Hello!
2011-08-02 Uros Bizjak ubiz...@gmail.com
PR target/47766
* doc/md.texi (stack_protect_set): The pattern moves ptr_mode value.
(stack_protect_test): The pattern compares ptr_mode value.
Tested on x86_64-pc-linux-gnu, committed to mainline (and soon release
Hello!
We should not call CPU_COUNT when not defined in glibc.
2011-08-03 Uros Bizjak ubiz...@gmail.com
* config/linux/proc.c (gomp_cpuset_popcount): Rename from
cpuset_popcount. No more static.
(gomp_init_num_threads): Update call to cpuset_popcount
On Wed, Aug 3, 2011 at 11:11 AM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Aug 03, 2011 at 10:36:47AM +0200, Uros Bizjak wrote:
We should not call CPU_COUNT when not defined in glibc.
Oops, forgot that old glibcs don't have it.
I don't like the prototype in libgomp.h, for one sched.h
On Fri, Aug 5, 2011 at 4:45 AM, H.J. Lu hongjiu...@intel.com wrote:
This patch adds -march=core-avx-i to support Intel Ivy Bridge. OK
for trunk and 4.6?
2011-08-04 H.J. Lu hongjiu...@intel.com
* config/i386/i386.c (processor_alias_table): Add core-avx-i.
*
there
is a problem in IRA/reload how offsetable operand addresses are
handled (I will post separate RFC patch for that).
2011-08-05 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (*pushmode2): Use o constraint instead
of m for operand 0. Add type and mode attribute
wrong in reload?
2011-08-05 Uros Bizjak ubiz...@gmail.com
PR target/49781
* config/i386/i386.c (ix86_decompose_address): Allow zero-extended
SImode addresses.
(ix86_print_operand_address): Handle zero-extended addresses.
(memory_address_length): Add length
../../../gcc-svn/trunk/libiberty/testsuite/test-expandargv.c: In
function ‘writeout_test’:
../../../gcc-svn/trunk/libiberty/testsuite/test-expandargv.c:211:
warning: ignoring return value of ‘fwrite’, declared with attribute
warn_unused_result
Attached patch fixes this warning.
2011-08-05 Uros
On Mon, May 28, 2012 at 3:37 PM, Marc Glisse marc.gli...@inria.fr wrote:
Ping? The rest of the patch has been approved already.
On Thu, 10 May 2012, Marc Glisse wrote:
Hello,
could an i386 maintainer take a look at the following testcase?
gcc/testsuite/ChangeLog
2012-05-08 Marc Glisse
Hello!
Ping?
Please at least add and URL to the patch, it took me some time to
found the latest version [1], I'm not even sure if it is the latest
version...
I assume that you cleared all issues with middle-end and scheduler
maintainers, it is not clear from the message.
+ (1) IMUL
On Sat, Jun 2, 2012 at 3:27 PM, H.J. Lu hongjiu...@intel.com wrote:
ix86_sched_reorder should skip debug insns. Tested on Linux/ia32.
OK to install?
2012-06-02 H.J. Lu hongjiu...@intel.com
PR bootstrap/53555
* config/i386/i386.c (ix86_sched_reorder) Skip debug insns.
OK.
On Mon, Jun 4, 2012 at 8:09 PM, H.J. Lu hongjiu...@intel.com wrote:
We should enable x32 run-time library if --with-abi={x32|mx32} is used
to configure GCC i[34567]86-*-* and x86_64-*-*. Tested on Linux/x86-64.
OK for trunk?
2012-06-04 H.J. Lu hongjiu...@intel.com
PR target/53575
On Tue, Jun 5, 2012 at 2:47 PM, H.J. Lu hjl.to...@gmail.com wrote:
We should enable x32 run-time library if --with-abi={x32|mx32} is used
to configure GCC i[34567]86-*-* and x86_64-*-*. Tested on Linux/x86-64.
Why all three ABIs here? Didn't user specify -with-abi=mx32 only, so
Hello!
2012-06-06 Uros Bizjak ubiz...@gmail.com
* gcc.target/i386/avx-os-support.h (XCR_XFEATURE_ENABLED_MASK): New.
(XSTATE_FP): Ditto.
(XSTATE_SSE): Ditto.
(XSTATE_YMM): Ditto.
(avx_os_support): Use new defines.
Tested on x86_64-pc-linux-gnu AVX
instructions for 32bit targets. There is no other
functional changes, although the change to real SSE FP ops is
tempting. ;)
2012-06-07 Uros Bizjak ubiz...@gmail.com
* config/i386/32/sfp-machine.h (__gcc_CMPtype, CMPtype,
_FP_KEEPNANFRACP, _FP_CHOOSENAN, FP_EX_INVALID, FP_EX_DENORM
On Thu, Jun 7, 2012 at 9:58 PM, Uros Bizjak ubiz...@gmail.com wrote:
Attached patch rewrites x86 soft-FP as an out-of-line function that
gets conditionally called when exception occurs. This rewrite removes
17 instances of the same code from libgcc. In addition, the patch
unifies a lot
Hello!
Similar to x86, this patch puts soft-FP exception handler out-of-line.
The patch also cleans asm constraints a bit (introduces +), but has
no other functional changes.
2012-06-12 Uros Bizjak ubiz...@gmail.com
* config/ia64/sfp-machine.h (__sfp_handle_exceptions): New
On Tue, Jun 12, 2012 at 12:31 PM, Uros Bizjak ubiz...@gmail.com wrote:
Similar to x86, this patch puts soft-FP exception handler out-of-line.
The patch also cleans asm constraints a bit (introduces +), but has
no other functional changes.
2012-06-12 Uros Bizjak ubiz...@gmail.com
Hello!
A couple of #defines can be moved to shared header, no need to mask
_fex with exception mask and fnstsw should be marked volatile, since
it depends on hidden FP status register.
2012-06-13 Uros Bizjak ubiz...@gmail.com
* config/i386/32/sfp-machine.h (_FP_NANSIGN_S
Hello!
This patch fixes a trivial oversight in the name of or family of
sync functions.
2012-06-12 Uros Bizjak ubiz...@gmail.com
* config/ia64/sync.md (fetchop_name): Change ior attribute to or.
Tested on ia64-unknown-linux-gnu. Is it still OK to squeeze this
trivial one-liner
On Wed, Jun 13, 2012 at 4:47 PM, H.J. Lu hongjiu...@intel.com wrote:
On i386, ix86_size_cost will be used for -Os, which has zero for
simultaneous_prefetches, prefetch_block, l1_cache_size and l2_cache_size.
This patch adds ix86_tune_cost and uses it for simultaneous_prefetches,
wait with it until 4.7.1 is released.
+2012-06-07 Venkataramanan Kumar venkataramanan.ku...@amd.com
+
The common way is to use:
Backport from mainline
2012-05-09 Uros Bizjak ubiz...@gmail.com
instead of the following line:
+ Backport from 2012-05-09 mainline r187354
Hello!
movd to/from MMX or SSE registers is SSE2 instruction. Also, remove
wrong x,x alternative.
2012-06-14 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (*zero_extendsidi2): Mark movd alternatives
SSE2 only. Remove x,x alternative.
(*zero_extendsidi2_rex64
On Thu, Jun 14, 2012 at 8:47 PM, Uros Bizjak ubiz...@gmail.com wrote:
movd to/from MMX or SSE registers is SSE2 instruction. Also, remove
wrong x,x alternative.
2012-06-14 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (*zero_extendsidi2): Mark movd alternatives
SSE2
Hello!
The testcase still fails on x86_64-pc-linux-gnu with:
FAIL: gcc.dg/tree-ssa/vrp68.c scan-tree-dump-times vrp1 link_error 1
since there are two calls to link_error.
2012-06-17 Uros Bizjak ubiz...@gmail.com
* gcc.dg/tree-ssa/vrp68.c: Fix scan-tree-dump-times argument order
Hello!
2012-06-17 Uros Bizjak ubiz...@gmail.com
* config/i386/sse.md (vcvtph2ps): Fix vec_select selector.
Tested on x86_64-pc-linux-gnu {,-m32}, committed to mainline SVN, will
be backported to release branches.
Uros.
Index: config/i386/sse.md
Hello!
2012-06-17 Uros Bizjak ubiz...@gmail.com
* config/i386/sfp-machine.h (FP_HANDLE_EXCEPTIONS): Use
__builtin_expect when checking for exceptions.
* config/ia64/sfp-machine.h (FP_HANDLE_EXCEPTIONS): Ditto.
Tested on x86_64-pc-linux-gnu {,-m32} and ia64-unknown
Hello!
Please note that you will probably hit PR33329, this is the reason
that we expand multiplications after reload. Please see [1] for
further explanation. There is gcc.target/i386/pr33329.c test to cover
this issue, but it is not effective anymore since the simplification
happens at tree
On Sun, Jun 17, 2012 at 8:37 PM, Uros Bizjak ubiz...@gmail.com wrote:
Hello!
Please note that you will probably hit PR33329, this is the reason
that we expand multiplications after reload. Please see [1] for
further explanation. There is gcc.target/i386/pr33329.c test to cover
this issue
Hello!
gcc.target/i386/pr33329.c is fully optimized with tree optimizers to a
constant. Attached patch increases array size to avoid
over-optimization and to perform intended RTL optimization check.
2012-06-17 Uros Bizjak ubiz...@gmail.com
* gcc.target/i386/pr33329.c (f): Increase
Hello!
Attached patch teaches gcc how to merge unaligned load (UNSPEC_MOVU)
with pcmpistr/pcmpestr instructions.
2012-06-18 Uros Bizjak ubiz...@gmail.com
PR target/53712
* config/i386/sse.md (*sse4_2_pcmpestr_unaligned): New.
(*sse4_2_pcmpistr_unaligned): New
Hello!
Following the patch that allows unaligned operands in pcmpestri [1],
we can substitute x86 asm in lex.c with equivalent builtin functions.
2011-06-18 Uros Bizjak ubiz...@gmail.com
* lex.c (search_line_sse42): Use __builtin_ia32_loaddqu and
__builtin_ia32_pcmpestri128
On Tue, Jun 19, 2012 at 12:07 AM, Richard Henderson r...@redhat.com wrote:
On 2012-06-18 13:19, Uros Bizjak wrote:
/* ??? The builtin doesn't understand that the PCMPESTRI read from
memory need not be aligned. */
- __asm (%vpcmpestri $0, (%1), %2
- : =c(index) : r
On Tue, Jun 19, 2012 at 8:38 AM, Uros Bizjak ubiz...@gmail.com wrote:
On Tue, Jun 19, 2012 at 12:07 AM, Richard Henderson r...@redhat.com wrote:
On 2012-06-18 13:19, Uros Bizjak wrote:
/* ??? The builtin doesn't understand that the PCMPESTRI read from
memory need not be aligned
Hello!
2012-06-19 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (FRNDINT_ROUNDING): New int iterator.
(rounding): New int attribute.
(ROUNDING): Ditto.
(frndintxf2_rounding): Macroize insn from
frndintxf2_{floor,ceil,trunc} using FRNDINT_ROUNDING
Hello!
2012-06-19 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (FIST_ROUNDING): New int iterator.
(rounding): Handle UNSPEC_FIST_{FLOOR,CEIL}.
(ROUNDING): Ditto.
(*fistmode2_rounding_1): Macroize insn from
*fistmode2_{floor,ceil}_1 using
On Mon, Jun 18, 2012 at 10:06 PM, Richard Henderson r...@redhat.com wrote:
Please note that you will probably hit PR33329, this is the reason
that we expand multiplications after reload. Please see [1] for
further explanation. There is gcc.target/i386/pr33329.c test to cover
this issue, but
Hello!
2012-06-20 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (rounding_insn): New int attribute.
(rounding_insnxf2): Macroize insn from
{floor,ceil,btrunc}xf2 using FRNDINT_ROUNDING int iterator.
(lrounding_insnxfmode2): Rename from lroundingxfmode2
Hello!
2012-06-20 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (rounding_insnmode2): Macroize expander
from {floor,ceil,btrunc}mode2 using FIST_ROUNDING int iterator.
(lrounding_insnMODEF:modeSWI48:mode2): Macroize expander
from l{floor,ceil
Hello!
2012-06-20 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (SINCOS): New int iterator.
(sincos): New int attribute.
(*sincosxf2_i387): Macroize insn from *{sin,cos}xf2_i387 using
SINCOS int iterator.
(*sincos_extendmodexf2_i387): Macroize insn
Hello!
During htab_delete (dropped_values), loc_exp_dep_pool
allocated objects might be accessed, so it is better to free the
pool afterwards.
Bootstrapped/regtested on i686-linux, ok for trunk?
Looks obvious.
The patch doesn't fix all writes to freed up memory, please see
comment #8
Hello!
The GNU assembler now (just as of today) accepts 'rep bsf ...' or 'rep bsr
...'
syntax. It's always better to put a prefix on the instruction itself
rather than to write 'rep; ...'. This changes 'rep; bsf ...' to 'rep bsf ...'
when the assembler accepts the latter.
2012-06-22
On Sat, Jun 23, 2012 at 12:00 AM, Roland McGrath mcgra...@google.com wrote:
Here is an alternative patch that just changes the configure test
controlling %; so it will elide the ; only for an assembler that
also accepts 'rep bsf', 'rep bsr', and 'rep ret', and just uses
%; for these cases too.
Hello!
2012-07-01 Uros Bizjak ubiz...@gmail.com
* c-pch.c (c_common_write_pch): Remove unused variables.
Tested on x86_64-pc-linux-gnu, committed to mainline SVN.
Uros.
Index: c-pch.c
===
--- c-pch.c (revision 189099
Hello!
gcc/
2012-07-01 Uros Bizjak ubiz...@gmail.com
* config/i386/xmmintrin.h (_mm_sfence): Use __builtin_ia32_pause.
libgomp/
2012-07-01 Uros Bizjak ubiz...@gmail.com
* config/linux/x86/futex.h (cpu_relax): Use __builtin_ia32_pause.
* testsuite/libgomp.c/sort-1
On Sun, Jul 1, 2012 at 3:10 PM, Uros Bizjak ubiz...@gmail.com wrote:
gcc/
2012-07-01 Uros Bizjak ubiz...@gmail.com
* config/i386/xmmintrin.h (_mm_sfence): Use __builtin_ia32_pause.
libgomp/
2012-07-01 Uros Bizjak ubiz...@gmail.com
* config/linux/x86/futex.h
On Tue, Jul 3, 2012 at 12:37 PM, Richard Guenther
richard.guent...@gmail.com wrote:
Based on the observation above, the patch is OK for mainline, but
please also handle rep nop case.
Here's the new version of the patch that does that. Note that someone
needs to commit this for me, since I
Hello!
Attached patch fixes PR 53811. gcc failed to check if symbol_ref
operands fits sibcall_insn_operand predicate. In case of
-mcmodel=large, we must emit indirect jump for a sibcall.
2012-07-03 Uros Bizjak ubiz...@gmail.com
PR target/53811
* config/i386/i386.c
Hello!
2012-07-03 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.c (ix86_option_override_internal): Fix wrong
code model string in the error message.
Committed.
Uros.
Index: config/i386/i386.c
===
--- config
Hello!
Several otherwise harmless may be used uninitialized warnings break
LTO profiled bootstrap. Attached patch fixes all these places,
enabling profiled LTO bootstrap to continue.
2012-07-04 Uros Bizjak ubiz...@gmail.com
* expmed.c (expand_mult): Initialize coeff and is_neg.
java
have to update
something, at least when profiling, before calling
cgraph_propagate_frequency.
2012-07-04 Uros Bizjak ubiz...@gmail.com
* ipa.c (symtab_remove_unreachable_nodes): Partially revert r187375
to not call cgraph_propagate_frequency if something was changed.
testsuite
Hello!
2012-07-06 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.c (construct_plt_address): Make static.
* config/i386/i386-protos.h (construct_plt_address): Remove.
Tested on x86_64-pc-linux-gnu, committed to mainline SVN.
Uros.
Index: i386-protos.h
On Fri, Jul 6, 2012 at 5:24 PM, Andi Kleen a...@firstfloor.org wrote:
Uros Bizjak ubiz...@gmail.com writes:
Sometimes, gcc generates:
leaq(%rbx,%rax), %rax
that is in fact equivalent (modulo flags reg clobber) to:
addq%rbx, %rax
Attached patch adds additional peephole2 patterns
Hello!
TARGET_BRANCH_PREDICTION_HINTS isn't used at all. This patch removes it.
Bootstrappedtested (incl. -m32) on x86_64-unknown-linux-gnu. OK for trunk?
This infrastructure can be used for future targets, so let's leave it as is.
Uros.
On Tue, Jul 10, 2012 at 9:04 AM, Steven Bosscher stevenb@gmail.com wrote:
TARGET_BRANCH_PREDICTION_HINTS isn't used at all. This patch removes it.
Bootstrappedtested (incl. -m32) on x86_64-unknown-linux-gnu. OK for trunk?
This infrastructure can be used for future targets, so let's leave
Hello!
A test is added. Is it ok?
gcc/testsuite/ChangeLog
2012-07-08 Dehao Chen de...@google.com
* gcc.dg/debug_info_inline.c: New test.
This test fails on alphaev68-unknown-linux-gnu [1] and other non-x86
targets [2] too:
FAIL: gcc.dg/debug_info_inline.c scan-assembler loc 1
Hello!
2012-07-12 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (QImode and HImode cmove splitters): Merge
QImode and HImode splitters. Use ix86_comparison_operator.
Explicitly match FLAGS_REG.
(DFmode cmove splitter): Explicitly match FLAGS_REG.
Tested
Hello!
A patch that's in my local tree for quite some time. This removes
redundant tests.
LTO bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2012-07-13 Richard Guenther rguent...@suse.de
* gimple.c (gimple_types_compatible_p_1): Remove redundant
Hello!
Recently, gcc was switched to enable TFmode moves in XMM registers
also for TARGET_SSE. Sync reg-reg move cost calculation.
2012-07-14 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.c (ix86_set_reg_reg_cost): Enable TFmode/TCmode
for TARGET_SSE.
Tested on x86_64-pc
On Mon, Jul 16, 2012 at 4:45 PM, Jakub Jelinek ja...@redhat.com wrote:
While backporting, I've noticed there is an extra semicolon that shouldn't
be there in -O0 version of _xabort. Ok for trunk?
2012-07-16 Jakub Jelinek ja...@redhat.com
* config/i386/rtmintrin.h (_xabort):
On Mon, Jul 16, 2012 at 4:32 PM, H.J. Lu hjl.to...@gmail.com wrote:
On Mon, Jul 16, 2012 at 7:14 AM, Kirill Yukhin kirill.yuk...@gmail.com
wrote:
Hello guys,
Here is a tiny patch which adds two new intrinsics which were
introduced in recent spec [1].
They're aliased to the existing
On Wed, Jul 18, 2012 at 7:34 PM, Andrew Haley a...@redhat.com wrote:
On 07/18/2012 05:30 PM, H.J. Lu wrote:
2012-07-16 H.J. Lu hongjiu...@intel.com
PR libjava/53973
* include/x86_64-signal.h (CHECK_67H_PREFIX): New.
(HANDLE_DIVIDE_OVERFLOW): Check and and skip 67h address
already tripped on this issue.
2012-07-18 Uros Bizjak ubiz...@gmail.com
* doc/tm.texi.in (MODE_AFTER): Add entity as the first macro argument.
* doc/tm.texi: Regenerate.
* mode-switching.c (optimize_mode_switching): Update MODE_AFTER call.
* config/sh/sh.h
On Wed, Jul 18, 2012 at 6:27 PM, H.J. Lu hongjiu...@intel.com wrote:
This patch defines FFI_SIZEOF_JAVA_RAW to 4 for x32, similar to MIPS n32.
It fixed:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53982
Here is the patch with updated ChangeLog entry. X32 has the same issue
as MIPS n32,
On Wed, Jul 18, 2012 at 9:10 PM, H.J. Lu hjl.to...@gmail.com wrote:
This patch defines FFI_SIZEOF_JAVA_RAW to 4 for x32, similar to MIPS n32.
It fixed:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53982
Here is the patch with updated ChangeLog entry. X32 has the same issue
as MIPS n32,
Hello!
This patch adds new intrinsics for new ADCX, ADOX, RDSEED and
PREFETCHW instructions, introduced here:
http://software.intel.com/en-us/avx/
+/* Copyright (C) 2012 Free Software Foundation, Inc.
+ Contributed by Michael Zolotukhin.
Please remove Contributed by lines from headers.
+
On Thu, Jul 19, 2012 at 11:10 AM, Jakub Jelinek ja...@redhat.com wrote:
On Thu, Jul 19, 2012 at 11:03:25AM +0200, Uros Bizjak wrote:
+ /* Set CF to specified value. */
+ emit_insn (gen_addqi3_cc(
+ gen_reg_rtx(QImode),
+ op1,
+ constm1_rtx
On Thu, Jul 19, 2012 at 11:51 AM, Jakub Jelinek ja...@redhat.com wrote:
You have a mnemonic clash here. prefetchw is not good name for a new
instruction, it clashes with existing 3dnow name. Intel will need to
fix the spec, you probably won't be able to change prefetchw encoding
in
On Thu, Jul 19, 2012 at 11:51 AM, Jakub Jelinek ja...@redhat.com wrote:
I think it is fine to change existing mm3dnow.h header, but not in the way
it has been changed in the patch. The thing is that unlike the newly added
intrinsics headers, mm3dnow.h is still publicly includable header, and
On Wed, Jul 18, 2012 at 8:31 PM, Uros Bizjak ubiz...@gmail.com wrote:
As with all other mode switching macros, we need to pass entity index
also to MODE_AFTER macro. In a multi-entity mode switching case, we
usually don't have same modes for all entities, and we should be able
to return
On Fri, Oct 28, 2011 at 10:57 AM, Jakub Jelinek ja...@redhat.com wrote:
On Thu, Oct 27, 2011 at 10:07:13PM +0200, Uros Bizjak wrote:
Please use expressive RTX forms for expanders, similar to the above
define_insn RTX. You can avoid calling gen_avx2_lshrvmode at the end
of c code. Also
Hello!
We can extend existing code attributes. Also, the patch includes some
stylistic changes in XOP shift patterns.
No functional changes.
2011-10-28 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (shift_insn): Rename code attribute from
shiftrt_insn. Also handle
On Fri, Oct 28, 2011 at 8:21 PM, Uros Bizjak ubiz...@gmail.com wrote:
Hello!
We can extend existing code attributes. Also, the patch includes some
stylistic changes in XOP shift patterns.
No functional changes.
Eh, the patch was the old one, added is additional patch with updated ChangeLog
Hello!
2011-10-29 Uros Bizjak ubiz...@gmail.com
* gcc.target/i386/fma-check.h (main): Use return 0 instead of exit (0).
* gcc.target/i386/fma4-check.h (main): Ditto.
* gcc.target/i386/xop-check.h (main): Ditto.
Committed as trivial change to mainline SVN.
Uros.
Index
Hello!
lshlv16qi3 is not a generic name for expander, and we have ashlv16qi3
for this. Attached patch adds lshrv16qi3 to generate logical
shift-right XOP instruction.
2011-10-29 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (lshlv16qi3): Remove expander.
(lshrv16qi3
Hello!
These pattern names are misleading, implying that these are logical
shift left and arithmetic shift left. They are not, they are shift
logical and shift arithmetic. Attached (trivial) patch renames
these patterns to the insn mnemonic they generate.
2011-10-29 Uros Bizjak ubiz
Hello!
2011-10-30 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (avx2_vec_dupmode): Macroize insn from
avx2_vec_dup{v8sf,v4sf} using VF1 mode iterator.
(vec_dupv4sf): Remove expander.
(vec_dupv4sf): Merge from *vec_dupv4sf and *vec_dupv4sf_avx
On Thu, Oct 27, 2011 at 6:42 PM, Uros Bizjak ubiz...@gmail.com wrote:
This patch updates the Go library to the most recent weekly release. I
think the only potential portability issues here are the use of the
ipv6_mreq struct. I'm not entirely sure the new exp/terminal package is
portable
On Tue, Nov 1, 2011 at 10:07 AM, Jakub Jelinek ja...@redhat.com wrote:
Similarly to the V{4,8}SFmode - unsigned V{4,8}SImode conversion
support for AVX this one adds V{2,4}DFmode - unsigned V{4,8}SImode
conversion.
Ok for trunk?
Please put expander function into i386.c. IMO, this expander
Hello!
On Mon, Oct 31, 2011 at 05:23:58PM -0500, Quentin Neill wrote:
Interested parties should view these threads from three years ago:
http://gcc.gnu.org/ml/gcc-patches/2008-11/threads.html#00145
http://gcc.gnu.org/ml/gcc-patches/2008-12/threads.html#00174
Testing on x86_64, okay to
On Tue, Nov 1, 2011 at 2:35 PM, Jakub Jelinek ja...@redhat.com wrote:
Similarly to the V{4,8}SFmode - unsigned V{4,8}SImode conversion
support for AVX this one adds V{2,4}DFmode - unsigned V{4,8}SImode
conversion.
Ok for trunk?
Please put expander function into i386.c. IMO, this
Hello!
Fix a typo.
2011-10-30 Uros Bizjak ubiz...@gmail.com
PR target/50940
* config/i386/i386.md (floatsimode2_vector_sse_with_temp splitter):
Compare ssevecmodemode with V4SFmode, not V4SImode.
Tested on x86_64-pc-linux-gnu, committed to mainline SVN.
Uros.
Index
Hello!
We have a nice utility function that can be used in int-float
splitter constraints.
2011-11-01 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (splitters for int-float conversion): Use
reg_or_subregno in splitter constraints.
Bootstrapped and regression tested
On Tue, Nov 1, 2011 at 11:00 PM, Jakub Jelinek ja...@redhat.com wrote:
On Tue, Nov 01, 2011 at 10:33:07PM +0100, Uros Bizjak wrote:
We have a nice utility function that can be used in int-float
splitter constraints.
2011-11-01 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md
On Wed, Nov 2, 2011 at 6:06 AM, Ian Lance Taylor i...@google.com wrote:
The problem with your proposal is that the output would be invalid Go,
because it would attempt to define the name _aa twice. However, it does
seem plausible that in most scenarios of this type it would be more
useful
Hello!
2011-11-02 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.c (bdesc_args) [IX86_BUILTIN_CVTTPD2DQ256]: Use
CODE_FOR_fix_truncv4dfv4si2, not CODE_FOR_fix_truncv4sfv4si2.
Tested on x86_64-pc-linux-gnu, committed to mainline SVN.
Uros.
Index: i386.c
Hello!
2011-11-02 Uros Bizjak ubiz...@gmail.com
* gcc.target/i386/avx-cvt-2.c (dg-options): Add -mtune=generic.
* gcc.target/i386/avx2-cvt-2.c (dg-options): Ditto.
* gcc.target/i386/sse2-cvt-2.c (dg-options): Ditto.
* gcc.target/i386/vectorize4-avx.c (dg-final
Hello!
Currently gcc will promote from QI/HI mode to SI mode various
operators but not minus. It will however promote a neg followed by an
add (add is in the current list of promotable operators and neg is
promoted around config/i386/i386.md:16904). This omission can cause
RAT stalls in
On Wed, Nov 2, 2011 at 5:50 PM, Ian Lance Taylor i...@google.com wrote:
#defines with arguments are not working at all. Please consider
following testcase:
You're right: this approach doesn't work for preprocessor macros with
arguments. Making those work via this approach would be much much
Hello!
Here is a fix for PR 50962. Fix improves AGU stall avoidance
optimization by adding opportunity to emit lea instead of mov when it
is profitable.
2011-11-03 Enkovich Ilya ilya.enkov...@intel.com
PR target/50962
* config/i386/i386-protos.h (ix86_use_lea_for_mov): New.
Hello!
No functional change.
2011-11-03 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md: Use {} for multi-line preparation statements.
Bootstrapped and regression tested on x86_64-pc-linux-gnu, committed
to mainline SVN.
Uros.
Index: i386.md
Hello!
For some reason, several expanders use operandN internal variable
instead of operands[N] from the array. Attached patch fixes this
oversight.
2011-11-03 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (lroundX87MODEF:modeSWI248x:mode2,
rintmode2, floormode2
On Thu, Nov 3, 2011 at 6:54 PM, Jakub Jelinek ja...@redhat.com wrote:
As mentioned in the last mail, the testcase in that patch
shows a bug in the unsigned int - float vectorization
on i?86/x86_64. E.g. 0x808fU converted to float
by scalar code is 0x1.02p+31, but by vector code
is
Hello!
Currently slim RTL dump prints STRICT_LOW_PART as:
48 strict_low_part=flags:CCZ!=0
Attached patch fixes these dumps to print:
48 strict_low_part(ax:QI)=flags:CCZ!=0
2011-11-03 Uros Bizjak ubiz...@gmail.com
* sched-vis.c (print_value): Handle STRICT_LOW_PART.
Tested
Hello!
Attached patch fixes all remaining i386 testsuite failures. The header
did say that we can include it also through immintrin.h ...
2011-11-07 Uros Bizjak ubiz...@gmail.com
* config/i386/f16cintrin: Remove extra _X86INTRIN_H_INCLUDED check.
Tested on x86_64-pc-linux-gnu {,-m32
/atomic-other-int128.c -O2 -g (test for
excess errors)
FAIL: gcc.dg/simulate-thread/atomic-other-int128.c -O3 -g (test for
excess errors)
FAIL: gcc.dg/simulate-thread/atomic-other-int128.c -Os -g (test for
excess errors)
2011-11-07 Uros Bizjak ubiz...@gmail.com
* lib/target
Hello!
2011-11-07 Uros Bizjak ubiz...@gmail.com
* gcc.target/i386/pr49781-1.c (dg-options): Add -mtune=generic.
Tested on x86_64-pc-linux-gnu, committed to mainline SVN.
Uros.
Index: gcc.target/i386/pr49781-1.c
On Fri, Nov 4, 2011 at 2:22 PM, Jakub Jelinek ja...@redhat.com wrote:
On Fri, Nov 04, 2011 at 12:21:49PM +0100, Richard Guenther wrote:
Ok. I guess it's ok to use builtins for now - I didn't think of
the memory reference issue ;)
Based on IRC discussion I'm posting an updated patch (both
On Mon, Nov 7, 2011 at 7:38 PM, Jakub Jelinek ja...@redhat.com wrote:
BUILT_IN_LRINT has been vectorized just using 16-byte
vectors, the following patch cures it (of course, for -m64
it unfortunately can't be vectorized, as long there is DImode
rather than SImode).
Hm...
Looking at
Hello!
Attached patch fixes omission from (int) rint () conversion to
__builtin_irint (). We can vectorize this function when out_mode is
SImode, no matter how the builtin is called. Throw in also
BUILT_IN_LLRINT, just for completion ...
2011-11-07 Uros Bizjak ubiz...@gmail.com
On Mon, Nov 7, 2011 at 10:20 PM, Jakub Jelinek ja...@redhat.com wrote:
I think it is at least more readable and perhaps for some CPUs could
be faster (for SandyBridge it is the same speed) if we emit a more
specialized insn over a more generic one.
Bootstrapped/regtested on x86_64-linux and
1001 - 1100 of 6068 matches
Mail list logo