Re: [PATCH V2, RFC] Fix PR62147 by passing finiteness information to RTL phase

Hi all, I've committed this and with one more change. --- gcc/loop-iv.c (revision 272731) +++ gcc/loop-iv.c (working copy) @@ -3004,7 +3004,7 @@ find_simple_exit (struct loop *loop, struct niter_ well. It results in incorrect predicate information on the exit condition

[PATCH v3 2/2] PR c/65403 - Add tests for -Wno-error=

* c-c++-common/pr65403-1.c: New test. * c-c++-common/pr65403-2.c: New test. --- gcc/testsuite/c-c++-common/pr65403-1.c | 10 ++ gcc/testsuite/c-c++-common/pr65403-2.c | 15 +++ 2 files changed, 25 insertions(+) create mode 100644

[PATCH v3 1/2] PR c/65403 - Ignore -Wno-error=

From: Manuel López-Ibáñez * opts-common.c (ignored_wnoerror_options): New global variable. * opts-global.c (print_ignored_options): Ignore -Wno-error= except if there are other diagnostics. * opts.c (enable_warning_as_error): Record ignored -Wno-error

[PATCH] rs6000: Enable -fvariable-expansion-in-unroller by default

Hi, We've done some experimenting and realized that the subject option almost always provide improved performance for Power when the loop unroller is enabled. So this patch turns that flag on by default for us. Bootstrapped and tested on powerpc64le-unknown-linux-gnu with no regressions. Is

Re: [PATCH] constrain one character optimization to one character stores (PR 90989)

On 6/26/19 4:31 PM, Jeff Law wrote: On 6/25/19 5:03 PM, Martin Sebor wrote: The caller ensures that handle_char_store is only called for stores to arrays (MEM_REF) or single elements as wide as char. Where? I don't see it, even after fixing the formatting in strlen_check_and_optimize_stmt

Re: [PATCH] Enable GCC support for AVX512_VP2INTERSECT.

On Wed, Jun 26, 2019 at 6:14 PM Rainer Orth wrote: > > Hi Hongtao, > > > Index: testsuite/lib/target-supports.exp > > === > > --- testsuite/lib/target-supports.exp (revision 272667) > > +++ testsuite/lib/target-supports.exp (working

RE: Use ODR for canonical types construction in LTO

Hi, This commit https://gcc.gnu.org/viewcvs/gcc?view=revision=272628 is breaking trunk LTO on some real benchmarks, so can it be fixed or reverted? For example, lto1: error: type variant differs by TYPE_CXX_ODR_P constant 256> unit-size constant 32> align:64 warn_if_not_align:0

Re: [RFA][tree-optimization/90883] Improve DSE to handle redundant calls

On 6/26/19 7:14 PM, Bill Schmidt wrote: > Looks like this patch breaks bootstrap. > > /home3/wschmidt/gcc/gcc-mainline-base/gcc/tree-ssa-dse.c: In function > 'void dse\ > _optimize_redundant_stores(gimple*)': > /home3/wschmidt/gcc/gcc-mainline-base/gcc/tree-ssa-dse.c:649:46: error: > ISO C++\ >  

Re: [RFA][tree-optimization/90883] Improve DSE to handle redundant calls

Looks like this patch breaks bootstrap. /home3/wschmidt/gcc/gcc-mainline-base/gcc/tree-ssa-dse.c: In function 'void dse\ _optimize_redundant_stores(gimple*)': /home3/wschmidt/gcc/gcc-mainline-base/gcc/tree-ssa-dse.c:649:46: error: ISO C++\  forbids converting a string constant to 'char*'

Re: [PATCH] warn on returning alloca and VLA (PR 71924, 90549)

On 6/18/19 9:19 PM, Martin Sebor wrote: > On 6/14/19 2:59 PM, Jeff Law wrote: [ big snip ] >> A COND_EXPR on the RHS of an assignment is valid gimple.  That's what we >> need to consider here -- what is and what is not valid gimple.  And its >> more likely that PHIs will be transformed into RHS 

[PATCH] C++20 constexpr lib part 2/3

Implement C++20 p0879 - Constexpr for swap and swap related functions. This is much smaller than the first but also basically marks swap and the algorithms that depend on swap as constexpr. It is similarly tested on x86_64-linux: $ make check -k -j4 $ make check

[PATCH 1/3] C++20 constexpr lib part 1/3

Here is the first of three patches for C++20 constexpr library. ?? Implement C++20 p0202 - Add constexpr Modifiers to Functions in and Headers. ??Implement C++20 p1023 - constexpr comparison operators for std::array. I believe I have answered peoples concerns with the last patch

[PATCH] Define std::chars_format enumeration type

This type isn't used anywhere yet, but will be needed for the floating-point overloads of to_chars and from_chars. * include/std/charconv (chars_format): Define bitmask type. * testsuite/20_util/to_chars/chars_format.cc: New test. Tested x86_64-linux, committed to trunk.

Re: [PATCH] RISC-V: Add -malign-data= option.

Hmm, may I suggest use "natural" rather than "abi" and 32bit or 64bit rather than "word"; it is not obvious what abi means and it is not obvious what word means here; it could be either 32bit or 64bit depending on the option. It's actually worse: in RISC-V "word" always means 32-bit

Re: [PATCH] constrain one character optimization to one character stores (PR 90989)

On 6/25/19 5:03 PM, Martin Sebor wrote: > > The caller ensures that handle_char_store is only called for stores > to arrays (MEM_REF) or single elements as wide as char. Where? I don't see it, even after fixing the formatting in strlen_check_and_optimize_stmt :-) > gimple *stmt = gsi_stmt

Re: [PATCH] Fix warnings seen by clang in gcc/symbol-summary.h.

On 6/26/19 12:46 AM, Martin Liška wrote: > Hi. > > The patch is about missing argument to function call and > unused arguments in symbol-summary.h. > > Patch can bootstrap on x86_64-linux-gnu and survives regression tests. > > Ready to be installed? > Thanks, > Martin > > gcc/ChangeLog: > >

Re: [PATCH 31/30] Update documentation for movmem to cpymem change

On 6/26/19 2:16 PM, Aaron Sawdey wrote: > On 6/25/19 4:43 PM, Jeff Law wrote: >> On 6/25/19 2:22 PM, acsaw...@linux.ibm.com wrote: >>> From: Aaron Sawdey >>> >>> * builtins.c (get_memory_rtx): Fix comment. >>> * optabs.def (movmem_optab): Change to cpymem_optab. >>> * expr.c

Re: [PATCH 23/30] Changes to rs6000

On Tue, Jun 25, 2019 at 03:22:32PM -0500, acsaw...@linux.ibm.com wrote: > From: Aaron Sawdey > > * config/rs6000/rs6000.md: (movmemsi) Change name to cpymemsi. This is fine. Thanks! Segher

Re: [PATCH] Remove quite obvious dead assignments.

On 6/26/19 4:57 AM, Martin Liška wrote: > Hi. > > I've spent some with clang-static-analyzer and I analyzed the warnings > reported. > As always wit analyzers, majority of the issues are false positives, however > it caught > couple of real issues: > >

[PATCH 31/30] Update documentation for movmem to cpymem change

On 6/25/19 4:43 PM, Jeff Law wrote: > On 6/25/19 2:22 PM, acsaw...@linux.ibm.com wrote: >> From: Aaron Sawdey >> >> * builtins.c (get_memory_rtx): Fix comment. >> * optabs.def (movmem_optab): Change to cpymem_optab. >> * expr.c (emit_block_move_via_cpymem): Change movmem to cpymem.

[PATCH] PR debug/90981 Empty .debug_addr crashes -gdwarf-5 -gsplit-dwarf

Even if there was no, or an empty address list we would try to generate and index for the .debug_addr section with -gdwarf-5 and -gsplit-dwarf. The skeleton DIE would also get a (dangling) DW_AT_addr_base in that case. PR debug/90981 * dwarf2out.c

Re: [PATCH] Fix misc stuff seen by clang-static-analyzer.

On 6/26/19 4:57 AM, Martin Liška wrote: > Hi. > > This small stuff handles a misc clang-static-analyzer issues. > > Patch can bootstrap on x86_64-linux-gnu and survives regression tests. > > Ready to be installed? > Thanks, > Martin > > gcc/lto/ChangeLog: > > 2019-06-26 Martin Liska > >

Re: [SVE] [fwprop] PR88833 - Redundant moves for WHILELO-based loops

Prathamesh Kulkarni writes: > On Wed, 26 Jun 2019 at 16:05, Richard Sandiford > wrote: >> >> Prathamesh Kulkarni writes: >> > On Tue, 25 Jun 2019 at 20:05, Richard Sandiford >> > wrote: >> >> >> >> Prathamesh Kulkarni writes: >> >> > On Mon, 24 Jun 2019 at 21:41, Prathamesh Kulkarni >> >> >

Re: [RFA][tree-optimization/90883] Improve DSE to handle redundant calls

On 6/26/19 5:53 AM, Richard Biener wrote: > On Wed, Jun 26, 2019 at 6:17 AM Jeff Law wrote: >> >> So based on the conversation in the BZ I cobbled together a patch to >> extend tree-ssa-dse.c to also detect redundant stores. >> >> To be clear, given two stores, the first store is dead if the

Re: [RFA][tree-optimization/90883] Improve DSE to handle redundant calls

On 6/26/19 5:53 AM, Richard Biener wrote: > On Wed, Jun 26, 2019 at 6:17 AM Jeff Law wrote: >> >> So based on the conversation in the BZ I cobbled together a patch to >> extend tree-ssa-dse.c to also detect redundant stores. >> >> To be clear, given two stores, the first store is dead if the

Re: [PATCH, PPC 2/2] Fix Darwin bootstrap after split of rs6000.c.

On Wed, Jun 26, 2019 at 04:58:06PM +0100, Iain Sandoe wrote: > The recent change in the file layout in rs6000 breaks Darwin bootstrap. > > To fix this we need to make the branch islands (or code) visible between > both files. I chose to keep the generation side in rs6000.c and move > the output

Re: [PATCH, PPC 1/2] Make sure the common gt-*.h files are built for all sub-targets.

Hi Iain, On Wed, Jun 26, 2019 at 04:57:12PM +0100, Iain Sandoe wrote: > The new gt-rs6000-logue.h is common to all sub-targets in the port, so > it needs to be added for them. > > It seems better to place the common target_gtfiles in the powerpc*-*-* > section, rather than duplicating them in

Re: [gomp4.5] Handle #pragma omp declare target link

Hi! On Mon, 14 Dec 2015 20:17:33 +0300, Ilya Verbin wrote: > Here is an updated patch [for "#pragma omp declare target link"] ..., that got committed long ago (trunk r231655), with additional changes later on. As has later been filed in PR81689, the test case added "libgomp.c/target-link-1.c

[Committed] Document revision 272667

WHen I committed revision 272667 last nigth, I forgot to commit the updated ChangeLog entries. This commit documents the changes. 2016-06-26 Steven G. Kargl * ChangeLog: Document revision 272667 2016-06-26 Steven G. Kargl * testsuite/ChangeLog: Document revision 272667

[PATCH, PPC 2/2] Fix Darwin bootstrap after split of rs6000.c.

The recent change in the file layout in rs6000 breaks Darwin bootstrap. To fix this we need to make the branch islands (or code) visible between both files. I chose to keep the generation side in rs6000.c and move the output routine to rs6000-logue.c, placing a reference to the islands vector in

[PATCH, PPC 1/2] Make sure the common gt-*.h files are built for all sub-targets.

The new gt-rs6000-logue.h is common to all sub-targets in the port, so it needs to be added for them. It seems better to place the common target_gtfiles in the powerpc*-*-* section, rather than duplicating them in sub-targets. This would make it less likely that a sub-target would be overlooked

Re: [PATCH][gcc] libgccjit: check result_type in gcc_jit_context_new_binary_op

David Malcolm writes: > On Tue, 2019-06-25 at 08:11 +, Andrea Corallo wrote: >> Hi, >> third version for this patch with the simplified test. >> >> make check-jit pass clean >> >> Bests >> Andrea >> >> 2019-06-09 Andrea Corallo andrea.cora...@arm.com >> >> * libgccjit.c

[PING] [PATCH] warn on returning alloca and VLA (PR 71924, 90549)

Ping: did my reply and updated patch resolve your concerns? https://gcc.gnu.org/ml/gcc-patches/2019-06/msg01106.html On 6/18/19 9:19 PM, Martin Sebor wrote: On 6/14/19 2:59 PM, Jeff Law wrote: On 6/4/19 1:40 PM, Martin Sebor wrote: On 6/3/19 5:24 PM, Martin Sebor wrote: On 5/31/19 2:46 PM,

Re: [PATCH] True IPA reimplementation of IPA-SRA

Hi, On Thu, Jun 13 2019, Jan Hubicka wrote: > Hi, > i read all changes except for ipa-sra itself. Here are some comments, > I will look at the remaining file next. > > Honza > > > diff --git a/gcc/cgraph.h b/gcc/cgraph.h > index 9a19d83fffb..3f838c08e76 100644 > --- a/gcc/cgraph.h > +++

[PATCH] Add new helper traits for signed/unsigned integer types

Reuse the __is_one_of alias in additional places, and define traits to check for signed/unsigned integer types so we don't have to duplicate those checks elsewhere. The additional overloads for std::byte in were reviewed by LEWG and considered undesirable, so this patch removes them. *

Re: [SVE] [fwprop] PR88833 - Redundant moves for WHILELO-based loops

On Wed, 26 Jun 2019 at 16:05, Richard Sandiford wrote: > > Prathamesh Kulkarni writes: > > On Tue, 25 Jun 2019 at 20:05, Richard Sandiford > > wrote: > >> > >> Prathamesh Kulkarni writes: > >> > On Mon, 24 Jun 2019 at 21:41, Prathamesh Kulkarni > >> > wrote: > >> >> > >> >> On Mon, 24 Jun

Re: [PATCH][gcc] libgccjit: check result_type in gcc_jit_context_new_binary_op

On Tue, 2019-06-25 at 08:11 +, Andrea Corallo wrote: > Hi, > third version for this patch with the simplified test. > > make check-jit pass clean > > Bests > Andrea > > 2019-06-09 Andrea Corallo andrea.cora...@arm.com > > * libgccjit.c (gcc_jit_context_new_binary_op): Check result_type

Re: [PATCH][gcc] libgccjit: add bitfield support

On Wed, 2019-06-26 at 11:07 +, Andrea Corallo wrote: > Hi David, > thanks for the suggestions. > Updated version for the bitfield libgccjit support patch here > addressing comments. > > test-error-gcc_jit_context_new_bitfield-invalid-width.c is reworked > and now assume that the long of the

[PR preprocessor/90927] Fixe dependency output

this patch fixes 90927. The assert triggers when the user uses both -MT and -MQ target options in an unfortunate order. I had thought about this and thought users wouldn't use both. You'd think by now that I've learnt the answer to 'Would a user ever do $X?', is 'Yes, of course they will'.

[PATCH 2/2] rs6000: Fix rs6000_keep_leaf_when_profiled

This function is called from elsewhere, so shouldn't be static. Committing. 2019-06-26 Segher Boessenkool * config/rs6000/rs6000-internal.h (rs6000_keep_leaf_when_profiled): New declaration. * config/rs6000/rs6000-logue.c (rs6000_keep_leaf_when_profiled): Remove

[PATCH 1/2] rs6000: Remove duplicated code

A large portion of the code moved from rs6000.c (to rs6000-logue.c) was accidentally retained. This fixes it. Committing. Segher 2019-06-26 Segher Boessenkool * rs6000.c: Fix previous commit, it missed some changes. --- gcc/config/rs6000/rs6000.c | 1279

Re: [RFA][tree-optimization/90883] Improve DSE to handle redundant calls

On Wed, Jun 26, 2019 at 6:17 AM Jeff Law wrote: > > So based on the conversation in the BZ I cobbled together a patch to > extend tree-ssa-dse.c to also detect redundant stores. > > To be clear, given two stores, the first store is dead if the later > store overwrites all the live bytes set by

Re: [PATCH] Remove quite obvious dead assignments.

Hi, On Wed, Jun 26 2019, Martin Liška wrote: > Hi. > > I've spent some with clang-static-analyzer and I analyzed the warnings > reported. > As always wit analyzers, majority of the issues are false positives, however > it caught > couple of real issues: > >

Re: [PATCH] Remove quite obvious dead assignments.

On Wed, Jun 26, 2019 at 12:57:15PM +0200, Martin Liška wrote: > --- a/gcc/asan.c > +++ b/gcc/asan.c > @@ -1713,8 +1713,8 @@ asan_emit_allocas_unpoison (rtx top, rtx bot, rtx_insn > *before) >rtx ret = init_one_libfunc ("__asan_allocas_unpoison"); >top = convert_memory_address (ptr_mode,

Re: [RFA] Handle _CHK builtins in tree-ssa-dse.c

On Tue, Jun 25, 2019 at 2:20 AM Jeff Law wrote: > > These are some minor improvements to tree-ssa-dse, in particular it adds > handling of the _CHK variants of the supported functions (memcpy, > memmove, memset). It's just something I noticed while poking at 90883. > > These don't trigger during

[PATCH] Fix ICE in PR90982

The following works around an issue in IPA SRA which does analysis twice, once on the original function body and once on the cloned one when replacing refs. If both end up not agreeing we end up with stale unreplaced refs - as in this case where earlier analysis had access to SSA range info

[PATCH] Kill lto_bitmap_alloc/free

This removes another global obstack and adjusts the single user, fixing inconsitent freeing on the way. Bootstrapped/tested on x86_64-unknown-linux-gnu, applied. Richard. 2019-06-26 Richard Biener * lto-streamer.h (lto_bitmap_alloc): Remove. (lto_bitmap_free): Likewise.

Re: [PATCH][gcc] libgccjit: add bitfield support

Hi David, thanks for the suggestions. Updated version for the bitfield libgccjit support patch here addressing comments. test-error-gcc_jit_context_new_bitfield-invalid-width.c is reworked and now assume that the long of the compiler compiling the test is of the same size of the libgccjit long.

[PATCH] Remove quite obvious dead assignments.

Hi. I've spent some with clang-static-analyzer and I analyzed the warnings reported. As always wit analyzers, majority of the issues are false positives, however it caught couple of real issues: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90973

[PATCH] Fix misc stuff seen by clang-static-analyzer.

Hi. This small stuff handles a misc clang-static-analyzer issues. Patch can bootstrap on x86_64-linux-gnu and survives regression tests. Ready to be installed? Thanks, Martin gcc/lto/ChangeLog: 2019-06-26 Martin Liska * lto-dump.c (struct symbol_entry): Add default dtor.

[libsanitizer] Fix sanitizer_common/sanitizer_posix_libcdep.cc compilation on Solaris 11.5

A recent Solaris 11.5 Beta build (st_047) introduced MADV_DONTDUMP, breaking the libsanitizer build. The fix is already upstream https://reviews.llvm.org/D62892 and I've now installed it on mainline and the gcc-9 branch after testing on i386-pc-solaris2.11 and sparc-sun-solaris2.11.

Re: [SVE] [fwprop] PR88833 - Redundant moves for WHILELO-based loops

Prathamesh Kulkarni writes: > On Tue, 25 Jun 2019 at 20:05, Richard Sandiford > wrote: >> >> Prathamesh Kulkarni writes: >> > On Mon, 24 Jun 2019 at 21:41, Prathamesh Kulkarni >> > wrote: >> >> >> >> On Mon, 24 Jun 2019 at 19:51, Richard Sandiford >> >> wrote: >> >> > >> >> > Prathamesh

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On Wed, Jun 26, 2019 at 12:19:28PM +0200, Uros Bizjak wrote: > > The patch isn't correct if TARGET_MMX_WITH_SSE, but not TARGET_AVX, because > > in that case it will push only that 8 and nothing else, while you really > > want to have 16 and 8 in that order, so that it tries to vectorize first > >

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On Wed, Jun 26, 2019 at 10:47 AM Jakub Jelinek wrote: > > On Wed, Jun 26, 2019 at 10:17:26AM +0200, Uros Bizjak wrote: > > Please note that the patch regresses > > > > FAIL: gcc.target/i386/sse2-vect-simd-11.c scan-tree-dump-times vect > > "vectorized [1-3] loops" 2 > > FAIL:

Re: [PATCH] Enable GCC support for AVX512_VP2INTERSECT.

Hi Hongtao, > Index: testsuite/lib/target-supports.exp > === > --- testsuite/lib/target-supports.exp (revision 272667) > +++ testsuite/lib/target-supports.exp (working copy) > @@ -7963,6 +7963,20 @@ > } "-mavx512bw" ] > } > >

Re: [SVE] [fwprop] PR88833 - Redundant moves for WHILELO-based loops

On Tue, 25 Jun 2019 at 20:05, Richard Sandiford wrote: > > Prathamesh Kulkarni writes: > > On Mon, 24 Jun 2019 at 21:41, Prathamesh Kulkarni > > wrote: > >> > >> On Mon, 24 Jun 2019 at 19:51, Richard Sandiford > >> wrote: > >> > > >> > Prathamesh Kulkarni writes: > >> > > @@ -1415,6 +1460,19

Re: [PATCH] Enable GCC support for AVX512_VP2INTERSECT.

On Wed, Jun 26, 2019 at 5:21 PM Martin Liška wrote: > > Hi. > > Started from r272668 I see: > > /tmp/ccqxwVjt.s: Assembler messages: > > /tmp/ccqxwVjt.s:22: Error: no such instruction: `vp2intersectq > .LC1(%rip),%zmm0,%k0' > > /tmp/ccqxwVjt.s:33: Error: no such instruction: `vp2intersectd >

[PING][AArch64] Use scvtf fbits option where appropriate

Ping, plus minor rework (mostly non-functional changes) gcc/ChangeLog: 2019-06-12 Joel Hutton * config/aarch64/aarch64-protos.h (aarch64_fpconst_pow2_recip): New prototype * config/aarch64/aarch64.c (aarch64_fpconst_pow2_recip): New function *

Re: [PATCH] Enable GCC support for AVX512_VP2INTERSECT.

Hi. Started from r272668 I see: /tmp/ccqxwVjt.s: Assembler messages: /tmp/ccqxwVjt.s:22: Error: no such instruction: `vp2intersectq .LC1(%rip),%zmm0,%k0' /tmp/ccqxwVjt.s:33: Error: no such instruction: `vp2intersectd .LC3(%rip),%zmm0,%k0' compiler exited with status 1 FAIL:

Re: [PATCH 05/30] Changes to arm

On 25/06/2019 21:22, acsaw...@linux.ibm.com wrote: > From: Aaron Sawdey > > * config/arm/arm-protos.h: Change movmem to cpymem in names. > * config/arm/arm.c (arm_movmemqi_unaligned, arm_gen_movmemqi, > gen_movmem_ldrd_strd, thumb_expand_movmemqi) Change movmem to cpymem. >

Re: [PATCH 02/30] Changes for aarch64

On 25/06/2019 21:22, acsaw...@linux.ibm.com wrote: > From: Aaron Sawdey > > * config/aarch64/aarch64-protos.h: Change movmem to cpymem. > * config/aarch64/aarch64.c (aarch64_expand_movmem): Change movmem > to cpymem. > * config/aarch64/aarch64.h: Change movmem to cpymem.

Re: [PATCH 01/30] Changes to machine independent code

On 26/06/2019 09:36, Richard Sandiford wrote: > Jeff Law writes: >> On 6/25/19 2:22 PM, acsaw...@linux.ibm.com wrote: >>> From: Aaron Sawdey >>> >>> * builtins.c (get_memory_rtx): Fix comment. >>> * optabs.def (movmem_optab): Change to cpymem_optab. >>> * expr.c

Re: [PATCH] Define midpoint and lerp functions for C++20 (P0811R3)

On 25/06/19 21:55 +0200, Rainer Orth wrote: Hi Jonathan, Doh, I looked in and saw that we get std::abs(double) from the Solaris headers, and then forgot and used it anyway. I'll replace that right away, thanks. Should be fixed at r272653. Tested x86_64-linux, committed to trunk. it did

Re: [PATCH] Move rust_{is_mangled,demangle_sym} to a private libiberty header.

Bootstrapped and tested on x86_64-unknown-linux-gnu. (Apologies for the delay, while I was able to run libiberty tests back when I submitted the patch, I wanted to make sure I can run the whole GCC testsuite, especially for more significant future contributions, so I had to wait until I had

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On Wed, Jun 26, 2019 at 10:17:26AM +0200, Uros Bizjak wrote: > Please note that the patch regresses > > FAIL: gcc.target/i386/sse2-vect-simd-11.c scan-tree-dump-times vect > "vectorized [1-3] loops" 2 > FAIL: gcc.target/i386/sse2-vect-simd-15.c scan-tree-dump-times vect > "vectorized [1-3] loops"

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On Wed, Jun 26, 2019 at 10:36 AM Richard Biener wrote: > > On June 26, 2019 10:25:44 AM GMT+02:00, Uros Bizjak wrote: > >On Wed, Jun 26, 2019 at 10:17 AM Uros Bizjak wrote: > >> > >> Now that TARGET_MMX_WITH_SSE is implemented, the compiler should be > >> able to auto-vectorize: > > > >On a

Re: [PATCH 0/3] RFC: Let debug stmts influence codegen@-Og

On 23/06/19 14:51 +0100, Richard Sandiford wrote: Also, the new mode is mostly orthogonal to the optimisation level (although it would in effect disable optimisations like loop vectorisation, until we have a way of representing debug info for vectorised loops). The third patch therefore adds an

Re: [PATCH 01/30] Changes to machine independent code

Jeff Law writes: > On 6/25/19 2:22 PM, acsaw...@linux.ibm.com wrote: >> From: Aaron Sawdey >> >> * builtins.c (get_memory_rtx): Fix comment. >> * optabs.def (movmem_optab): Change to cpymem_optab. >> * expr.c (emit_block_move_via_cpymem): Change movmem to cpymem. >>

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On June 26, 2019 10:25:44 AM GMT+02:00, Uros Bizjak wrote: >On Wed, Jun 26, 2019 at 10:17 AM Uros Bizjak wrote: >> >> Now that TARGET_MMX_WITH_SSE is implemented, the compiler should be >> able to auto-vectorize: > >On a related note, following slightly changed testcase: > >void >foo (char

Re: [PATCH] Move rust_{is_mangled,demangle_sym} to a private libiberty header.

Hi Mark, Valgrind is definitely on my upstreaming list, alongside GDB, LLDB and Linux perf. You can see the preliminary version here: https://gist.github.com/eddyb/c41a69378750a433767cf53fe2316768 (do not use it yet, I still want to tweak it a bit more before upstreaming it, soon, and I want

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On June 26, 2019 10:17:26 AM GMT+02:00, Uros Bizjak wrote: >Now that TARGET_MMX_WITH_SSE is implemented, the compiler should be >able to auto-vectorize: > >void >foo (char *restrict r, char *restrict a) >{ > for (int i = 0; i < 8; i++) >r[i] += a[i]; >} > >Attached patch enables the

Re: [RFC PATCH, i386]: Autovectorize 8-byte vectors

On Wed, Jun 26, 2019 at 10:17 AM Uros Bizjak wrote: > > Now that TARGET_MMX_WITH_SSE is implemented, the compiler should be > able to auto-vectorize: On a related note, following slightly changed testcase: void foo (char *restrict r, char *restrict a) { for (int i = 0; i < 24; i++) r[i]

Re: [PATCH] Fix AVX512* wrong-code due to *_vinsert_0 (PR target/90991)

On Wed, Jun 26, 2019 at 10:11 AM Jakub Jelinek wrote: > > Hi! > > The following testcase is miscompiled starting with my PR85480 change. > While it is perfectly fine to use "xm" or "vm" constraints for the source > operand when the other operand is "C", we rely on the AVX/AVX512 behavior > that

[RFC PATCH, i386]: Autovectorize 8-byte vectors

Now that TARGET_MMX_WITH_SSE is implemented, the compiler should be able to auto-vectorize: void foo (char *restrict r, char *restrict a) { for (int i = 0; i < 8; i++) r[i] += a[i]; } Attached patch enables the conversion and produces: foo: movq(%rdi), %xmm1 movq

[PATCH] Fix AVX512* wrong-code due to *_vinsert_0 (PR target/90991)

Hi! The following testcase is miscompiled starting with my PR85480 change. While it is perfectly fine to use "xm" or "vm" constraints for the source operand when the other operand is "C", we rely on the AVX/AVX512 behavior that most 128-bit or 256-bit vector instructions clear the upper bits of

Re: [PATCH] RISC-V: Add -malign-data= option.

On Tue, 25 Jun 2019 15:58:53 PDT (-0700), pins...@gmail.com wrote: On Tue, Jun 25, 2019 at 3:46 PM Ilia Diachkov wrote: Hello, This patch adds new machine specific option -malign-data={word,abi} to RISC-V port. The option switches alignment of global variables and constants of

Re: [PATCH] Add to same comdate group only if set (PR middle-end/90899)

PING^1 On 6/18/19 10:58 AM, Martin Liška wrote: > Hi. > > The patch is quite obvious, it copies the same what we do in > another IPA passes. > > Patch can bootstrap on x86_64-linux-gnu and survives regression tests. > > Ready to be installed? > Thanks, > Martin > > gcc/ChangeLog: > >

Re: [PATCH] Fix PR90914

Hi. Just for the record, the patch is responsible for a significant debug info growth for 434.zeusmp with -O2 -flto and other options: https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=227.75.4 Thanks, Martin

Re: [PATCH 12/30] Changes to i386

On Tue, Jun 25, 2019 at 10:27 PM wrote: > > From: Aaron Sawdey > > * config/i386/i386-expand.c (expand_set_or_movmem_via_loop, > expand_set_or_movmem_via_rep, expand_movmem_epilogue, > expand_setmem_epilogue_via_loop, expand_set_or_cpymem_prologue, >

[PATCH] doc: Fix opindex for -W options

@opindex -Wxxx is wrong; it should be @opindex Wxxx. Committing as trivial and obvious. 2019-06-26 Segher Boessenkool * doc/invoke.texi (Warning Options): Fix some @opindex syntax. --- gcc/doc/invoke.texi | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git

[PATCH] Fix warnings seen by clang in gcc/symbol-summary.h.

Hi. The patch is about missing argument to function call and unused arguments in symbol-summary.h. Patch can bootstrap on x86_64-linux-gnu and survives regression tests. Ready to be installed? Thanks, Martin gcc/ChangeLog: 2019-06-25 Martin Liska * symbol-summary.h (traverse):