; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86,X86-SSE ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=X86,X86-AVX,X86-AVX1 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=X86,X86-AVX,X86-AVX2 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512vl | FileCheck %s --check-prefixes=X86,X86-AVX512,X86-AVX512VL ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512fp16 | FileCheck %s --check-prefixes=X86,X86-AVX512,X86-AVX512FP16 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefixes=X86,X86-AVX512,X86-AVX512VLDQ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X64,X64-SSE ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=X64,X64-AVX,X64-AVX1 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=X64,X64-AVX,X64-AVX2 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl | FileCheck %s --check-prefixes=X64,X64-AVX512,X64-AVX512VL ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512fp16 | FileCheck %s --check-prefixes=X64,X64-AVX512,X64-AVX512FP16 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefixes=X64,X64-AVX512,X64-AVX512VLDQ ; ; 128-bit Vectors ; define <2 x double> @fneg_v2f64(<2 x double> %p) nounwind { ; X86-SSE-LABEL: fneg_v2f64: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: xorps {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0 ; X86-SSE-NEXT: retl ; ; X86-AVX-LABEL: fneg_v2f64: ; X86-AVX: # %bb.0: ; X86-AVX-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0, %xmm0 ; X86-AVX-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v2f64: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: vpxorq {{\.?LCPI[0-9]+_[0-9]+}}{1to2}, %xmm0, %xmm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v2f64: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}{1to2}, %xmm0, %xmm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v2f64: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}{1to2}, %xmm0, %xmm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v2f64: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: xorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 ; X64-SSE-NEXT: retq ; ; X64-AVX-LABEL: fneg_v2f64: ; X64-AVX: # %bb.0: ; X64-AVX-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 ; X64-AVX-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v2f64: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpxorq {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to2}, %xmm0, %xmm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v2f64: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to2}, %xmm0, %xmm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v2f64: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to2}, %xmm0, %xmm0 ; X64-AVX512VLDQ-NEXT: retq %t = fsub <2 x double> , %p ret <2 x double> %t } define <4 x float> @fneg_v4f32(<4 x float> %p) nounwind { ; X86-SSE-LABEL: fneg_v4f32: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: xorps {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0 ; X86-SSE-NEXT: retl ; ; X86-AVX1-LABEL: fneg_v4f32: ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0, %xmm0 ; X86-AVX1-NEXT: retl ; ; X86-AVX2-LABEL: fneg_v4f32: ; X86-AVX2: # %bb.0: ; X86-AVX2-NEXT: vbroadcastss {{.*#+}} xmm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX2-NEXT: vxorps %xmm1, %xmm0, %xmm0 ; X86-AVX2-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v4f32: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: vpxord {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %xmm0, %xmm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v4f32: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %xmm0, %xmm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v4f32: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %xmm0, %xmm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v4f32: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: xorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 ; X64-SSE-NEXT: retq ; ; X64-AVX1-LABEL: fneg_v4f32: ; X64-AVX1: # %bb.0: ; X64-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 ; X64-AVX1-NEXT: retq ; ; X64-AVX2-LABEL: fneg_v4f32: ; X64-AVX2: # %bb.0: ; X64-AVX2-NEXT: vbroadcastss {{.*#+}} xmm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX2-NEXT: vxorps %xmm1, %xmm0, %xmm0 ; X64-AVX2-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v4f32: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpxord {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v4f32: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v4f32: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0 ; X64-AVX512VLDQ-NEXT: retq %t = fsub <4 x float> , %p ret <4 x float> %t } define <8 x half> @fneg_v8f16(ptr %p) nounwind { ; X86-SSE-LABEL: fneg_v8f16: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-SSE-NEXT: movaps (%eax), %xmm0 ; X86-SSE-NEXT: xorps {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0 ; X86-SSE-NEXT: retl ; ; X86-AVX1-LABEL: fneg_v8f16: ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX1-NEXT: vmovaps (%eax), %xmm0 ; X86-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0, %xmm0 ; X86-AVX1-NEXT: retl ; ; X86-AVX2-LABEL: fneg_v8f16: ; X86-AVX2: # %bb.0: ; X86-AVX2-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX2-NEXT: vpbroadcastw {{.*#+}} xmm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX2-NEXT: vpxor (%eax), %xmm0, %xmm0 ; X86-AVX2-NEXT: retl ; ; X86-AVX512-LABEL: fneg_v8f16: ; X86-AVX512: # %bb.0: ; X86-AVX512-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX512-NEXT: vpbroadcastw {{.*#+}} xmm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX512-NEXT: vpxor (%eax), %xmm0, %xmm0 ; X86-AVX512-NEXT: retl ; ; X64-SSE-LABEL: fneg_v8f16: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps (%rdi), %xmm0 ; X64-SSE-NEXT: xorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 ; X64-SSE-NEXT: retq ; ; X64-AVX1-LABEL: fneg_v8f16: ; X64-AVX1: # %bb.0: ; X64-AVX1-NEXT: vmovaps (%rdi), %xmm0 ; X64-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 ; X64-AVX1-NEXT: retq ; ; X64-AVX2-LABEL: fneg_v8f16: ; X64-AVX2: # %bb.0: ; X64-AVX2-NEXT: vpbroadcastw {{.*#+}} xmm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX2-NEXT: vpxor (%rdi), %xmm0, %xmm0 ; X64-AVX2-NEXT: retq ; ; X64-AVX512-LABEL: fneg_v8f16: ; X64-AVX512: # %bb.0: ; X64-AVX512-NEXT: vpbroadcastw {{.*#+}} xmm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX512-NEXT: vpxor (%rdi), %xmm0, %xmm0 ; X64-AVX512-NEXT: retq %v = load <8 x half>, ptr %p, align 16 %nnv = fsub <8 x half> , %v ret <8 x half> %nnv } ; ; 256-bit Vectors ; define <4 x double> @fneg_v4f64(<4 x double> %p) nounwind { ; X86-SSE-LABEL: fneg_v4f64: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: movaps {{.*#+}} xmm2 = [-0.0E+0,-0.0E+0] ; X86-SSE-NEXT: xorps %xmm2, %xmm0 ; X86-SSE-NEXT: xorps %xmm2, %xmm1 ; X86-SSE-NEXT: retl ; ; X86-AVX1-LABEL: fneg_v4f64: ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0 ; X86-AVX1-NEXT: retl ; ; X86-AVX2-LABEL: fneg_v4f64: ; X86-AVX2: # %bb.0: ; X86-AVX2-NEXT: vbroadcastsd {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX2-NEXT: vxorps %ymm1, %ymm0, %ymm0 ; X86-AVX2-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v4f64: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: vpxorq {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %ymm0, %ymm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v4f64: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %ymm0, %ymm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v4f64: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %ymm0, %ymm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v4f64: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps {{.*#+}} xmm2 = [-0.0E+0,-0.0E+0] ; X64-SSE-NEXT: xorps %xmm2, %xmm0 ; X64-SSE-NEXT: xorps %xmm2, %xmm1 ; X64-SSE-NEXT: retq ; ; X64-AVX1-LABEL: fneg_v4f64: ; X64-AVX1: # %bb.0: ; X64-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0 ; X64-AVX1-NEXT: retq ; ; X64-AVX2-LABEL: fneg_v4f64: ; X64-AVX2: # %bb.0: ; X64-AVX2-NEXT: vbroadcastsd {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX2-NEXT: vxorps %ymm1, %ymm0, %ymm0 ; X64-AVX2-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v4f64: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpxorq {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %ymm0, %ymm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v4f64: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %ymm0, %ymm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v4f64: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %ymm0, %ymm0 ; X64-AVX512VLDQ-NEXT: retq %t = fsub <4 x double> , %p ret <4 x double> %t } define <8 x float> @fneg_v8f32(<8 x float> %p) nounwind { ; X86-SSE-LABEL: fneg_v8f32: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: movaps {{.*#+}} xmm2 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-SSE-NEXT: xorps %xmm2, %xmm0 ; X86-SSE-NEXT: xorps %xmm2, %xmm1 ; X86-SSE-NEXT: retl ; ; X86-AVX1-LABEL: fneg_v8f32: ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0 ; X86-AVX1-NEXT: retl ; ; X86-AVX2-LABEL: fneg_v8f32: ; X86-AVX2: # %bb.0: ; X86-AVX2-NEXT: vbroadcastss {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX2-NEXT: vxorps %ymm1, %ymm0, %ymm0 ; X86-AVX2-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v8f32: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: vpxord {{\.?LCPI[0-9]+_[0-9]+}}{1to8}, %ymm0, %ymm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v8f32: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}{1to8}, %ymm0, %ymm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v8f32: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}{1to8}, %ymm0, %ymm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v8f32: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps {{.*#+}} xmm2 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-SSE-NEXT: xorps %xmm2, %xmm0 ; X64-SSE-NEXT: xorps %xmm2, %xmm1 ; X64-SSE-NEXT: retq ; ; X64-AVX1-LABEL: fneg_v8f32: ; X64-AVX1: # %bb.0: ; X64-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0 ; X64-AVX1-NEXT: retq ; ; X64-AVX2-LABEL: fneg_v8f32: ; X64-AVX2: # %bb.0: ; X64-AVX2-NEXT: vbroadcastss {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX2-NEXT: vxorps %ymm1, %ymm0, %ymm0 ; X64-AVX2-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v8f32: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpxord {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm0, %ymm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v8f32: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm0, %ymm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v8f32: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm0, %ymm0 ; X64-AVX512VLDQ-NEXT: retq %t = fsub <8 x float> , %p ret <8 x float> %t } define <16 x half> @fneg_v16f16(ptr %p) nounwind { ; X86-SSE-LABEL: fneg_v16f16: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-SSE-NEXT: movaps {{.*#+}} xmm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-SSE-NEXT: movaps (%eax), %xmm0 ; X86-SSE-NEXT: xorps %xmm1, %xmm0 ; X86-SSE-NEXT: xorps 16(%eax), %xmm1 ; X86-SSE-NEXT: retl ; ; X86-AVX1-LABEL: fneg_v16f16: ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX1-NEXT: vmovups (%eax), %ymm0 ; X86-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0 ; X86-AVX1-NEXT: retl ; ; X86-AVX2-LABEL: fneg_v16f16: ; X86-AVX2: # %bb.0: ; X86-AVX2-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX2-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX2-NEXT: vpxor (%eax), %ymm0, %ymm0 ; X86-AVX2-NEXT: retl ; ; X86-AVX512-LABEL: fneg_v16f16: ; X86-AVX512: # %bb.0: ; X86-AVX512-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX512-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX512-NEXT: vpxor (%eax), %ymm0, %ymm0 ; X86-AVX512-NEXT: retl ; ; X64-SSE-LABEL: fneg_v16f16: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps {{.*#+}} xmm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-SSE-NEXT: movaps (%rdi), %xmm0 ; X64-SSE-NEXT: xorps %xmm1, %xmm0 ; X64-SSE-NEXT: xorps 16(%rdi), %xmm1 ; X64-SSE-NEXT: retq ; ; X64-AVX1-LABEL: fneg_v16f16: ; X64-AVX1: # %bb.0: ; X64-AVX1-NEXT: vmovups (%rdi), %ymm0 ; X64-AVX1-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0 ; X64-AVX1-NEXT: retq ; ; X64-AVX2-LABEL: fneg_v16f16: ; X64-AVX2: # %bb.0: ; X64-AVX2-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX2-NEXT: vpxor (%rdi), %ymm0, %ymm0 ; X64-AVX2-NEXT: retq ; ; X64-AVX512-LABEL: fneg_v16f16: ; X64-AVX512: # %bb.0: ; X64-AVX512-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX512-NEXT: vpxor (%rdi), %ymm0, %ymm0 ; X64-AVX512-NEXT: retq %v = load <16 x half>, ptr %p, align 16 %nnv = fsub <16 x half> , %v ret <16 x half> %nnv } ; ; 512-bit Vectors ; define <8 x double> @fneg_v8f64(<8 x double> %p) nounwind { ; X86-SSE-LABEL: fneg_v8f64: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: pushl %ebp ; X86-SSE-NEXT: movl %esp, %ebp ; X86-SSE-NEXT: andl $-16, %esp ; X86-SSE-NEXT: subl $16, %esp ; X86-SSE-NEXT: movaps {{.*#+}} xmm3 = [-0.0E+0,-0.0E+0] ; X86-SSE-NEXT: xorps %xmm3, %xmm0 ; X86-SSE-NEXT: xorps %xmm3, %xmm1 ; X86-SSE-NEXT: xorps %xmm3, %xmm2 ; X86-SSE-NEXT: xorps 8(%ebp), %xmm3 ; X86-SSE-NEXT: movl %ebp, %esp ; X86-SSE-NEXT: popl %ebp ; X86-SSE-NEXT: retl ; ; X86-AVX-LABEL: fneg_v8f64: ; X86-AVX: # %bb.0: ; X86-AVX-NEXT: vbroadcastsd {{.*#+}} ymm2 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX-NEXT: vxorps %ymm2, %ymm0, %ymm0 ; X86-AVX-NEXT: vxorps %ymm2, %ymm1, %ymm1 ; X86-AVX-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v8f64: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: vpxorq {{\.?LCPI[0-9]+_[0-9]+}}{1to8}, %zmm0, %zmm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v8f64: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}{1to8}, %zmm0, %zmm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v8f64: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}{1to8}, %zmm0, %zmm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v8f64: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps {{.*#+}} xmm4 = [-0.0E+0,-0.0E+0] ; X64-SSE-NEXT: xorps %xmm4, %xmm0 ; X64-SSE-NEXT: xorps %xmm4, %xmm1 ; X64-SSE-NEXT: xorps %xmm4, %xmm2 ; X64-SSE-NEXT: xorps %xmm4, %xmm3 ; X64-SSE-NEXT: retq ; ; X64-AVX-LABEL: fneg_v8f64: ; X64-AVX: # %bb.0: ; X64-AVX-NEXT: vbroadcastsd {{.*#+}} ymm2 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX-NEXT: vxorps %ymm2, %ymm0, %ymm0 ; X64-AVX-NEXT: vxorps %ymm2, %ymm1, %ymm1 ; X64-AVX-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v8f64: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpxorq {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %zmm0, %zmm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v8f64: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %zmm0, %zmm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v8f64: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %zmm0, %zmm0 ; X64-AVX512VLDQ-NEXT: retq %t = fsub <8 x double> , %p ret <8 x double> %t } define <16 x float> @fneg_v16f32(<16 x float> %p) nounwind { ; X86-SSE-LABEL: fneg_v16f32: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: pushl %ebp ; X86-SSE-NEXT: movl %esp, %ebp ; X86-SSE-NEXT: andl $-16, %esp ; X86-SSE-NEXT: subl $16, %esp ; X86-SSE-NEXT: movaps {{.*#+}} xmm3 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-SSE-NEXT: xorps %xmm3, %xmm0 ; X86-SSE-NEXT: xorps %xmm3, %xmm1 ; X86-SSE-NEXT: xorps %xmm3, %xmm2 ; X86-SSE-NEXT: xorps 8(%ebp), %xmm3 ; X86-SSE-NEXT: movl %ebp, %esp ; X86-SSE-NEXT: popl %ebp ; X86-SSE-NEXT: retl ; ; X86-AVX-LABEL: fneg_v16f32: ; X86-AVX: # %bb.0: ; X86-AVX-NEXT: vbroadcastss {{.*#+}} ymm2 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX-NEXT: vxorps %ymm2, %ymm0, %ymm0 ; X86-AVX-NEXT: vxorps %ymm2, %ymm1, %ymm1 ; X86-AVX-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v16f32: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: vpxord {{\.?LCPI[0-9]+_[0-9]+}}{1to16}, %zmm0, %zmm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v16f32: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}{1to16}, %zmm0, %zmm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v16f32: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}{1to16}, %zmm0, %zmm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v16f32: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps {{.*#+}} xmm4 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-SSE-NEXT: xorps %xmm4, %xmm0 ; X64-SSE-NEXT: xorps %xmm4, %xmm1 ; X64-SSE-NEXT: xorps %xmm4, %xmm2 ; X64-SSE-NEXT: xorps %xmm4, %xmm3 ; X64-SSE-NEXT: retq ; ; X64-AVX-LABEL: fneg_v16f32: ; X64-AVX: # %bb.0: ; X64-AVX-NEXT: vbroadcastss {{.*#+}} ymm2 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX-NEXT: vxorps %ymm2, %ymm0, %ymm0 ; X64-AVX-NEXT: vxorps %ymm2, %ymm1, %ymm1 ; X64-AVX-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v16f32: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpxord {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to16}, %zmm0, %zmm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v16f32: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to16}, %zmm0, %zmm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v16f32: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to16}, %zmm0, %zmm0 ; X64-AVX512VLDQ-NEXT: retq %t = fsub <16 x float> , %p ret <16 x float> %t } define <32 x half> @fneg_v32f16(ptr %p) nounwind { ; X86-SSE-LABEL: fneg_v32f16: ; X86-SSE: # %bb.0: ; X86-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-SSE-NEXT: movaps {{.*#+}} xmm3 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-SSE-NEXT: movaps (%eax), %xmm0 ; X86-SSE-NEXT: xorps %xmm3, %xmm0 ; X86-SSE-NEXT: movaps 16(%eax), %xmm1 ; X86-SSE-NEXT: xorps %xmm3, %xmm1 ; X86-SSE-NEXT: movaps 32(%eax), %xmm2 ; X86-SSE-NEXT: xorps %xmm3, %xmm2 ; X86-SSE-NEXT: xorps 48(%eax), %xmm3 ; X86-SSE-NEXT: retl ; ; X86-AVX1-LABEL: fneg_v32f16: ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX1-NEXT: vbroadcastss {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX1-NEXT: vxorps (%eax), %ymm1, %ymm0 ; X86-AVX1-NEXT: vxorps 32(%eax), %ymm1, %ymm1 ; X86-AVX1-NEXT: retl ; ; X86-AVX2-LABEL: fneg_v32f16: ; X86-AVX2: # %bb.0: ; X86-AVX2-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX2-NEXT: vpbroadcastw {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX2-NEXT: vpxor (%eax), %ymm1, %ymm0 ; X86-AVX2-NEXT: vpxor 32(%eax), %ymm1, %ymm1 ; X86-AVX2-NEXT: retl ; ; X86-AVX512VL-LABEL: fneg_v32f16: ; X86-AVX512VL: # %bb.0: ; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX512VL-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX512VL-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0 ; X86-AVX512VL-NEXT: vpxorq (%eax), %zmm0, %zmm0 ; X86-AVX512VL-NEXT: retl ; ; X86-AVX512FP16-LABEL: fneg_v32f16: ; X86-AVX512FP16: # %bb.0: ; X86-AVX512FP16-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX512FP16-NEXT: vpbroadcastw {{.*#+}} zmm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX512FP16-NEXT: vpxorq (%eax), %zmm0, %zmm0 ; X86-AVX512FP16-NEXT: retl ; ; X86-AVX512VLDQ-LABEL: fneg_v32f16: ; X86-AVX512VLDQ: # %bb.0: ; X86-AVX512VLDQ-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-AVX512VLDQ-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X86-AVX512VLDQ-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0 ; X86-AVX512VLDQ-NEXT: vpxorq (%eax), %zmm0, %zmm0 ; X86-AVX512VLDQ-NEXT: retl ; ; X64-SSE-LABEL: fneg_v32f16: ; X64-SSE: # %bb.0: ; X64-SSE-NEXT: movaps {{.*#+}} xmm3 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-SSE-NEXT: movaps (%rdi), %xmm0 ; X64-SSE-NEXT: xorps %xmm3, %xmm0 ; X64-SSE-NEXT: movaps 16(%rdi), %xmm1 ; X64-SSE-NEXT: xorps %xmm3, %xmm1 ; X64-SSE-NEXT: movaps 32(%rdi), %xmm2 ; X64-SSE-NEXT: xorps %xmm3, %xmm2 ; X64-SSE-NEXT: xorps 48(%rdi), %xmm3 ; X64-SSE-NEXT: retq ; ; X64-AVX1-LABEL: fneg_v32f16: ; X64-AVX1: # %bb.0: ; X64-AVX1-NEXT: vbroadcastss {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX1-NEXT: vxorps (%rdi), %ymm1, %ymm0 ; X64-AVX1-NEXT: vxorps 32(%rdi), %ymm1, %ymm1 ; X64-AVX1-NEXT: retq ; ; X64-AVX2-LABEL: fneg_v32f16: ; X64-AVX2: # %bb.0: ; X64-AVX2-NEXT: vpbroadcastw {{.*#+}} ymm1 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX2-NEXT: vpxor (%rdi), %ymm1, %ymm0 ; X64-AVX2-NEXT: vpxor 32(%rdi), %ymm1, %ymm1 ; X64-AVX2-NEXT: retq ; ; X64-AVX512VL-LABEL: fneg_v32f16: ; X64-AVX512VL: # %bb.0: ; X64-AVX512VL-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX512VL-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0 ; X64-AVX512VL-NEXT: vpxorq (%rdi), %zmm0, %zmm0 ; X64-AVX512VL-NEXT: retq ; ; X64-AVX512FP16-LABEL: fneg_v32f16: ; X64-AVX512FP16: # %bb.0: ; X64-AVX512FP16-NEXT: vpbroadcastw {{.*#+}} zmm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX512FP16-NEXT: vpxorq (%rdi), %zmm0, %zmm0 ; X64-AVX512FP16-NEXT: retq ; ; X64-AVX512VLDQ-LABEL: fneg_v32f16: ; X64-AVX512VLDQ: # %bb.0: ; X64-AVX512VLDQ-NEXT: vpbroadcastw {{.*#+}} ymm0 = [-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0,-0.0E+0] ; X64-AVX512VLDQ-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0 ; X64-AVX512VLDQ-NEXT: vpxorq (%rdi), %zmm0, %zmm0 ; X64-AVX512VLDQ-NEXT: retq %v = load <32 x half>, ptr %p, align 16 %nnv = fsub <32 x half> , %v ret <32 x half> %nnv } ;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line: ; X64: {{.*}} ; X86: {{.*}}