User:Realazthat/Projects wishlist/LLVM/APInt for CBE: Difference between revisions
Content added Content deleted
No edit summary |
No edit summary |
||
Line 1: | Line 1: | ||
Implement arbitrary precision integers for the C backend by using a template library with fixed precision integers, compiling it to bitcode, using the generated function calls on C structs representing the integers, and running the optimization passes on them again. |
Implement arbitrary precision integers for the C backend by using a template library with fixed precision integers, compiling it to bitcode, using the generated function calls on C structs representing the integers, and running the optimization passes on them again. |
||
It would be implemented as a pass, and thus would benefit all backends incapable of representing AP. |
|||
==References== |
==References== |
||
* http://muntl.sourceforge.net/ |
* http://muntl.sourceforge.net/ |
Latest revision as of 23:12, 13 October 2010
Implement arbitrary precision integers for the C backend by using a template library with fixed precision integers, compiling it to bitcode, using the generated function calls on C structs representing the integers, and running the optimization passes on them again.
It would be implemented as a pass, and thus would benefit all backends incapable of representing AP.
References
- http://muntl.sourceforge.net/
- Template library for AP integers
- wp:Computational_complexity_of_mathematical_operations