RE: New code merge optimization?

[Sebastian Perta <Sebastian dot Perta at renesas dot com>]

Hi,

Thank you!
As the description says this finds equivalent functions, I would like to find identical sequences inside
functions but at least this will provide all the up to date tools to compare code sequences, thank you again!

Best Regards,
Sebastian


> -----Original Message-----
> From: Martin Jambor [mailto:mjambor@suse.cz]
> Sent: 18 January 2018 15:18
> To: Sebastian Perta <Sebastian.Perta@renesas.com>; gcc-
> patches@gcc.gnu.org
> Cc: loki@gcc.gnu.org
> Subject: Re: New code merge optimization?
>
> Hi,
>
> On Thu, Jan 18 2018, Sebastian Perta wrote:
> > Hello,
> >
> > I am interested in implementing a new pass in gcc to merge identical
> > sequences of code in GCC to be used mainly for RL78.
> > The commercial RL78 compilers have such algorithms implemented and
> they make
> > quite good use of it.
> > Opportunities arise from the limited capabilities of RL78, for other targets
> > this might be a lot less useful.
> >
> > A while ago I found the following:
> > https://www.gnu.org/software/gcc/projects/cfo.html
> > And I ported all algorithms to gcc 4.9.2 and tried it on RL78 and RX and
> > this is what I found out:
> > For RX: no visible improvements with any of them
> > For RL78: some minor improvements only with -frtl-seqbastr:
> > Compiling all the C files from gcc/testsuite/gcc.c-torture/execute/*c  with
> > "-Os" and "-Os  -frtl-seqabstr" (using the modified gcc 4.9.2)
> > The algorithm was effective only in 60 files(out of 1643 files, that's only
> > 0.03% of the files currently present in gcc/testsuite/gcc.c-torture/execute)
> > On those 60 files I got an average of 6.5% improvement with the best
> > improvement for pr58574.c (36.4%).
> >
> > What do you think: is it worthwhile porting this to the trunk or I will just
> > waste my time?
> > Or should I start fresh? Maybe start from here:
> > http://llvm.org/docs/MergeFunctions.html?
>
> Martin Liška contributed identical code folding to GCC quite a few years
> ago now.  Look up the -fipa-icf option.  If that is not enough for you,
> starting from that seems more natural.
>
> Martin
>
> >
> > Gimple or rtl?
> > I suppose the preferred way will be to do this in gimple; however based on
> > what I did so far,  it's more likely to find identical sequences in rtl (for
> > RL78).
> >
> > Any thoughts? Thank you!
> >
> > Best Regards,
> > Sebastian



Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.