Type Systems for Multithreaded Software

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=bQ-NJ5IG1E4


Duration: 1:09:22
35 views
1

Developing correct multithreaded software is very challenging, due to the potential for unintended interference between threads. We present type systems for verifying two key non-interference properties in multithreaded software: race-freedom and atomicity. Verifying atomicity is particularly valuable since atomic procedures can be understood according to their sequential semantics, which significantly simplifies subsequent (formal and informal) correctness arguments. We will describe our experience applying these type systems and corresponding type inference algorithms to standard multithreaded benchmarks and other applications, and illustrate some defects revealed by this approach.



Other Videos By Microsoft Research


2016-09-06Natural Scene Categorization in Humans and Computers
2016-09-06Pair Programming Re-Design
2016-09-06An Interface to Support Multi-faceted Information Seeking and Targeted Relevance Feedback
2016-09-06Dependable Messaging in Sensor Networks
2016-09-06Discriminative Graphical Models for Structured Data Prediction
2016-09-06Dealing with data: classification, clustering and ranking
2016-09-06Toward the next generation interaction space: Tangible Space Initiative
2016-09-06G2 Microsystems Presents to Microsoft
2016-09-06Optimal Dynamic Auctions
2016-09-06Systems Challenges in Tiered Sensor Networks
2016-09-06Type Systems for Multithreaded Software
2016-09-06A Multi-Corpus Evaluation of Dynamic Markov Coding for Spam Filtering [1/3]
2016-09-06Brazil: After Neo-liberalism
2016-09-06Mobility for Sensing Uncertainty Reduction in Sensor Networks
2016-09-06One Dimensional DLA
2016-09-06Fitting a C^m Smooth function to data
2016-09-06Email activity management: A machine learning approach
2016-09-06Satisfaction: How Every Great Company Listens to the Voice of the Customer
2016-09-06Applying Data Mining Techniques to Computer Systems
2016-09-06Local chromatic number of quadrangulation of surfaces
2016-09-06Location-Based Activity Recognition


Tags:
microsoft research