|[2 earlier articles]|
|Re: Nondeterministic compilers? firstname.lastname@example.org (2004-02-27)|
|Re: Nondeterministic compilers? email@example.com (Franck Pissotte) (2004-02-27)|
|Re: Nondeterministic compilers? firstname.lastname@example.org (Alex Colvin) (2004-02-27)|
|Re: Nondeterministic compilers? email@example.com (Joachim Durchholz) (2004-03-02)|
|Re: Nondeterministic compilers? firstname.lastname@example.org (Walter) (2004-03-02)|
|Re: Nondeterministic compilers? email@example.com (2004-03-06)|
|Re: Nondeterministic compilers? firstname.lastname@example.org (norm rubin) (2004-03-11)|
|From:||"norm rubin" <email@example.com>|
|Date:||11 Mar 2004 12:50:27 -0500|
|Posted-Date:||11 Mar 2004 12:50:27 EST|
Non deterministic compilers can be viewed in at least three ways:
Non-deterministic compile time:
The ibm pl8 compler used a non deterministic hash code algorithnm,
which effected compile time performance but not generated code.
Non-deterministic code generated:
The Fx32 translator that I built at dec (translated x86 to alpha
binary) was non-deterministic It had a time bound on translation, so
that the compiler would be allowed to do more work on a fast machine.
Non-deterministic code changes as the program exeuctes.
The sun hotspot java compiler, as well as many other jit compilers,
recompile more aggressively routines that took a long time, so the
generated code might vary based on program input.
> glen herrmannsfeldt <firstname.lastname@example.org> writes:
> |> Does anyone know about any compilers that are not deterministic?
Return to the
Search the comp.compilers archives again.