Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 8

Liczba wyników na stronie
first rewind previous Strona / 1 next fast forward last
Wyniki wyszukiwania
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 1 next fast forward last
EN
This paper describes a novel Polyhedral Source-to-Source Compiler (PSSC) that enables automatic recognition of parallel regions of C/C++ code and annotating them with OpenMP/OpenACC pragmas. The proposed source-to-source compiler uses polyhedral model to detect and optimize parallel loops. Loop optimization is done on intermediate code representation by Polly compiler and then it is mapped to original source code. This approach allows combining the simplicity and efficiency of Intermediate Representation (IR) code optimization with readability of output code. Experimental results show that the proposed compiler is able to reach the comparable performance to the original Polly compiler.
PL
Artykuł opisuje nowatorski kompilator typu source-to-source, który wykorzystuje model polihedralny do automatycznego wykrywania kodu C/C++, który może być wykonywany równolegle. Fragmenty kodu źródłowego, które mogą zostać zrównoleglone, są opatrywane pragmami OpenMP/OpenACC. Opisywany kompilator śledzi zmiany jakie zostały wprowadzone w kodzie pośrednim przez kompilator Polly, a następnie odwzoruje te transformacje w kodzie źródłowym. Przedstawione w artykule podejście umożliwia połączenie zalet wynikających z optymalizowania kodu pośredniego z możliwością łatwego przenoszenia na różne platformy kodu wysokopoziomowego. Przeprowadzone pomiary wydajności wykazały, że opracowany kompilator pozwala zrównoleglić kod wysokopoziomowy równie wydajnie jak bazowy kompilator Polly.
EN
Nowadays high energy physics scientists build and design systems that are complex in terms of the huge amount of subsystems and individual components. A single subsystem may consist of a few tens of thousands digital and analogue channels and sensors. As a result, the data rates captured in modern systems may result in gigabytes per second. Complex systems could generate various alarms and provide other diagnostic information. Consequently, a huge number of variables are needed to control and monitor the system. It could be a real challenge to provide access to all alarms and diagnostic information in systems composed of thousands of channels. In this sense, it is necessary to develop a methodology of designing Human Machine Interfaces (HMI) that will be simple to use and allow describing of relatively complex systems. This paper describes an HMI scheme able to obtain and present data from High Energy Physics systems. The purpose of this paper is to evaluate HMI panels dedicated for complex systems. The prototype HMI uses the demonstration PXIe-based Neutron Flux Monitor (NFM) developed by the Department of Microelectronics and Computer Science. This NFM is going to provide essential information for plasma operation in the ITER plant. The HMI involves a Graphical User Interface and an Alarm Management Scheme, all based on the Experimental Physics and Industrial Control System (EPICS) framework. The Graphical User Interface (GUI) includes the use of several tools provided by the Control System Studio as well as JavaScript, rules and actions to dynamically present data to the operators. In regards to alarm management, a scheme is proposed to efficiently handle alarms by presenting the relevant information and controls to quickly react to alarms.
PL
Obecnie systemy wykorzystywane w eksperymentach fizyki wielkich energii składają się ze znacznej liczby podsystemów i komponentów. Pojedynczy podsystem może obsługiwać nawet kilkadziesiąt tysięcy kanałów cyfrowych i analogowych oraz sensorów, w wyniku czego ilość danych zbieranych przez system jest liczona w gigabajtach na sekundę. Skomplikowane systemy mogą generować różne alarmy i dostarczać inne informacje diagnostyczne. W związku z tym, sterowanie i monitorowanie systemów wymaga wielu zmiennych. Umożliwienie dostępu do wszystkich alarmów i informacji diagnostycznych w systemach składających się z tysięcy kanałów stanowi poważne wyzwanie. Z tego względu niezbędne jest opracowanie metodologii projektowania paneli operatorskich HMI (Human Machine Interface) zbudowanych w prosty sposób, jednak pozwalających na obsługę stosunkowo złożonych systemów. Artykuł przedstawia system HMI dedykowany do zbierania i wizualizacji danych z systemów fizyki wielkich energii. Celem niniejszego artykułu jest ocena HMI dedykowanego dla skomplikowanych systemów. Opisywany prototypowy panel HMI wykorzystuje system monitorowania strumienia neutronów NFM (Neutron Flux Monitor), bazujący na standardzie PXIe, opracowany w Katedrze Mikroelektroniki i Technik Informatycznych. System NFM będzie dostarczał najważniejsze informacje dla użytkowania plazmy w ośrodku ITER. HMI składa się z graficznego interfejsu użytkownika GUI (Graphical User Interface) oraz Systemu Zarządzania Alarmami, zbudowanego w oparciu o platformę EPICS (Experimental Physics and Industrial Control System). Interfejs HMI korzysta z narzędzi dostarczanych przez CSS (Control System Studio) oraz języka JavaScript, reguł i akcji, aby na bieżąco prezentować dane operatorom. W kwestii obsługi alarmów, zaproponowano schemat prezentacji właściwych informacji i metod obsługi, pozwalający w wydajny sposób zarządzać alarmami.
EN
The data acquisition systems must be capable of process all the data produced by the source to ensure the highest level of accuracy, especially when it deals with hard real-time system monitoring task. However, the production of data is faster than the process to acquire and to process such a data. Using concurrency approach is an alternative to obtain the required level of performance and data processing. This paper presents the comparison between various C++ frameworks that by using multithreading technology and ringbuffer data structure allow data transfer in concurrent way. The comparison is based on the time interval between the instant when data is published and the instant when the data is gathered. These latency measurements have been taken using the configuration of one producer and two consumers for all evaluated frameworks. The results show that using standard C++ libraries to develop a simple framework it is possible to achieve suitable performance in order to fulfill the requirements of the high performance data acquisition systems described.
EN
As a result of a growing interest in the xTCA systems by research centers conducting High Energy Physics (HEP) experiments, the PICMG xTCA for Physics Coordinating Committee is developing a new, attractive standards. They are specifically dedicated to control and data acquisition systems of HEP applications. These new specifications define a number of extensions to the Advanced Telecommunications Computing Architecture (ATCA) and Micro Telecommunications Computing Architecture ( μ TCA) standards. The MicroTCA for Physics specification, named MTCA.4, defines some new solutions in order to simplify an application of μ TCA specification in HEP facilities. The most important of them is μ RTM module which can be connected to the Advanced Mezzanine Card (AMC) and provides more usable area and additional I/O connectors at the rear of the shelf. The MTCA.4 also introduces additional channels dedicated for timing and synchronization signals that are available on backplane. Due to the fact that xTCA for Physics is a new specification which has not been officially published, both the AMC and μ RTM modules and the software for the MMC with possibility of μ RTM handling are not yet available. For this reason the preparation of firmware for the Module Management Controller (MMC) for AMC modules compatible with MicroTCA for Physics specification is required. This paper presents a structure of the MMC for the μ TC (MicroTCA-based Controller) board. It also describes the microcontroller software which fulfills the role of the MMC.
EN
The High Energy Physics (HEP) experiments, due to their large scale, required performance and precision, have to be controlled by complex, distributed control systems. The systems are responsible for processing thousands of signals from various sensors of different types. Very often, one of the data sources applied in such systems are visible light/infrared cameras or other imaging sensors, which provide substantial information about studied phenomena. High data throughput for camera systems require dedicated mechanisms for data collecting and processing. Moreover, the images from cameras should be also available to system operator. It needs the support from both operator panels interface and control application which should provide data in the dedicated format. The paper presents two different approaches to image distribution, processing and visualisation applied in distributed control systems. Discussed is the issue of support for cameras and image data implemented in the Distributed Object Oriented Control System (DOOCS) and an example control system designed to the needs of image acquisition system on the base of the Experimental Physics and Industrial Control System (EPICS) environment.
6
EN
The paper describes an universal module for video stream acquisition from fast cameras with Camera Link interface. The first version of the referenced standard defines three configurations: Base, Medium and Full. The developed module supports all of them achieving transmission speeds up to 5.44 Gb/s for raw image data in the Full configuration. The module is designed according to FPGA Mezzanine Card (FMC) standard and can cooperate with carrier boards containing High-Pin Count (HPC) version of the connector. The module was tested with the TEWS TAMC-641 module.
EN
The Advanced Telecommunications Computing Architecture (ATCA) and Micro Telecommunications Computing Architecture (μTCA) standards, intended for high-performance applications, offer an array of features that are compelling from the industry use perspective, like high reliability (99,999%) or hot-swap support. The standards incorporate the Intelligent Platform Management Interface (IPMI) for the purpose o advanced diagnostics and operation control. This standard imposes support for non-volatile Field Replaceable Unit (FRU) information for specific components of an ATCA/μTCA-based system, which would typically include description of a given component. The Electronic Keying (EK) mechanism is capable of using this information for ensuring more reliable cooperation of the components. The FRU Information for the ATCA/μTCA implementation elements may be of sophisticated structure. This paper focuses on a software tool facilitating the process of assembling this information, the goal of which is to make it more effective and less error-prone.
EN
The Advanced Telecommunications Computing Architecture (ATCA) specification allows to meet the newest trends in high speed communication technologies. Furthermore, it provides manageability, availability and exceptional reliability at 99.999% level. Therefore, this architecture is perfect to use in control systems of complex projects like the Free-Electron Laser in Hamburg (FLASH) or the X-ray Free Electron Laser (X-FEL) that work with high-frequency signal processing, use high-speed communication protocols such as PCIe, Gigabit Ethernet or RocketIO and require continuous stable operation. Modular construction allows for flexible system configuration. What is worth emphasizing, in contrast to previous solutions like, VME (Versa Module Eurocard), it is possible to change the system configuration without the need for power shutdown. Hot-plug functionality is delivered by Intelligent Platform Management Interface (IPMI) system. For this reason, all ATCA units, which can be replaced in the field should have Intelligent Platform Management Controller (IPMC) that provides IPMI functionality. The IPMC is responsible for management and control of module on which it is placed. Due to the fact that the IPMC operates independently of another components, it allows to activation and deactivation other functional blocks of the module. For this purpose if has to communicate with the Shelf Manager according to IPMI standard. The IPMC monitors also vital operational parameters such as temperature, voltage, current and reports all abnormalities to the Shelf Manager. This paper presents the IPMC for ATCA Carrier Board with three AMC Bays. All IPMC functionalities required by ATCA specification are in this case fulfilled by microcontroller dedicated for IPMI management in xTCA systems.
first rewind previous Strona / 1 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.