<html><body>
<p><font size="2" face="sans-serif">Commit r225795.</font><br>
<br>
<font size="2" face="sans-serif">Olivier</font><br>
<br>
<img width="16" height="16" src="cid:1__=0ABBF75FDFC0DE818f9e8a93df938@us.ibm.com" border="0" alt="Inactive hide details for Hal Finkel ---01/13/2015 03:55:35 AM-------- Original Message ----- > From: "Olivier H Sallenave" <oh"><font size="2" color="#424282" face="sans-serif">Hal Finkel ---01/13/2015 03:55:35 AM-------- Original Message ----- > From: "Olivier H Sallenave" <ohsallen@us.ibm.com></font><br>
<br>
<font size="1" color="#5F5F5F" face="sans-serif">From: </font><font size="1" face="sans-serif">Hal Finkel <hfinkel@anl.gov></font><br>
<font size="1" color="#5F5F5F" face="sans-serif">To: </font><font size="1" face="sans-serif">Olivier H Sallenave/Watson/IBM@IBMUS</font><br>
<font size="1" color="#5F5F5F" face="sans-serif">Cc: </font><font size="1" face="sans-serif"><llvm-commits@cs.uiuc.edu>, Owen Anderson <resistor@mac.com></font><br>
<font size="1" color="#5F5F5F" face="sans-serif">Date: </font><font size="1" face="sans-serif">01/13/2015 03:55 AM</font><br>
<font size="1" color="#5F5F5F" face="sans-serif">Subject: </font><font size="1" face="sans-serif">Re: [PATCH] [LLVMdev] More FMA folding opportunities</font><br>
<hr width="100%" size="2" align="left" noshade style="color:#8091A5; "><br>
<br>
<br>
<tt><font size="2">----- Original Message -----<br>
> From: "Olivier H Sallenave" <ohsallen@us.ibm.com><br>
> To: "Owen Anderson" <resistor@mac.com><br>
> Cc: "Hal Finkel" <hfinkel@anl.gov>, llvm-commits@cs.uiuc.edu<br>
> Sent: Monday, January 12, 2015 2:01:28 PM<br>
> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> <br>
> Hi,<br>
> <br>
> <br>
> Attached is the suggested patch.<br>
<br>
LGTM.<br>
<br>
-Hal<br>
<br>
> <br>
> <br>
> Thanks,<br>
> Olivier<br>
> <br>
> <br>
> <br>
> -----Owen Anderson <resistor@mac.com> wrote: -----<br>
> <br>
> To: Hal Finkel <hfinkel@anl.gov><br>
> From: Owen Anderson <resistor@mac.com><br>
> Date: 01/11/2015 10:50PM<br>
> Cc: Olivier H Sallenave/Watson/IBM@IBMUS, llvm-commits@cs.uiuc.edu<br>
> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> <br>
> <br>
> > On Jan 9, 2015, at 11:19 PM, Hal Finkel <hfinkel@anl.gov> wrote:<br>
> > <br>
> > ----- Original Message -----<br>
> >> From: "Olivier H Sallenave" <ohsallen@us.ibm.com><br>
> >> To: "Hal Finkel" <hfinkel@anl.gov><br>
> >> Cc: llvm-commits@cs.uiuc.edu<br>
> >> Sent: Friday, January 9, 2015 11:59:07 AM<br>
> >> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> >> <br>
> >> Hi Hal,<br>
> >> <br>
> >> I am wondering if we should move the FP_EXTEND cases to the<br>
> >> PPC-specific combiner:<br>
> >> <br>
> >> (fadd (fpext (fmul x, y)), z) -> (fma (fpext x), (fpext y), z)<br>
> >> <br>
> >> We know that for PPC the FPEXT nodes will be removed, but it'll<br>
> >> likely not be the case for other architectures. For those, you end<br>
> >> up with one FMADD rather than one FMUL and one FADD, but also have<br>
> >> two FPEXT nodes rather than one. I would not expect this to take<br>
> >> more cycles to run, but there's no way to systematically check<br>
> >> that<br>
> >> since the cost models are not necessarily accurate.<br>
> >> <br>
> >> In the end, the "aggressive" flag only ensures that you can<br>
> >> combine<br>
> >> one FMUL node with many FADD nodes.<br>
> >> <br>
> >> What do you think?<br>
> > <br>
> > I think we already have too much code in the PPCISelLowering that<br>
> > could be in the target-independent combiner, appropriately<br>
> > guarded. We already have TLI hooks for isTruncateFree, isZExtFree,<br>
> > isFNegFree, isFAbsFree, and we could certainly add an isFPExtFree<br>
> > hook at well.<br>
> <br>
> That would make a lot of sense to me.<br>
> <br>
> —Owen<br>
> <br>
> > -Hal<br>
> > <br>
> >> <br>
> >> Olivier<br>
> >> <br>
> >> Inactive hide details for Hal Finkel ---01/08/2015 08:36:56<br>
> >> PM-------- Original Message ----- > From: "Hal Finkel"<br>
> >> <hfinkel@anlHal Finkel ---01/08/2015 08:36:56 PM-------- Original<br>
> >> Message ----- > From: "Hal Finkel" <hfinkel@anl.gov><br>
> >> <br>
> >> From: Hal Finkel <hfinkel@anl.gov><br>
> >> To: Owen Anderson <resistor@mac.com><br>
> >> Cc: Olivier H Sallenave/Watson/IBM@IBMUS,<br>
> >> <llvm-commits@cs.uiuc.edu><br>
> >> Date: 01/08/2015 08:36 PM<br>
> >> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> >> <br>
> >> <br>
> >> <br>
> >> <br>
> >> ----- Original Message -----<br>
> >>> From: "Hal Finkel" <hfinkel@anl.gov><br>
> >>> To: "Owen Anderson" <resistor@mac.com><br>
> >>> Cc: "Olivier H Sallenave" <ohsallen@us.ibm.com>,<br>
> >>> llvm-commits@cs.uiuc.edu<br>
> >>> Sent: Thursday, January 8, 2015 6:48:34 PM<br>
> >>> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> >>> <br>
> >>> ----- Original Message -----<br>
> >>>> From: "Hal Finkel" <hfinkel@anl.gov><br>
> >>>> To: "Owen Anderson" <resistor@mac.com><br>
> >>>> Cc: "Olivier H Sallenave" <ohsallen@us.ibm.com>,<br>
> >>>> llvm-commits@cs.uiuc.edu<br>
> >>>> Sent: Thursday, January 8, 2015 6:19:14 PM<br>
> >>>> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> >>>> <br>
> >>>> ----- Original Message -----<br>
> >>>>> From: "Owen Anderson" <resistor@mac.com><br>
> >>>>> To: "Hal Finkel" <hfinkel@anl.gov><br>
> >>>>> Cc: "Olivier H Sallenave" <ohsallen@us.ibm.com>,<br>
> >>>>> llvm-commits@cs.uiuc.edu<br>
> >>>>> Sent: Thursday, January 8, 2015 6:16:57 PM<br>
> >>>>> Subject: Re: [PATCH] [LLVMdev] More FMA folding opportunities<br>
> >>>>> <br>
> >>>>> <br>
> >>>>> <br>
> >>>>> <br>
> >>>>> <br>
> >>>>> <br>
> >>>>> On Jan 8, 2015, at 6:06 PM, Hal Finkel < hfinkel@anl.gov ><br>
> >>>>> wrote:<br>
> >>>>> <br>
> >>>>> Olivier, Aditya is right. You need to add the FP_EXTEND to the<br>
> >>>>> operands of the resulting FMA as appropriate. Can you please<br>
> >>>>> fix<br>
> >>>>> that? We can also move those under the "aggressive" flag for<br>
> >>>>> now.<br>
> >>>>> <br>
> >>>>> <br>
> >>>>> Can you fix or revert this ASAP, please? The missing extends<br>
> >>>>> are<br>
> >>>>> breaking our branches pretty badly.<br>
> >>>> <br>
> >>>> Yep, will do in a few minutes.<br>
> >>> <br>
> >>> I moved the transformations behind the aggressive flag in r225485<br>
> >>> --<br>
> >>> that should unbreak your builds while I fix the extends.<br>
> >> <br>
> >> FP_EXTEND nodes added as of r225492. Thanks for pointing this out!<br>
> >> <br>
> >> -Hal<br>
> >> <br>
> >>> <br>
> >>> -Hal<br>
> >>> <br>
> >>>> <br>
> >>>> -Hal<br>
> >>>> <br>
> >>>>> <br>
> >>>>> <br>
> >>>>> —Owen<br>
> >>>> <br>
> >>>> --<br>
> >>>> Hal Finkel<br>
> >>>> Assistant Computational Scientist<br>
> >>>> Leadership Computing Facility<br>
> >>>> Argonne National Laboratory<br>
> >>>> <br>
> >>>> _______________________________________________<br>
> >>>> llvm-commits mailing list<br>
> >>>> llvm-commits@cs.uiuc.edu<br>
> >>>> </font></tt><tt><font size="2"><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a></font></tt><tt><font size="2"><br>
> >>>> <br>
> >>> <br>
> >>> --<br>
> >>> Hal Finkel<br>
> >>> Assistant Computational Scientist<br>
> >>> Leadership Computing Facility<br>
> >>> Argonne National Laboratory<br>
> >>> <br>
> >>> _______________________________________________<br>
> >>> llvm-commits mailing list<br>
> >>> llvm-commits@cs.uiuc.edu<br>
> >>> </font></tt><tt><font size="2"><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a></font></tt><tt><font size="2"><br>
> >>> <br>
> >> <br>
> >> --<br>
> >> Hal Finkel<br>
> >> Assistant Computational Scientist<br>
> >> Leadership Computing Facility<br>
> >> Argonne National Laboratory<br>
> >> <br>
> >> <br>
> >> <br>
> > <br>
> > --<br>
> > Hal Finkel<br>
> > Assistant Computational Scientist<br>
> > Leadership Computing Facility<br>
> > Argonne National Laboratory<br>
> > <br>
> > _______________________________________________<br>
> > llvm-commits mailing list<br>
> > llvm-commits@cs.uiuc.edu<br>
> > </font></tt><tt><font size="2"><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a></font></tt><tt><font size="2"><br>
> <br>
> <br>
<br>
-- <br>
Hal Finkel<br>
Assistant Computational Scientist<br>
Leadership Computing Facility<br>
Argonne National Laboratory<br>
<br>
</font></tt><br>
</body></html>