octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Question of JITC


From: lyh.kernel
Subject: Re: Question of JITC
Date: Thu, 8 Aug 2013 16:11:22 +0800

Hello Max,

There are three JITC entry points currently - one for simple for statement, one for while statement and the other one for do_multi_index_op. I think do_multi_index_op JIT compiling would be invoked in function call/matrix indexing, but I have no idea how to trigger it. Would you mind to give me a example?

Thanks a lot


2013/8/8 Max Brister <address@hidden>
On Sun, Aug 4, 2013 at 11:10 AM, lyh.kernel <address@hidden> wrote:
>
> Hello Max,
>
> I want to look into the source code of type inference. I noticed that you have mentioned three papers about type inference in your blog. Would you mind to tell me which paper should I take a took? And which part of source should I traced?
>
> Thanks a lot

The approach to type inference was inspired by the Falcon MATLAB
compiler [1]. Practically, implementation ended up being similar to
the Sparse Conditional Constant Propagation algorithm (SCCP)[2]. The
Conditional part of SCCP is currently limited to proving error checks
are unnecessary, however.

[1]  L. D. Rose and D. Padua. Techniques for the Translation of MATLAB
Programs into Fortran 90. ACM Transactions on Programming Languages
and Systems, 21(2):286–323, Mar. 1999.
[1] http://www.cs.rice.edu/~keith/512/2011/Lectures/L19-SCCP-1up.pdf

--
Max Brister


reply via email to

[Prev in Thread] Current Thread [Next in Thread]