[Bug target/59839] New: i686 AVX2 gather intrinsic not compiling with -O0

[dejan.crnila at dewesoft dot si]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59839

Bug ID: 59839
   Summary: i686 AVX2 gather intrinsic not compiling with -O0
   Product: gcc
   Version: unknown
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: target
  Assignee: unassigned at gcc dot gnu.org
  Reporter: dejan.crnila at dewesoft dot si

Created attachment 31850
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31850&action=edit
preprocesed source

the following source does

#include 

void test(const float* data)
{
__m256i i = _mm256_set1_epi32(1);
__m256 d = _mm256_i32gather_ps(data, i, 1);
}

does not compile with

gcc -mavx2 -O0 -c test.c

note: it does compile with -O1 or above

error:

test.c: In function 'test':
test.c:7:1: error: unrecognizable insn:
 }
 ^
(insn 28 27 0 2 (parallel [
(set (mem/c:V8SF (plus:SI (reg/f:SI 54 virtual-stack-vars
(const_int -64 [0xffc0])) [0 d+0
(unspec:V8SF [
(reg:V8SF 61 [ D.7004 ])
(mem:SF (unspec:SI [
(reg/f:SI 71)
(reg:V8SI 60 [ D.7005 ])
(const_int 1 [0x1])
] UNSPEC_VSIBADDR) [0 S4 A8])
(mem:BLK (scratch) [0 A8])
(reg:V8SF 59 [ D.7004 ])
] UNSPEC_GATHER))
(clobber (scratch:V8SF))
]) test.c:6 -1
 (nil))
test.c:7:1: internal compiler error: in extract_insn, at recog.c:2150
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://sourceforge.net/projects/mingwbuilds/> for instructions.


gcc -v:

Using built-in specs.
COLLECT_GCC=c:\mingw\bin\gcc.exe
COLLECT_LTO_WRAPPER=c:/mingw/bin/../libexec/gcc/i686-w64-mingw32/4.8.0/lto-wrapper.exe
Target: i686-w64-mingw32
Configured with: ../../../src/gcc-4.8.0/configure --host=i686-w64-mingw32
--build=i686-w64-mingw32 --target=i686-w64-mingw32 --prefix=/mingw32
--with-sysroot=/t
emp/x32-480-win32-sjlj-r2/mingw32 --enable-shared --enable-static
--enable-targets=all --enable-multilib --enable-languages=c,c++,fortran,lto
--enable-libstdcxx
-time=yes --enable-threads=win32 --enable-libgomp --enable-lto
--enable-graphite --enable-checking=release --enable-fully-dynamic-string
--enable-version-specif
ic-runtime-libs --enable-sjlj-exceptions --disable-isl-version-check
--disable-cloog-version-check --disable-libstdcxx-pch --disable-libstdcxx-debug
--disable-b
ootstrap --disable-rpath --disable-win32-registry --disable-nls
--disable-werror --disable-symvers --with-gnu-as --with-gnu-ld
--with-arch-32=i686 --with-arch-6
4=nocona --with-tune-32=generic --with-tune-64=core2
--with-host-libstdcxx='-static -lstdc++' --with-libiconv --with-system-zlib
--with-gmp=/temp/mingw-prereq/i
686-w64-mingw32-static --with-mpfr=/temp/mingw-prereq/i686-w64-mingw32-static
--with-mpc=/temp/mingw-prereq/i686-w64-mingw32-static
--with-isl=/temp/mingw-prere
q/i686-w64-mingw32-static
--with-cloog=/temp/mingw-prereq/i686-w64-mingw32-static
--enable-cloog-backend=isl --with-pkgversion='rev2, Built by MinGW-builds proj
ect' --with-bugurl=http://sourceforge.net/projects/mingwbuilds/ CFLAGS='-O2
-pipe -I/temp/x32-480-win32-sjlj-r2/libs/include
-I/temp/mingw-prereq/x32-zlib/inclu
de -I/temp/mingw-prereq/i686-w64-mingw32-static/include' CXXFLAGS='-O2 -pipe
-I/temp/x32-480-win32-sjlj-r2/libs/include
-I/temp/mingw-prereq/x32-zlib/include -I
/temp/mingw-prereq/i686-w64-mingw32-static/include' CPPFLAGS= LDFLAGS='-pipe
-L/temp/x32-480-win32-sjlj-r2/libs/lib -L/temp/mingw-prereq/x32-zlib/lib
-L/temp/mi
ngw-prereq/i686-w64-mingw32-static/lib
-L/temp/x32-480-win32-sjlj-r2/mingw32/opt/lib'
Thread model: win32
gcc version 4.8.0 (rev2, Built by MinGW-builds project)

[Bug tree-optimization/66580] New: max reduction does not auto vectorize

[dejan.crnila at dewesoft dot si]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66580

Bug ID: 66580
   Summary: max reduction does not auto vectorize
   Product: gcc
   Version: 5.1.0
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: tree-optimization
  Assignee: unassigned at gcc dot gnu.org
  Reporter: dejan.crnila at dewesoft dot si
  Target Milestone: ---

following code does (max reduction) not auto vectorize on x64/mingw gcc

#include 
#include 

int main()
{
float a[] __attribute__((aligned(0x20))) = {0, 1, 2, 3, 4, 5, 6, 7, 8,
9, 10, 11, 12, 13, 14, 15, 16};
float maxval = -FLT_MAX;
for (size_t i = 0; i < 16; i++)
maxval = a[i] > maxval ? a[i] : maxval;

printf("%g\n", maxval);
}

compiled as 

gcc test.c -O3 -Wall -o test.exe -ftree-vectorizer-verbose=6 
-fopt-info-vec-all -ffast-math -mavx

not supported or bug?