RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Law; GCC > > Patches; Wilco Dijkstra; Michael Meissner; nd > > Subject: RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > > numbers in GIMPLE. > > > > On Wed, 23 Aug 2017, Tamar Christina wrote: > > > > > Hi Richard, > > > > > &g

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > numbers in GIMPLE. > > On Wed, 23 Aug 2017, Tamar Christina wrote: > > > Hi Richard, > > > > Thanks for the feedback, > > > > > > > > I think you should split up your work a bit. The piece

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Wed, 23 Aug 2017, Tamar Christina wrote: > Hi Richard, > > Thanks for the feedback, > > > > > I think you should split up your work a bit. The pieces from > > fold_builtin_* you remove and move to lowering that require no control flow > > like __builtin_isnan (x) -> x UNORD x should move to

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Hi Richard, Thanks for the feedback, > > I think you should split up your work a bit. The pieces from > fold_builtin_* you remove and move to lowering that require no control flow > like __builtin_isnan (x) -> x UNORD x should move to match.pd patterns (aka > foldings that will then be applied

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

gt; Dijkstra; Michael Meissner; nd > Subject: RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers > in GIMPLE. > > > -Original Message- > > From: Richard Biener [mailto:rguent...@suse.de] > > Sent: 08 June 2017 19:23 > > To: Joseph Myers > &g

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

sner; nd > Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > numbers in GIMPLE. > > On June 8, 2017 6:44:01 PM GMT+02:00, Joseph Myers > wrote: > >On Thu, 8 Jun 2017, Richard Biener wrote: > > > >> For a built-in this is generally valid. F

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > numbers in GIMPLE. > > On June 8, 2017 6:44:01 PM GMT+02:00, Joseph Myers > wrote: > >On Thu, 8 Jun 2017, Richard Biener wrote: > > > >> For a built-in this is generally valid. For plain isnan it depends > >o

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Hi Tamar, > I have reverted the patch in commit r 249050 until I can fix the late > expansion > and IBM long double issues. thanks. Just for the record, Solaris/SPARC (also big-endian) was affected as well: https://gcc.gnu.org/ml/gcc-testresults/2017-06/msg00908.html Rainer -

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

>To: Christophe Lyon; Markus Trippelsdorf > >Cc: Joseph Myers; Jeff Law; GCC Patches; Wilco Dijkstra; > >rguent...@suse.de; Michael Meissner; nd > >Subject: RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > >numbers in GIMPLE. > > > &g

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

> Cc: Tamar Christina; Christophe Lyon; Markus Trippelsdorf; Jeffrey Law; Wilco > Dijkstra; Michael Meissner; nd; GCC Patches > Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > numbers in GIMPLE. > > On Thu, 8 Jun 2017, David Edelsohn wrote: > > > Tamar

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 8 Jun 2017, David Edelsohn wrote: > Tamar, > > This patch has caused a large number of new failures on AIX, including > compiler ICEs. It also caused ICEs building glibc for all powerpc configurations (i.e., configurations using IBM long double). https://sourceware.org/ml/libc-testresu

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Tamar, This patch has caused a large number of new failures on AIX, including compiler ICEs. Thanks, David

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On June 8, 2017 6:44:01 PM GMT+02:00, Joseph Myers wrote: >On Thu, 8 Jun 2017, Richard Biener wrote: > >> For a built-in this is generally valid. For plain isnan it depends >on >> what the standard says. >> >> You have to support taking the address of isnan anyway and thus >> expanding to a l

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 8 Jun 2017, Richard Biener wrote: > For a built-in this is generally valid. For plain isnan it depends on > what the standard says. > > You have to support taking the address of isnan anyway and thus > expanding to a library call in that case. Why doesn't that not work? In the case o

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

t;To: Christophe Lyon; Markus Trippelsdorf >Cc: Joseph Myers; Jeff Law; GCC Patches; Wilco Dijkstra; >rguent...@suse.de; Michael Meissner; nd >Subject: RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like >numbers in GIMPLE. > >Thanks, I'm looking at the failure. >

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

de; Michael Meissner; nd Subject: RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. Thanks, I'm looking at the failure. My final validate seems to have only run the GCC tests. > -Original Message- > From: Christophe Lyon [mailto:christophe.l...@li

RE: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

eff Law; GCC Patches; Wilco Dijkstra; > rguent...@suse.de; Michael Meissner; nd > Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like > numbers in GIMPLE. > > On 8 June 2017 at 12:30, Markus Trippelsdorf > wrote: > > On 2017.01.19 at 18:20 +, Joseph

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On 8 June 2017 at 12:30, Markus Trippelsdorf wrote: > On 2017.01.19 at 18:20 +, Joseph Myers wrote: >> On Thu, 19 Jan 2017, Tamar Christina wrote: >> >> > Hi Joseph, >> > >> > I made the requested changes and did a quick pass over the rest >> > of the fp cases. >> >> I've no further comments,

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On 2017.01.19 at 18:20 +, Joseph Myers wrote: > On Thu, 19 Jan 2017, Tamar Christina wrote: > > > Hi Joseph, > > > > I made the requested changes and did a quick pass over the rest > > of the fp cases. > > I've no further comments, but watch out for any related test failures > being reporte

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

May 2, 2017 10:09 AM To: Bernhard Reutner-Fischer; Joseph Myers Cc: Jeff Law; Wilco Dijkstra; rguent...@suse.de; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. Ping. Hi All, Since this didn't manage to get in for GCC 7, I

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Mon, 23 Jan 2017, Tamar Christina wrote: > In the testcase I've also had to remove `-funsafe-math-optimizations` > because at least on AArch64 this sets the LZ bit in the FPCR which > causes the fcmp to collapse small subnormal numbers to 0 and so iszero > would fail. That's not an appropri

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

imizations. From: Joseph Myers Sent: Thursday, January 19, 2017 6:20:35 PM To: Tamar Christina Cc: Jeff Law; GCC Patches; Wilco Dijkstra; rguent...@suse.de; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. On Thu, 19 Jan 2

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 19 Jan 2017, Tamar Christina wrote: > Hi Joseph, > > I made the requested changes and did a quick pass over the rest > of the fp cases. I've no further comments, but watch out for any related test failures being reported. -- Joseph S. Myers jos...@codesourcery.com

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. On Thu, 19 Jan 2017, Tamar Christina wrote: > > The old code set orig_arg before converting IBM long double to double. > > Your code sets it after the conversion. The

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 19 Jan 2017, Tamar Christina wrote: > > The old code set orig_arg before converting IBM long double to double. > > Your code sets it after the conversion. The old code set min_exp based on > > a string set from REAL_MODE_FORMAT (orig_mode)->emin - 1; your code uses > > the adjusted mode.

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

ut then creating the real using build_real (type, rmin) using the adjusted type, shouldn't it not be getting truncated? I've updated and attached the patch. Tamar From: Joseph Myers Sent: Thursday, January 19, 2017 2:46:06 PM To: Tamar Christina

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 19 Jan 2017, Tamar Christina wrote: > > Also, I don't think the call to perform_ibm_extended_fixups in > > is_subnormal is correct. Subnormal for IBM long double is *not* the same > > as subnormal double high part. Likewise it's incorrect in is_normal as > > well. > > The calls to is_ze

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

iszero and issubnormal. Tamar. From: Joseph Myers Sent: Wednesday, January 18, 2017 5:05:07 PM To: Jeff Law Cc: Tamar Christina; GCC Patches; Wilco Dijkstra; rguent...@suse.de; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Wed, 18 Jan 2017, Joseph Myers wrote: > Generally, I don't see tests added that these new functions are correct > for float, double and long double, which would detect such issues if run > for a target with IBM long double. Specifically, I think gcc.dg/tg-tests.h should have tests added for

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Also, I don't think the call to perform_ibm_extended_fixups in is_subnormal is correct. Subnormal for IBM long double is *not* the same as subnormal double high part. Likewise it's incorrect in is_normal as well. Generally, I don't see tests added that these new functions are correct for flo

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Since the patch adds new built-in functions __builtin_issubnormal and __builtin_iszero, it also needs to update c-typeck.c:convert_arguments to make those functions remove excess precision. This is mentioned in the PRs 77925 and 77926 for addition of those functions (which as I noted in

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

! Kind Regards, Tamar From: Joseph Myers Sent: Thursday, December 15, 2016 7:03:27 PM To: Tamar Christina Cc: Jeff Law; GCC Patches; Wilco Dijkstra; rguent...@suse.de; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE l

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. Hi Jeff, I wasn't sure if you saw the updated patch attached to the previous email or if you just hadn't had the time to look at it yet. Cheers, Tamar

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

rs Cc: GCC Patches; Wilco Dijkstra; rguent...@suse.de; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. On 12/15/2016 03:14 AM, Tamar Christina wrote: > >> On a high level, presumably there's no real value in keeping

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On 12/15/2016 03:14 AM, Tamar Christina wrote: On a high level, presumably there's no real value in keeping the old code to "fold" fpclassify. By exposing those operations as integer logicals for the fast path, if the FP value becomes a constant during the optimization pipeline we'll see the r

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

s so far! Kind Regards, Tamar From: Joseph Myers Sent: Thursday, December 15, 2016 7:03:27 PM To: Tamar Christina Cc: Jeff Law; GCC Patches; Wilco Dijkstra; rguent...@suse.de; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPL

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 15 Dec 2016, Tamar Christina wrote: > > Note that on some systems we even disable 64bit floating point support. > > I suspect this check needs a little re-thinking as I don't think that > > checking for a specific UNITS_PER_WORD is correct, nor is checking the > > width of the type. I'm n

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

> On a high level, presumably there's no real value in keeping the old > code to "fold" fpclassify. By exposing those operations as integer > logicals for the fast path, if the FP value becomes a constant during > the optimization pipeline we'll see the reinterpreted values flowing > into the new

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On 11/25/2016 05:18 AM, Tamar Christina wrote: Hi Joseph, I have updated the patch with the changes, w.r.t to the formatting, there are tabs there that seem to be rendered at 4 spaces wide. In my editor setup at 8 spaces they are correct. Various random comments, mostly formatting. I do think

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. Ping? Is there anything else I need to do for the patch or is it OK for trunk? Just a note, it is in my queue of things to look at. GIven it was posted well before stage1 close I think it deserves the opportunity to move forward (even

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. Hi Joseph, I have updated the patch with the changes, w.r.t to the formatting, there are tabs there that seem to be rendered at 4 spaces wide. In my editor setup at 8 spaces they are

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. Hi Joseph, I have updated the patch with the changes, w.r.t to the formatting, there are tabs there that seem to be rendered at 4 spaces wide. In my editor setup at 8 spaces they are correct

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

, November 24, 2016 6:28:18 PM To: Tamar Christina Cc: GCC Patches; Wilco Dijkstra; rguent...@suse.de; l...@redhat.com; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. On Thu, 24 Nov 2016, Tamar Christina wrote: > @@ -11499,6 +11503,53 @@

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Thu, 24 Nov 2016, Tamar Christina wrote: > @@ -11499,6 +11503,53 @@ to classify. GCC treats the last argument as > type-generic, which > means it does not do default promotion from float to double. > @end deftypefn > > +@deftypefn {Built-in Function} int __builtin_isnan (...) > +This buil

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

:43 PM To: Tamar Christina Cc: GCC Patches; Wilco Dijkstra; rguent...@suse.de; l...@redhat.com; Michael Meissner; nd Subject: Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE. On Fri, 11 Nov 2016, Tamar Christina wrote: > is_infinite and is_zero have been crea

Re: [PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

On Fri, 11 Nov 2016, Tamar Christina wrote: > is_infinite and is_zero have been created. This patch also introduces two new > intrinsics __builtin_iszero and __builtin_issubnormal. And so the ChangeLog entry needs to include: PR middle-end/77925 PR middle-end/77926 (in addition

[PATCH][GCC][PATCHv3] Improve fpclassify w.r.t IEEE like numbers in GIMPLE.

Hi All, This is v3 of the patch which adds an optimized route to the fpclassify builtin for floating point numbers which are similar to IEEE-754 in format. The patch has been rewritten to do it in GIMPLE instead of a fold. As part of the implementation optimized versions of is_normal, is_subnorma