Discrete event simulation is the most suitable type of simulation for analyzing a complex system where changes happen at discrete time instants. Discrete event simulation is a major experimental methodology in several scientific and engineering domains. Unfortunately, a conventional discrete event simulator cannot meet with increasing demands of computational or the structural complexities of modern systems such as peer-to-peer (P2P) systems; therefore parallel discrete event simulation has been a focus of researchers for several decades. Unfortunately, no simulator is regarded as a standard which can satisfy the demands of all kinds of applications. Thus while given a simulator yields good performance for a specific kind of applications, it may failed to be efficient for other kinds of applications. Furthermore, although technological advancements have been made in the multi-core computing hardware, none of the mainstream P2P discrete event simulators is designed to support parallel simulation that exploits multi-core architectures. The proposed HydraNetSim parallel discrete event simulator (PDES) is a step toward addressing these issues. Developing a simulator which can support very large numbers of nodes to realize a massive P2P system, and can also execute in parallel is a non-trivial task. The literature review in this thesis gives a broad overview of prevailing approaches to dealing with the tricky problems of simulating a massive, large, and rapidly changing system, and provides a foundation for adopting a suitable architecture for developing a PDES. HydraNetSim is a discrete event simulator which allows parallel simulation and exploits the capabilities of parallelization of modern computing hardware. It is based on a novel master/slave paradigm. It divides the simulation model into a number of specific slaves (a cluster of processes) considering the number of cores provided by the underlying computing hardware. Each slave can be assigned to a specific CPU on a different core. Synchronization of the slaves is achieved by proposing a variant of the classic Null-Message Algorithm (NMA) with a focus on keeping the synchronization overhead as low as possible. Furthermore, HydraNetSim provides log information for debugging purposes and introduces a new mechanism of gathering and writing simulation results to a database. The experimental results show that the sequential counterpart of HydraNetSim (SDES) takes 41.6% more time than HydraNetSim-2Slave and 23.6% than HydraNetSim-3Slave. HydraNetSim-2Slave is 1.42 times faster, consumes 1.18 times more memory, and supports 2.02 times more nodes than a sequential discrete event simulator (SDES). Whereas, HydraNetSim-3Slave executes 1.24 times faster, consumes 2.08 times more memory, and supports 3.04 times more nodes than SDES. The scaling factor of HydraNetSim is ⌈(β-1)*102.04%⌉ of the maximum numbered of nodes supported by SDES; where β is the number of slaves. / Diskret händelsesimulering är den mest passande typen av simulering för att analysera ett komplext system där förändringar sker i diskreta tidpunkter. Diskret händelsesimulering är en stor experimentell metod i flera vetenskapliga och tekniska områden. Tyvärr kan en konventionell diskret händelse simulator uppfyller inte med ökande krav på beräkningsprogram eller strukturella komplexiteten av moderna system som peer-to-peer (P2P) system, och därför parallellt diskret händelse simulering har varit ett fokus för forskare under flera årtionde. Tyvärr ingen simulator ansåg som en standard som kan uppfylla kraven på alla typer av applikationer. Så samtidigt få en simulator ger bra prestanda för en specifik typ av applikationer kan det inte vara effektivt för andra typer av applikationer. Även om tekniska framsteget har gjorts i multi-core datorhårdvara, är ingen av de vanliga P2P händelsestyrd simulatorer för att stödja parallella simulering som utnyttjar flera kärnor arkitekturer. Den föreslagna HydraNetSim parallella diskret händelse simulator (PDES) är ett steg mot att fokusera på dessa frågor. Utveckla en simulator som kan stödja ett mycket stort antal noder för att realisera en massiv P2P-system, och kan även utföra parallellt är en icke-trivial uppdrag. Litteraturstudien i denna tesen ger en bred översikt över aktuell metoder för att hantera de svåra problem som simulerar en massiv, stor och snabbt ändra system och ger en grund för att adoptera en passande struktur för att utveckla ett PDES. HydraNetSim är en diskret händelse simulator som gör det möjligt parallellt simulering och utnyttjar funktionerna i parallellisering av modern datorhårdvara. Det är baserat på en ny master / slav paradigm. Den delar simuleringsmodellen i ett antal specifika slavar (ett kluster av processer) med tanke på antalet kärnor som tillhandahålls av den underliggande datorhårdvara. Varje slav kan tilldelas en specifik CPU på en annan kärna. Synkronisering av slavarna uppnås genom att föreslå en variant av det klassiska Null-Message Algorithm (NMA) med fokus på att hålla simuleringen overhead så lågt som möjligt. Dessutom ger HydraNetSim log information för felsökning ändamål och inför en ny mekanism för att samla in och skriva simuleringar resultat till en databas. De experimentella resultaten visar att den sekventiella motsvarigheten till HydraNetSim (SDES) tar 41,6% mer tid än HydraNetSim-2Slave och 23,6% mindre än HydraNetSim-3Slave. HydraNetSim-2Slave är 1,42 gånger snabbare, förbrukar 1,18 gånger mer minne, och stöder 2.02 gånger fler noder än en sekventiell händelsestyrd simulator (SDES). I HydraNetSim-3Slave kör 1.24 gånger snabbare, förbrukar 2,08 gånger mer minne, och stöder 3,04 gånger fler noder än SDES. Skalfaktorn av HydraNetSim är ⌈(β-1)*102.04%⌉ av den maximala numrerade noder som stöds av SDES; där β är antalet slavar.

Parallelizing the development cycles of hardware and software is becoming the industry’s norm for reducing time to market for electronic devices. In the absence of hardware, software development is based on a virtual platform; a fully functional software model of a system under development, able to execute unmodified code. A Transaction Level Model, expressed with the SystemC TLM 2.0 language, is one of the many possible ways for constructing a virtual platform. Under SystemC’s simulation engine, hardware and software is being co-simulated. However, the sequential nature of the reference implementation of the SystemC’s simulation kernel, is a limiting factor. Poor simulation performance often constrains the scope and depth of the design decisions that can be evaluated. It is the main objective of this thesis’ project to demonstrate the feasibility of parallelizing the co-simulation of hardware and software using Transaction Level Models, outside SystemC’s reference simulation environment. The major obstacle identified is the preservation of causal relations between simulation events. The solution is obtained by using the process synchronization mechanism known as the Chandy/Misra/Bryantt algorithm. To demonstrate our approach and evaluate under which conditions a speedup can be achieved, we use the model of a cache-coherent, symmetric multiprocessor executing a synthetic application. Two versions of the model are used for the comparison; the parallel version, based on the Message Passing Interface 3.0, which incorporates the synchronization algorithm and an equivalent sequential model based on SystemC TLM 2.0. Our results indicate that by adjusting the parameters of the synthetic application, a certain threshold is reached, above which a significant speedup against the sequential SystemC simulation is observed. Although performed manually, the transformation of a SystemC TLM 2.0 model into a parallel MPI application is deemed feasible.

The execution of parallel discrete event simulation across metacomputing infrastructures is examined. A master/worker architecture for parallel discrete event simulation is proposed providing robust executions under a dynamic set of services with system-level support for fault tolerance, semi-automated client-directed load balancing, portability across heterogeneous machines, and the ability to run codes on idle or time-sharing clients without significant interaction by users. Research questions and challenges associated with issues and limitations with the work distribution paradigm, targeted computational domain, performance metrics, and the intended class of applications to be used in this context are analyzed and discussed. A portable web services approach to master/worker parallel discrete event simulation is proposed and evaluated with subsequent optimizations to increase the efficiency of large-scale simulation execution through distributed master service design and intrinsic overhead reduction. New techniques for addressing challenges associated with optimistic parallel discrete event simulation across metacomputing such as rollbacks and message unsending with an inherently different computation paradigm utilizing master services and time windows are proposed and examined. Results indicate that a master/worker approach utilizing loosely coupled resources is a viable means for high throughput parallel discrete event simulation by enhancing existing computational capacity or providing alternate execution capability for less time-critical codes.