Latencies in Linux and FreeBSD kernels with different schedulers – O(1), CFS, 4BSD, ULE
نویسندگان
چکیده
This paper is a study of scheduler latencies in different versions of Linux 2.6 kernel with emphasis set on its usage in real-time systems. It tries to find the optimal kernel configuration leading to minimal latencies using some soft real-time tuning options. Benchmark tests under heavy load show differences between kernels and also between different scheduling policies. We compare Linux kernel 2.6.22 with a long-acting O(1) scheduler with the to date latest Linux kernel 2.6.28 that already uses a completely new CFS scheduler. Not only a scheduler, but also other kernel options can lead to latency reducing. We compare a kernel compiled for server employment and then tuned to act as a soft real time operating system. For better comparison we perform selected benchmarks also on FreeBSD 7.1 kernel compiled with the older 4BSD scheduler and with the newer ULE scheduler. ULE scheduler was improved in the version 7 of FreeBSD, so we compare it also with ULE in FreeBSD 6.3. The emphasis of this paper is set on finding a scheduler with minimal latencies on tested hardware.
منابع مشابه
On Intelligent Mitigation of Process Starvation In Multilevel Feedback Queue Scheduling
CPU time-share process schedulers for computer operating systems have existed since Corbato published his paper on the Compatible Time Sharing System in 1962 [8]. With this new type of scheduler came the need to effectively divide CPU time between N processes, where N could be 2 or more processes. Modern time-sharing process schedulers which have been developed in the decades since have been de...
متن کاملULE: A Modern Scheduler for FreeBSD
The existing thread scheduler in FreeBSD was well suited towards the computing environment that it was developed in. As the priorities and hardware targets of the project have changed, new features and scheduling properties were required. This paper presents ULE, a scheduler that is designed with modern hardware and requirements in mind. Prior to discussing ULE, the designs of several other sch...
متن کاملMitigating starvation of Linux CPU-bound processes in the presence of network I/O
In prior research work, it has been demonstrated that Linux can starve CPU-bound processes in the presence of network I/O. The starvation of Linux CPU-bound processes occurs under the two Linux schedulers, namely the 2.6 O(1) scheduler and the more recent 2.6 Completely Fair Scheduler (CFS). In this paper, we analyze the underlying root causes of this starvation problem and we propose effective...
متن کاملCurriculum Vitæ
My research and development activities are reported in the following sections (more details can be found in my homepage). The practical outcomes of these activities can be summarized as follows. • I am the author, developer and copyright holder of Quick Fair Queueu-ing Plus (QFQ+), one of the most accurate and fast packet schedulers available in Linux. • I am one of the authors, developers and ...
متن کاملA Cooperative File System
The Cooperative File System (CFS) is a new peer-to-peer read-only storage system that provides provable guarantees for the efficiency, robustness, and load-balance of file storage and retrieval. CFS does this with a completely decentralized architecture that can scale to large systems. CFS servers provide a distributed hash table (DHash) for block storage. CFS clients interpret DHash blocks as ...
متن کامل