Abstract. Performance profiling of MPI programs generates overhead during execution that introduces error in profile measurements. It is possible to track and remove overhead online, but it is necessary to communicate execution delay between processes to correctly adjust their interdependent timing. We demonstrate the first implementation of a onlne measurement overhead compensation system for profiling MPI programs. This is implemented in the TAU performance systems. It requires novel techniques for delay communication in the use of MPI. The ability to reduce measurement error is demonstrated for problematic test cases and real applications
We have developed an environment, based upon robust, existing, open source software, for tuning appl...
Abstract. Tracing parallel programs to observe their performance introduces in-trusion as the result...
This article presents a class library for detecting typical performance problems in event traces of ...
Abstract. Performance profiling generates measurement overhead during parallel program execution. Me...
In this paper we describe the difficulties inherent in making accurate, reproducible measurements of...
Abstract. In this paper we describe the difficulties inherent in making accurate, reproducible measu...
We have developed a new MPI benchmark package called MPIBench that uses a very precise and portable ...
Abstract. Workload characterization is an important technique that helps us understand the performan...
The desire for high performance on scalable parallel systems is increasing the complexity and the...
This paper presents an automatic counter instrumentation and pro ling module added to the MPI librar...
An MPI profiling library is a standard mechanism for intercepting MPI calls by applications. Profili...
Event tracing of parallel programs can provide valuable information about program performance. The d...
An MPI profiling library is a standard mechanism for intercepting MPI calls by applications. Profili...
The amount of parallelism in modern supercomputers currently grows from generation to generation. Fu...
Applications on todays massively parallel supercom-puters rely on performance analysis tools to guid...
We have developed an environment, based upon robust, existing, open source software, for tuning appl...
Abstract. Tracing parallel programs to observe their performance introduces in-trusion as the result...
This article presents a class library for detecting typical performance problems in event traces of ...
Abstract. Performance profiling generates measurement overhead during parallel program execution. Me...
In this paper we describe the difficulties inherent in making accurate, reproducible measurements of...
Abstract. In this paper we describe the difficulties inherent in making accurate, reproducible measu...
We have developed a new MPI benchmark package called MPIBench that uses a very precise and portable ...
Abstract. Workload characterization is an important technique that helps us understand the performan...
The desire for high performance on scalable parallel systems is increasing the complexity and the...
This paper presents an automatic counter instrumentation and pro ling module added to the MPI librar...
An MPI profiling library is a standard mechanism for intercepting MPI calls by applications. Profili...
Event tracing of parallel programs can provide valuable information about program performance. The d...
An MPI profiling library is a standard mechanism for intercepting MPI calls by applications. Profili...
The amount of parallelism in modern supercomputers currently grows from generation to generation. Fu...
Applications on todays massively parallel supercom-puters rely on performance analysis tools to guid...
We have developed an environment, based upon robust, existing, open source software, for tuning appl...
Abstract. Tracing parallel programs to observe their performance introduces in-trusion as the result...
This article presents a class library for detecting typical performance problems in event traces of ...