[llvm-dev] Fwd: cfl-aa

[George Burgess IV via llvm-dev]

+Jia, who worked on CFLAA this summer.

Hi!

CFLAA was in its very very early stages at the time of the 3.7 release.
Even now, as we're getting ready for 3.9, it's still experimental (though
hopefully quite a bit more accurate than it was last year). So, yes, in 3.7
and 3.8, CFLAA was *very* conservative. :)

If you'd like numbers on what the current CFL implementations can do, you
can find them at writeup.pdf here: https://github.com/grievejia/GSoC2016

No clue what the status of scev is/was, so I can't talk much about that.

Thanks,
George

On Mon, Aug 29, 2016 at 9:36 PM, Vitor Paisante via llvm-dev <
llvm-dev at lists.llvm.org> wrote:

> dear LLVMers,
>     I am trying to use some of the LLVM alias analyses, and I would like
> to check two things with you: is scev-aa being maintained in LLVM 3.7?
> Second question: I run cfl-aa, and I got a very small number of pointer
> disambiguation (no alias) with it. My results for SPEC CINT 2006 follow
> below. Is this low number of no alias responses something to be excepted?
> Below the results that I got for SPEC with cfl-aa:
>
> TotalQueries   |   Name   |   NoAliasResponses
> 31944   |   470.lbm   |   0
> 49133   |   429.mcf   |   42
> 95098   |   473.astar   |   0
> 146301   |   462.libquantum   |   5
> 428082   |   458.sjeng   |   9773
> 808471   |   433.milc   |   2163
> 1787190   |   450.soplex   |   72
> 2472234   |   401.bzip2   |   229
> 2574217   |   456.hmmer   |   1833
> 3492577   |   445.gobmk   |   8480
> 3685838   |   444.namd   |   616
> 12943554   |   471.omnetpp   |   422
> 20068605   |   464.h264ref   |   8593
> 23849576   |   400.perlbench   |   99316
> 37779455   |   447.dealII   |   11204
> 186008992   |   403.gcc   |   404828
>
>     I am finding these results weird because I was expecting a larger
> number of no-alias responses. For instance, I got only 404,828 responses
> out of 186,008,992 queries. Has anyone gotten similar, or different results?
>
> Regards,
> Vitor Mendes Paisante
>
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160829/ce926092/attachment.html>