Integrating Profiler Data
| Thesis title in Czech: | Integrování profilovacích dat |
|---|---|
| Thesis title in English: | Integrating Profiler Data |
| Academic year of topic announcement: | 2007/2008 |
| Thesis type: | diploma thesis |
| Thesis language: | angličtina |
| Department: | Department of Software Engineering (32-KSI) |
| Supervisor: | prof. Ing. Petr Tůma, Dr. |
| Author: | hidden - assigned and confirmed by the Study Dept. |
| Date of registration: | 17.10.2007 |
| Date of assignment: | 17.10.2007 |
| Date and time of defence: | 26.05.2008 00:00 |
| Date of electronic submission: | 26.05.2008 |
| Date of proceeded defence: | 26.05.2008 |
| Opponents: | RNDr. Leo Galamboš, Ph.D. |
| Guidelines |
| The existing support for performance profiling is fragmented among compiler tools, virtual machine tools, operating systems and middleware. The tools exhibit various degree of overhead, configurability and other properties, which makes them difficult to use in large projects. The goal of this assignment is to devise a framework for performance profiling that would allow to adaptively integrate information from various sources, maintaining minimum overhead and maximum configurability. Specifically:
1. Study existing profilers, performance data formats [2-7], proposed infrastructure for performance data collection [1]. 2. Design the data model and the data format for performance data that will allow integration of information from various sources (languages, virtual machines, operating systems, hosts), compatible with [1] and able to handle large and sparse datasets. 3. Design an interface that can make common profilers accessible to the infrastructure [1]. 4. Deliver a working implementation on Linux supporting native code and at least one virtual machine, well documented and tested. A pragmatic compromise between accessing performance data online using the interface and offline using the data model and data format will be made to overcome technical issues. |
| References |
| [1] Bulej L.: Connector-based Performance Data Collection for Component Applications, Ph.D. Thesis, Charles University, 2007.
[2] gprof, http://www.gnu.org [3] oprofile, http://oprofile.sf.net [4] dtrace, http://www.sun.com/bigadmin/content/dtrace [5] hprof, http://java.sun.com/developer/technicalArticles/Programming/HPROF.html [6] JVMTI, http://java.sun.com/j2se/1.5.0/docs/guide/jvmti/index.html [7] NetCDF, http://www.unidata.ucar.edu/software/netcdf |