skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: A Framework to Develop Symbolic Performance Models of Parallel Applications

Abstract

Performance and workload modeling has numerous uses at every stage of the high-end computing lifecycle: design, integration, procurement, installation and tuning. Despite the tremendous usefulness of performance models, their construction remains largely a manual, complex, and time-consuming exercise. We propose a new approach to the model construction, called modeling assertions (MA), which borrows advantages from both the empirical and analytical modeling techniques. This strategy has many advantages over traditional methods: incremental construction of realistic performance models, straightforward model validation against empirical data, and intuitive error bounding on individual model terms. We demonstrate this new technique on the NAS parallel CG and SP benchmarks by constructing high fidelity models for the floating-point operation cost, memory requirements, and MPI message volume. These models are driven by a small number of key input parameters thereby allowing efficient design space exploration of future problem sizes and architectures.

Authors:
 [1];  [1]
  1. ORNL
Publication Date:
Research Org.:
Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Sponsoring Org.:
USDOE Office of Science (SC)
OSTI Identifier:
989553
DOE Contract Number:
DE-AC05-00OR22725
Resource Type:
Conference
Resource Relation:
Conference: 5th International Workshop on Performance Modeling, Evaluation, and Optimization of Parallel and Distributed Systems, Rhodes, Greece, 20060429, 20060429
Country of Publication:
United States
Language:
English
Subject:
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; BENCHMARKS; CONSTRUCTION; DESIGN; EVALUATION; EXPLORATION; OPTIMIZATION; PERFORMANCE; PROCUREMENT; SIMULATION; TUNING; VALIDATION; PARALLEL PROCESSING; COMPUTERS; COMPUTER CODES; PROGRAMMING

Citation Formats

Alam, Sadaf R, and Vetter, Jeffrey S. A Framework to Develop Symbolic Performance Models of Parallel Applications. United States: N. p., 2006. Web.
Alam, Sadaf R, & Vetter, Jeffrey S. A Framework to Develop Symbolic Performance Models of Parallel Applications. United States.
Alam, Sadaf R, and Vetter, Jeffrey S. Sun . "A Framework to Develop Symbolic Performance Models of Parallel Applications". United States. doi:.
@article{osti_989553,
title = {A Framework to Develop Symbolic Performance Models of Parallel Applications},
author = {Alam, Sadaf R and Vetter, Jeffrey S},
abstractNote = {Performance and workload modeling has numerous uses at every stage of the high-end computing lifecycle: design, integration, procurement, installation and tuning. Despite the tremendous usefulness of performance models, their construction remains largely a manual, complex, and time-consuming exercise. We propose a new approach to the model construction, called modeling assertions (MA), which borrows advantages from both the empirical and analytical modeling techniques. This strategy has many advantages over traditional methods: incremental construction of realistic performance models, straightforward model validation against empirical data, and intuitive error bounding on individual model terms. We demonstrate this new technique on the NAS parallel CG and SP benchmarks by constructing high fidelity models for the floating-point operation cost, memory requirements, and MPI message volume. These models are driven by a small number of key input parameters thereby allowing efficient design space exploration of future problem sizes and architectures.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Sun Jan 01 00:00:00 EST 2006},
month = {Sun Jan 01 00:00:00 EST 2006}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • The results of a study of a family of parallel symbolic architectures executing several parallel applications are presented. The class of architectures being simulated is characterized by a shared memory structure, by a hierarchical interconnect, and by clustered processors. Speedup measurements were obtained from six different application kernels. Measurements were also performed to assess the degradation of speedup as a function of the interconnection delays, and to study the effect of different scheduling algorithms. The results presented support the claim that the proposed architecture would be a powerful parallel symbolic computation system. The paper discusses processor starvation, fine grain parallelism,more » uneven loads, foreign reference, schedule and indeterminate computation with respect to the applications chosen.« less
  • Multi-resolution validation of hierarchical performance models of scientific applications is critical primarily for two reasons. First, the step-by-step validation determines the correctness of all essential components or phases in a science simulation. Second, a model that is validated at multiple resolution levels is the very first step to generate predictive performance models, for not only existing systems but also for emerging systems and future problem sizes. We present the design and validation of hierarchical performance models of two scientific benchmarks using a new technique called the modeling assertions (MA). Our MA prototype framework generates symbolic performance models that can bemore » evaluated efficiently by generating the equivalent model representations in Octave and MATLAB. The multi-resolution modeling and validation is conducted on two contemporary, massively-parallel systems, XT3 and Blue Gene/L system. The workload distribution and the growth rates predictions generated by the MA models are confirmed by the experimental data collected on the MPP platforms. In addition, the physical memory requirements that are generated by the MA models are verified by the runtime values on the Blue Gene/L system, which has 512 MBytes and 256 MBytes physical memory capacity in its two unique execution modes.« less
  • No abstract prepared.
  • No abstract prepared.
  • Sparse parallel factorization is among the most complicated and irregular algorithms to analyze and optimize. Performance depends both on system characteristics such as the floating point rate, the memory hierarchy, and the interconnect performance, as well as input matrix characteristics such as such as the number and location of nonzeros. We present LUsim, a simulation framework for modeling the performance of sparse LU factorization. Our framework uses micro-benchmarks to calibrate the parameters of machine characteristics and additional tools to facilitate real-time performance modeling. We are using LUsim to analyze an existing parallel sparse LU factorization code, and to explore amore » latency tolerant variant. We developed and validated a model of the factorization in SuperLU_DIST, then we modeled and implemented a new variant of slud, replacing a blocking collective communication phase with a non-blocking asynchronous point-to-point one. Our strategy realized a mean improvement of 11percent over a suite of test matrices.« less