This title appears in the Scientific Report :
2010
Please use the identifier:
http://hdl.handle.net/2128/4315 in citations.
A Generic Binary Instrumenter and Heuristics to Select Relevant Instrumentation Points
A Generic Binary Instrumenter and Heuristics to Select Relevant Instrumentation Points
Various tools targeting High Performance Computing applications, e.g., Scalasca, have been developed to aid software engineers in their analysis and subsequently in the improvement of their application’s performance. This thesis focuses on the very first step of application analysis, introducing a g...
Saved in:
Personal Name(s): | Mußler, Jan (Corresponding author) |
---|---|
Contributing Institute: |
Jülich Supercomputing Center; JSC |
Imprint: |
Jülich
Forschungszentrum Jülich GmbH Zentralbibliothek, Verlag
2010
|
Physical Description: |
XI, 91 p. |
Dissertation Note: |
Masterarbeit |
Document Type: |
Master Thesis |
Research Program: |
Computational Science and Mathematical Methods Scientific Computing |
Series Title: |
Berichte des Forschungszentrums Jülich
4335 |
Link: |
OpenAccess |
Publikationsportal JuSER |
Various tools targeting High Performance Computing applications, e.g., Scalasca,
have been developed to aid software engineers in their analysis and subsequently
in the improvement of their application’s performance.
This thesis focuses on the very first step of application analysis, introducing a
generic binary instrumenter, developed to support different measurement systems
and fulfill their demands imposed on the instrumentation. The definition of code
fragments uses a subset of C to expose Dyninst’s code generation and provide
compatibility with different tools. In order to improve the selection process, the
instrumenter also features rule-based filtering, permitting a targeted analysis and a reduction of instrumentation overhead that otherwise perturbs the measurement .... |