Performance Comparison of EPICS IOC and MARTe in a Hard Real-Time Control Application A.Barbalace 1, G. Manduchi 1, A. Neto 2, G. De Tommasi 3, D.F. Valcárcel 2, F. Sartori 4 1 Consorzio RFX, Associazione EURATOM-ENEA sulla Fusione, Corso Stati Uniti 4, I Padova, Italy 2 Associação EURATOM/IST, Instituto de Plasmas e Fusão Nuclear, Av. Rovisco Pais, Lisboa, Portugal 3 Associazione EURATOM-ENEA-CREATE, Università di Napoli Federico II, I Napoli, Italy 4 Fusion for Energy, Barcelona, Spain ph , Fax poster PFE - 13
input/output board NI6255 (PXI external rack) workstation HP Compaq dc 7900 CMT, x86 Intel Core 2 Duo GHz, 3 MB L2, 3 GB RAM Linux version RT patched (rt-24), 1kHz system clock What we have done Similarities and Differences Case Study EPICS IOCMARTe modular components based architecturesRecordsGAMs current configuration is defined in a text fileReCompileReLoading computational resource modelNo ControlFull Control
Reference Prog. EPICS IOC MARTe min19.0 us30.0 us20.9 us MAX28.3 us60.5 us31.3 us Avg22.3 us39.3 us25.0 us min, MAX and average overall IO latency 1kHz2kHz5kHz10kHz Reference Prog. 2%4%8%14% EPICS IOC4% 14%28% MARTe4%7%14%29% CPU load at different clock frequencies First Results Comparing with a Reference Program the added latency due to MARTe in respect of the reference program is on average 2.7 µs. the added latency due to EPICS is 17 µs.
Conclusions Adding network queries EPICS IOCMARTe Lesson Learned MARTe provides a shorter and, above all, more bounded latency; EPICS can be used as well for non highly demanding real-time applications. Linux with the Real-Time patches appears to be well suited to develop hard Real-Time applications making it a strong candidate for real-time control in large systems.