NUMA Time Warp

Abstract

It is well known that Time Warp may suffer from large usage of memory, which may hamper the efficiency of the memory hierarchy. To cope with this issue, several approaches have been devised, mostly based on the reduction of the amount of used virtual memory, e.g., by the avoidance of checkpointing and the exploitation of reverse computing. In this article we present an orthogonal solution aimed at optimizing the latency for memory access operations when running Time Warp systems on Non-Uniform Memory Access (NUMA) multi-processor/multi-core computing systems. More in detail, we provide an innovative Linux-based architecture allowing per simulation-object management of memory segments made up by disjoint sets of pages, and supporting both static and dynamic binding of the memory pages reserved for an individual object to the different NUMA nodes, depending on what worker thread is in charge of running that simulation object along a given wall-clock-time window. Our proposal not only manages the virtual pages used for the live state image of the simulation object, rather, it also copes with memory pages destined to keep the simulation object's event buffers and any recoverability data. Further, the architecture allows memory access optimization for data (messages) exchanged across the different simulation objects running on the NUMA machine. Our proposal is fully transparent to the application code, thus operating in a seamless manner. Also, a free software release of our NUMA memory manager for Time Warp has been made available within the open source ROOT-Sim simulation platform. Experimental data for an assessment of our innovative proposal are also provided in this article

Closing the Gap Between Lookahead and Checkpointing to Provide Hybrid Synchronization

“…We intend to evaluate the system with more realistic scenarios in the future. Finally, we see as an exciting future work to allow dynamic changes which affect the end-to-end behavior [Pellegrini and Quaglia 2015]. The deployment of part of the synchronization computation into the network infrastructure [Bosshart et al 2014] and Integrating specification languages such as DEVS [Syriani et al 2011] to define the behavior of LPs are also in perspective.…”

OS-Based NUMA Optimization: Tackling the Case of Truly Multi-thread Applications with Non-partitioned Virtual Page Accesses

“…Beyond kernel-level software, NUMA-aware allocators have also been considered in HPC, e.g., in parallel scientific applications based on explicit data partitioning [17], [23]. In such a scenario, the application-level allocator resorts to NUMA-specific system calls to support the (dynamic) binding of logical pages to the NUMA nodes where the threads touching a specific partition of the overall application state are running (or are dynamically migrated to).…”

“…This leads to the scenario where the virtual pages used for the objects' representation and for their recoverability data are accessed in shared mode by the threads (depending on runtime decisions of the load-sharing policy and on how it clusters the objects across the worker threads). (D) It already embeds a NUMA oriented memory management support (see [23]) that operates in user-space, which can be exploited for a comparative analysis of our current proposal.…”