From: kyukhin Date: Mon, 21 Mar 2016 10:51:04 +0000 (+0000) Subject: PR target/70293 X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=08b2b6926d41676e047e23660d51eca9e03c93de;p=thirdparty%2Fgcc.git PR target/70293 gcc/ * config/i386 (define_insn "*vec_dup"/AVX2): Block third alternative for AVX-512VL target, gcc/testsuite/ * gcc.target/i386/pr70293.c: New test. git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@234363 138bc75d-0d04-0410-961f-82ee72b054a4 --- diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 7a918672563a..02c258019399 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,9 @@ +2016-03-21 Kirill Yukhin + + PR target/70293 + * config/i386/sse.md: (define_insn "*vec_dup"/AVX2): + Block third alternative for AVX-512VL target, + 2016-03-21 Martin Liska PR hsa/70234 diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md index b25c2465ebf7..44141ea5d97f 100644 --- a/gcc/config/i386/sse.md +++ b/gcc/config/i386/sse.md @@ -17419,7 +17419,8 @@ vbroadcast\t{%1, %0|%0, %1} vbroadcast\t{%x1, %0|%0, %x1} #" - [(set_attr "type" "ssemov") + [(set_attr "isa" "*,*,noavx512vl") + (set_attr "type" "ssemov") (set_attr "prefix_extra" "1") (set_attr "prefix" "maybe_evex") (set_attr "mode" "")]) diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog index f7ca6848fd04..9f4d811710f7 100644 --- a/gcc/testsuite/ChangeLog +++ b/gcc/testsuite/ChangeLog @@ -1,3 +1,8 @@ +2016-03-21 Kirill Yukhin + + PR target/70293 + * gcc.target/i386/pr70293.c: New test. + 2016-03-21 Richard Biener PR tree-optimization/70288 diff --git a/gcc/testsuite/gcc.target/i386/pr70293.c b/gcc/testsuite/gcc.target/i386/pr70293.c new file mode 100644 index 000000000000..4510166ead86 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr70293.c @@ -0,0 +1,38 @@ +/* PR target/70293 */ +/* { dg-do compile } */ +/* { dg-require-effective-target lp64 } */ +/* { dg-options "-mtune=westmere -mavx512vl -O2" } */ + +typedef short __v8hi __attribute__((__vector_size__(16))); +typedef int __v8hu __attribute__((__vector_size__(16))); +typedef long __m128i __attribute__((__vector_size__(16))); +__m128i _mm_madd_epi16___B, _mm_mullo_epi16___A, + scaled_bilinear_scanline_sse2_8888_8_8888_OVER_xmm_b, + scaled_bilinear_scanline_sse2_8888_8_8888_OVER___trans_tmp_16, + scaled_bilinear_scanline_sse2_8888_8_8888_OVER___trans_tmp_13; +int _mm_srli_epi16___B, scaled_bilinear_scanline_sse2_8888_8_8888_OVER_m, + scaled_bilinear_scanline_sse2_8888_8_8888_OVER_dst, + scaled_bilinear_scanline_sse2_8888_8_8888_OVER_wt; +__m128i _mm_set_epi16(); +void _mm_cvtsi128_si32(); +void +scaled_bilinear_scanline_sse2_8888_8_8888_OVER(int p1) { + __m128i __trans_tmp_12, __trans_tmp_6, __trans_tmp_5, xmm_x = _mm_set_epi16(); + int mask; + __trans_tmp_5 = (__m128i){scaled_bilinear_scanline_sse2_8888_8_8888_OVER_wt}; + __trans_tmp_6 = (__m128i)(__v8hi){p1, p1, p1, p1, p1, p1, p1, p1}; + while (scaled_bilinear_scanline_sse2_8888_8_8888_OVER_dst) { + scaled_bilinear_scanline_sse2_8888_8_8888_OVER_m = mask++; + if (scaled_bilinear_scanline_sse2_8888_8_8888_OVER_m) { + __trans_tmp_12 = + (__m128i)((__v8hu)_mm_mullo_epi16___A * (__v8hu)__trans_tmp_6); + scaled_bilinear_scanline_sse2_8888_8_8888_OVER_xmm_b = __trans_tmp_12; + scaled_bilinear_scanline_sse2_8888_8_8888_OVER___trans_tmp_13 = + (__m128i)__builtin_ia32_psrlwi128((__v8hi)xmm_x, _mm_srli_epi16___B); + scaled_bilinear_scanline_sse2_8888_8_8888_OVER___trans_tmp_16 = + (__m128i)__builtin_ia32_pmaddwd128((__v8hi)__trans_tmp_5, + (__v8hi)_mm_madd_epi16___B); + _mm_cvtsi128_si32(); + } + } +}