiving for offline postprocessing to real-time processing. One of the roles of such systems is machine protection. A relevant application of vision diagnostics is the wall and divertor temperature monitoring and hot spot detection. However, algorithms for hot spot detection are computationally costly. To achieve real-time performance at the required time resolution for all these experiments, evaluating and validating the newest technologies is vital. This work applies heterogeneous computing techniques based on the OpenCL standard to the real-time hot spot detection problem and obtains the performance values in a Micro Telecommunications Computer Architecture (MTCA) platform. OpenCL reduces the development time, improves portability, and simplifies the evaluation and validation of each part of the algorithm to find the best-suited device in the heterogeneous system. The proposed solution enables balancing the computational load between a field-programmable gate array (FPGA) and a graphical processing unit (GPU). The algorithm has been adapted and optimized, taking profit on the particularities of each platform.
10 March 2023 in IEEE Transactions on Nuclear Science
The Micro Telecommunications Computing Architecture (MTCA) standard is widely used in developing advanced data acquisition and processing solutions in the big physics community. The number of applications implemented using commercial Advanced Mezzanine Cards (AMC) using AMD-Xilinx and Intel® Field Programmable Gate Array (FPGA) systems on chips is growing due to the flexibility and scalability of these reconfigurable hardware devices and their suitability to implement intelligent applications using artificial intelligence and machine learning techniques. The paper presents the specific design methodologies for hardware acceleration proposed by both FPGA manufacturers. Comparative results are obtained from two different software/hardware setups using two different AMCs, one based on Intel® FPGA Arria 10 and another based on Xilinx ZynqMP. The paper illustrates the process of how to modify the board support package, required by the hardware acceleration methodology, to implement the JESD204B and Low-Voltage Differential Signaling (LVDS) interfaces with the FPGA Mezzanine Card (FMC) modules containing the ADCs, to prepare the AMC cards to implement such kind of applications. The data acquisition and processing implementation inside these reference designs, with both languages OpenCL and High-Level Synthesis (HLS), is described. An important feature, needed for many applications in the big physics field, is the interface with the Experimental Physics and Industrial Control System (EPICS) software framework using the ITER Nominal Device Support framework, which is briefly described.
28 January 2025 in IEEE Transactions on Nuclear Science
The ITER International Fusion Experiment Organization is implementing the Real-Time Framework (RTF) to facilitate the development, deployment, and execution of instrumentation and control (I&C) applications optimized for real-time performance using the GNU/Linux-based ITER CODAC Core System software distribution. This contribution examines the feasibility of using hardware acceleration techniques with Field Programmable Gate Arrays (FPGAs) to implement real-time applications in the RTF that require specific compute-intensive functions. By combining the use of languages such as High-Level Synthesis (HLS) and Open Computing Language (OpenCL) with FPGA devices, specific hardware architectures can be implemented to solve certain computational problems to gain performance and limit latency. This work shows the methodology used to integrate HLS and OpenCL in the ITER CODAC Core System and the results obtained in terms of execution time for two common processing operations, vector addition, and matrix multiplication, using a commercial off-the-shelf FPGA-based device.