Combining Static and Dynamic Data in Code Visualization

Combining Static and Dynamic Data in Code Visualization David Eng Sable Research Group, Mc. Gill University PASTE 2002 Charleston, South Carolina November 18 -19, 2002

Presentation outline Introduction n Framework overview n JIL n SOOT n STEP n JIMPLEX n Future work n

Introduction n Code visualization: why and where? n n Motivation: n n Program understanding. Compiler development. Visualization of an optimizable IR. Expose both static & dynamic program attributes which are hidden in the raw IR. Design goals: n Intuitive, portable, & modular.

Framework overview Java class files My. Class. java Tools with JIL support SOOT JIL documents static data (JIL) Visualizers with JIL support JIMPLEX (XSL/HTML) DTD JVM/JVMPI (execution profile) STEP dynamic data (JIL) My. Visualizer (my language) DTD (a) (b) (c) (d)

Java Intermediate Language (JIL) XML-based representation of code. n A metalanguage used to describe other languages and their characteristics. n Encapsulates IRs as code objects which can be extended with both static and dynamic data. n

JIL document structure myclass element legend base required methods fields history name, returns, access name, type, access version, created runtime invokes runtime accesses contributing tools extension parameters labels locals statements exceptions name, type line number, label type, handler uses statement count uses jimple declaration definitions flow information live variables in/out invoke targets

Benefits of using XML Extensible: add or remove elements without breaking existing structure. n Supported by existing parsers, APIs, software, databases, servers, etc. n Portable across platforms and networks. n Human readable, editable. n

JIL n Let’s take a look at an example JIL document…

Static vs. dynamic data n Static data: Collected at compile-time n Program attributes n Flow information n Analysis results n Basis for program understanding We chose SOOT as a source of static data… n n

SOOT n n n Java optimization framework and API. Uses intermediate languages (IRs) to perform analyses and transformations. Provides several output formats, including JIL. Java source (. java) SOOT Baf, Grimp, and Jimple Baf source (. baf) Grimp source (. grimp) javac Java class (. class) analysis & optimization Jimple source (. jimple) JIL document (. xml)

SOOT as a source of JIL SOOT was modified to generate Jimple as XML. n The resulting XML contains general program characteristics and flow information. n Static analysis results can be linked to applicable code elements. n

SOOT elements in JIL n As a source of static data: Flow information n Live variable analysis n Variable uses and definitions n Class hierarchy analysis n Variable type analysis n

Static vs. dynamic data n Dynamic data: Collected at runtime n Profiling results n Variable access and use counts n Temporal data n Basis for evaluation n Especially important for OOP languages n

STEP Toolkit for extensible profiling. n Encodes and compresses trace data in a trace definition language. n Java class (. class) STEP Profiling agent (JVMPI, Kaffe, SOOT) event pipe EVolve trace format JIL document (. xml)

STEP as a source of JIL STEP as a source of profiling data. n Visualizers becomes consumers at the backend of the event pipe. n STEP traces are read and converted to JIL. n JIMPLEX has full control over what elements are included and in what format. n

STEP elements in JIL n As a source of runtime dynamic data: n field accesses n method invocations n object allocations

Separation of data from data Static and dynamic data are generated as independent XML sources. n Trace results from different profiling runs can be swapped in and out. n Data sources can live on separate servers. n

Merging XML data sources To visualize both static and dynamic elements simultaneously, the data is merged into a single source. n XML can be merged on the fly using XSL. n

Separation of data and interface Interfaces are independent of data. n Can be implemented using any combination of languages or technologies. n One possible interface will be presented here… n

JIMPLEX n n n Web-based visualization interface implemented using HTML and XSL. Uses JIL documents as data sources, allowing them to exist remotely. Runs on any machine equipped with a web browser. Portable across platforms and networks. Modular and extensible.

JIMPLEX interface (a) Server A: JIL document storage myclass. xml (JIL document) web server XML data Server B: JIMPLEX server (b) jimplex. xsl (XSL stylesheet) XML/XSL transformation web server HTML data (c) Client 1: JIMPLEX web client JIMPLEX (web page) Client n: JIMPLEX … web client JIMPLEX (web page)

Example Interface

JIMPLEX interface n Let’s take a look at the interface…

Contributions n JIL document format n n SOOT extensions n n JIL document output format STEP extensions n n DTD specifications JIL document generating backend JIMPLEX visualizer n XSL stylesheets

Future Work: JIL n n n Extract more runtime attributes from STEP, such as temporal data. Use XML merging to support multiple STEP streams, allowing multi-dimensional profiling. Use XML differentiation to support program and trace comparison, allowing better evaluation of optimizations.

Future Work: UI n n n Any updates to JIL should be reflected in the UI (ie. multi-dimensional profiling). With the inclusion of more and more data, an application-based interface would provide better performance. Profiling and other dynamic data are more suited towards a graphical representation.

Links n Java Intermediate Language (JIL): n n n SOOT: n n n http: //www. sable. mcgill. ca/jil/ JIL papers, resources, and the JIMPLEX visualizer. http: //www. sable. mcgill. ca/soot/ SOOT papers, documentation, and downloads. STEP: n n http: //www. sable. mcgill. ca/step/ STEP papers, more to come.