Quantifying the cost of context switch

Li, Chuanpeng; Ding, Chen; Shen, Kai

doi:10.1145/1281700.1281702

Cited by 143 publications

(65 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A context switch can have a significant cost, although the cost may depend vastly on a variety of factors such as the processor type, workload, and memory access patterns of the applications between which the context switch is performed. As discussed in [20,5], the overhead of a context switch comes from several aspectsthe processor registers need to be saved and restored, cache and translation lookaside buffer (TLB) entries for the evicted process need to be flushed and then reloaded for the incoming process, and the processor pipeline must be flushed.…”

Section: Performance Overhead Of Fusementioning

confidence: 99%

Performance and extension of user space file systems

Rajgarhia

Gehani

2010

Proceedings of the 2010 ACM Symposium on Applied Computing

View full text Add to dashboard Cite

Several efforts have been made over the years for developing file systems in user space. Many of these efforts have failed to make a significant impact as measured by their use in production systems. Recently, however, user space file systems have seen a strong resurgence. FUSE is a popular framework that allows file systems to be developed in user space while offering ease of use and flexibility.In this paper, we discuss the evolution of user space file systems with an emphasis on FUSE, and measure its performance using a variety of test cases. We also discuss the feasibility of developing file systems in high-level programming languages, by using as an example Java bindings for FUSE that we have developed. Our benchmarks show that FUSE offers adequate performance for several kinds of workloads.

show abstract

Section: Performance Overhead Of Fusementioning

confidence: 99%

Performance and extension of user space file systems

Rajgarhia

Gehani

2010

Proceedings of the 2010 ACM Symposium on Applied Computing

View full text Add to dashboard Cite

show abstract

“…The cost of a context switch can vary significantly [35,23]. On our platform, a simple benchmark where two threads take turns blocking each other on a synchronization variable estimates the context switch duration to be 3.3 µs.…”

Section: Thread Scheduling Overheadmentioning

confidence: 99%

Constructing performance model of JMS middleware platform

Martinec

Marek

Steinhauser

et al. 2014

Proceedings of the 5th ACM/SPEC International Conference on Performance Engineering

View full text Add to dashboard Cite

Middleware performance models are useful building blocks in the performance models of distributed software applications. We focus on performance models of messaging middleware implementing the Java Message Service standard, showing how certain system design properties -including pipelined processing and message coalescing -interact to create performance behavior that the existing models do not capture accurately. We construct a performance model of the ActiveMQ messaging middleware that addresses the outlined issues and discuss how the approach extends to other middleware implementations.

show abstract

“…As the algorithm is intended for real-time simulation, at 30 frames (steps) per second it must perform 30 × 160 = 4800 convolutions per second, or one every 200 microseconds. This interval is of the same order of magnitude as the context switch latency on a typical desktop operating system [7]. When benchmarking the fluid solver using ThreadScope [4], we noticed that for a grid of size 100 × 100 it was spending over half its time stalled while waiting for worker threads to be scheduled.…”

Section: Scheduling and Smallness Hintsmentioning

confidence: 99%

Guiding parallel array fusion with indexed types

et al. 2012

View full text Add to dashboard Cite

We present a refined approach to parallel array fusion that uses indexed types to specify the internal representation of each array. Our approach aids the client programmer in reasoning about the performance of their program in terms of the source code. It also makes the intermediate code easier to transform at compile-time, resulting in faster compilation and more reliable runtimes. We demonstrate how our new approach improves both the clarity and performance of several end-user written programs, including a fluid flow solver and an interpolator for volumetric data.

show abstract

Quantifying the cost of context switch

Cited by 143 publications

References 4 publications

Performance and extension of user space file systems

Performance and extension of user space file systems

Constructing performance model of JMS middleware platform

Guiding parallel array fusion with indexed types

Contact Info

Product

Resources

About