This thesis concerns the issue of completness detection of instruction set description for the LLVM compiler, or the ability of a compiler to generate target program for every valid source program in the appropriate high-level programming language. On the basis of regular tree grammar theory and several scietific theses that also concern this issue, formal tool for inclusion checking of two grammars. Furthermore a method for automatic extraction of the two grammars from the instruction set description has been devised, as a result of which the tool can be used for checking completeness of instruction selection. In combination with checking completeness of the legalization process of the LLVM compiler, which preceeds the instruction selection, it should be feaseable to check completeness of most compiler parts dependent on the target architecture.
| Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:235470 |
| Date | January 2012 |
| Creators | Nagy, Michal |
| Contributors | Dolíhal, Luděk, Hruška, Tomáš |
| Publisher | Vysoké učení technické v Brně. Fakulta informačních technologií |
| Source Sets | Czech ETDs |
| Language | Czech |
| Detected Language | English |
| Type | info:eu-repo/semantics/masterThesis |
| Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.2043 seconds