Re: [RFC][AARCH64][PATCH 5/5] add aarch64_loop_unroll_adjust to limit partial unrolling in rtl based on strided-loads in loop

Hi Ramana On 15 September 2017 at 18:40, Ramana Radhakrishnan wrote: > On Fri, Sep 15, 2017 at 2:33 AM, Kugan Vivekanandarajah > wrote: >> This patch adds aarch64_loop_unroll_adjust to limit partial unrolling >> in rtl based on

Re: [RFC][AARCH64][PATCH 2/5]: Add number of hw prefetchers available to cpu_prefetch_tune

Hi Andrew, On 15 September 2017 at 13:20, Andrew Pinski wrote: > On Thu, Sep 14, 2017 at 6:28 PM, Kugan Vivekanandarajah > wrote: >> This patch adds number of hw prefetchers available to >> cpu_prefetch_tune so it can be used in loop

[RFA] Implement __VA_OPT__

This implements __VA_OPT__, a new preprocessor feature added in C++2A. The paper can be found here: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0306r4.html I am not completely sure that I have handled the error reporting correctly. I chose to allow __VA_OPT__ generally, on the

Re: [PATCH] Factor out division by squares and remove division around comparisons (1/2)

On 15 September 2017 18:50:26 CEST, Jeff Law wrote: >On 09/13/2017 03:20 PM, Wilco Dijkstra wrote: >> Jeff Law wrote: >>> On 09/06/2017 03:55 AM, Jackson Woodruff wrote: On 08/30/2017 01:46 PM, Richard Biener wrote: >> >rdivtmp = 1 / (y*C); >tem = x

Re: GCC 5.5 Status Report (2017-09-16)

On Sat, Sep 16, 2017 at 3:03 PM, Jakub Jelinek wrote: > Status > == > > The GCC 5 branch is still open for regression and documentation fixes > but it's about time to close the branch with a last release from it. > Thus at the end of the next week I plan to do a RC for GCC

Re: Add a vect_get_num_copies helper routine

"H.J. Lu" writes: > On Thu, Sep 14, 2017 at 4:22 AM, Richard Sandiford > wrote: >> This patch adds a vectoriser helper routine to calculate how >> many copies of a vector statement we need. At present this >> is always: >> >>

[committed] PR82228: Move ncopies calculation in vectorizable_live_operation

...which should have been after the early exit for non-vectorised statements. (Actually it could be after the !vec_stmt test in current sources, but one of the later SVE patches needs it to be before, and that's also the standard position in other vectorizable_* routines.) Tested on

GCC 5.5 Status Report (2017-09-16)

Status == The GCC 5 branch is still open for regression and documentation fixes but it's about time to close the branch with a last release from it. Thus at the end of the next week I plan to do a RC for GCC 5.5 following with a release and the branch closing game. Please consider

Backports to 5.x

Hi! I've backported following patches to 5.x, bootstrapped/regtested them on x86_64-linux and i686-linux and committed. Jakub 2017-09-16 Jakub Jelinek Backported from mainline 2017-06-21 Jakub Jelinek PR c++/81154

[Patch, Fortran] PR 82018: -Wextra should imply -Wconversion-extra

Hi all, here is a small patch that enables -Wconversion-extra with -Wextra and updates the documentation. Up to gfortran 5, the warning in the test case was enabled by -Wconversion, and thus with -Wall. That was changed by Thomas in r224190, which makes complete sense to me. However, I think

Re: Add a vect_get_num_copies helper routine

On Thu, Sep 14, 2017 at 4:22 AM, Richard Sandiford wrote: > This patch adds a vectoriser helper routine to calculate how > many copies of a vector statement we need. At present this > is always: > > LOOP_VINFO_VECT_FACTOR (loop_vinfo) / TYPE_VECTOR_SUBPARTS

Re: Backports for GCC 6 branch

> One more that I've just tested. On which platform? Here's what I have on x86_64-suse-linux: FAIL: gcc.dg/asan/pr81224.c -O0 (internal compiler error) FAIL: gcc.dg/asan/pr81224.c -O0 (test for excess errors) FAIL: gcc.dg/asan/pr81224.c -O1 (internal compiler error) FAIL:

Re: [PATCH] Add comments to struct cgraph_thunk_info

On 15 September 2017 18:20:33 CEST, Pierre-Marie de Rodat wrote: >On 09/15/2017 06:10 PM, Jeff Law wrote: >> OK. >> jeff > >Committed. Thanks! + * for result-adjusting thinks, the FIXED_OFFSET adjustment is done after s/think/thunk/ TIA

Re: Transform (x / y) != 0 to x >=y and (x / y) == 0 to x < y if x, y are unsigned

Wilco Dijkstra writes: > Marc Glisse wrote: > >> The question is whether, having computed c=a/b, it is cheaper to test >> a> I think it is usually the second one, but not for all types on all >> targets. Although since >> you mention VRP, it is easier to do further