test/CodeGen/X86/avx512dqvl-intrinsics.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -disable-peephole -mtriple=i686-unknown-unknown -mattr=+avx512dq,+avx512vl --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
; RUN: llc < %s -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx512dq,+avx512vl --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64

declare <2 x i64> @llvm.x86.avx512.mask.cvtpd2qq.128(<2 x double>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvt_pd2qq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_pd2qq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtpd2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x7b,0xc8]
; X86-NEXT:    vcvtpd2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x7b,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_pd2qq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtpd2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x7b,0xc8]
; X64-NEXT:    vcvtpd2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x7b,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.cvtpd2qq.256(<4 x double>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvt_pd2qq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_pd2qq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtpd2qq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x7b,0xc8]
; X86-NEXT:    vcvtpd2qq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x7b,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_pd2qq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtpd2qq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x7b,0xc8]
; X64-NEXT:    vcvtpd2qq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x7b,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.128(<2 x double>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvt_pd2uqq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_pd2uqq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtpd2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x79,0xc8]
; X86-NEXT:    vcvtpd2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x79,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_pd2uqq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtpd2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x79,0xc8]
; X64-NEXT:    vcvtpd2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x79,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.256(<4 x double>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvt_pd2uqq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_pd2uqq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtpd2uqq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x79,0xc8]
; X86-NEXT:    vcvtpd2uqq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x79,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_pd2uqq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtpd2uqq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x79,0xc8]
; X64-NEXT:    vcvtpd2uqq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x79,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvt_ps2qq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtps2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0xc8]
; X86-NEXT:    vcvtps2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtps2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0xc8]
; X64-NEXT:    vcvtps2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

define <2 x i64> @test_int_x86_avx512_cvt_ps2qq_128_load(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvt_ps2qq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvt_ps2qq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvt_ps2qq_128_load(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvt_ps2qq_128_load(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvt_ps2qq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvt_ps2qq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}


define <2 x i64> @test_int_x86_avx512_cvt_ps2qq_128_load_2(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvt_ps2qq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvt_ps2qq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvt_ps2qq_128_load_2(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvt_ps2qq_128_load_2(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvt_ps2qq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvt_ps2qq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_cvt_ps2qq_128_load_3(<4 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvt_ps2qq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvt_ps2qq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvt_ps2qq_128_load_3(<4 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvt_ps2qq_128_load_3(<4 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvt_ps2qq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2qq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7b,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvt_ps2qq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2qq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7b,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

declare <4 x i64> @llvm.x86.avx512.mask.cvtps2qq.256(<4 x float>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvt_ps2qq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtps2qq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x7b,0xc8]
; X86-NEXT:    vcvtps2qq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x7b,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtps2qq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x7b,0xc8]
; X64-NEXT:    vcvtps2qq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x7b,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvtps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvt_ps2uqq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtps2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0xc8]
; X86-NEXT:    vcvtps2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtps2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0xc8]
; X64-NEXT:    vcvtps2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

define <2 x i64> @test_int_x86_avx512_cvt_ps2uqq_128_load(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvt_ps2uqq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvt_ps2uqq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvt_ps2uqq_128_load(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvt_ps2uqq_128_load(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvt_ps2uqq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvt_ps2uqq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_cvt_ps2uqq_128_load_2(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvt_ps2uqq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvt_ps2uqq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvt_ps2uqq_128_load_2(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvt_ps2uqq_128_load_2(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvt_ps2uqq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvt_ps2uqq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_cvt_ps2uqq_128_load_3(<4 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvt_ps2uqq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvt_ps2uqq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvt_ps2uqq_128_load_3(<4 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvt_ps2uqq_128_load_3(<4 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvt_ps2uqq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvtps2uqq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x79,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvt_ps2uqq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvtps2uqq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x79,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

declare <4 x i64> @llvm.x86.avx512.mask.cvtps2uqq.256(<4 x float>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvt_ps2uqq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtps2uqq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x79,0xc8]
; X86-NEXT:    vcvtps2uqq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x79,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtps2uqq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x79,0xc8]
; X64-NEXT:    vcvtps2uqq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x79,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvtps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_qq2ps_128(<2 x i64> %x0, <4 x float> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfc,0x09,0x5b,0xc8]
; X86-NEXT:    vcvtqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfc,0x08,0x5b,0xc0]
; X86-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfc,0x09,0x5b,0xc8]
; X64-NEXT:    vcvtqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfc,0x08,0x5b,0xc0]
; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

define <4 x float>@test_int_x86_avx512_mask_cvt_qq2ps_128_zext(<2 x i64> %x0, <4 x float> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_128_zext:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfc,0x09,0x5b,0xc8]
; X86-NEXT:    vcvtqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfc,0x08,0x5b,0xc0]
; X86-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_128_zext:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfc,0x09,0x5b,0xc8]
; X64-NEXT:    vcvtqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfc,0x08,0x5b,0xc0]
; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 %x2)
  %res1 = shufflevector <4 x float> %res, <4 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
  %res2 = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 -1)
  %res3 = shufflevector <4 x float> %res2, <4 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
  %res4 = fadd <4 x float> %res1, %res3
  ret <4 x float> %res4
}

define <4 x float>@test_int_x86_avx512_mask_cvt_qq2ps_256(<4 x i64> %x0, <4 x float> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_256:
; X86:       # %bb.0:
; X86-NEXT:    vcvtqq2ps %ymm0, %xmm2 # encoding: [0x62,0xf1,0xfc,0x28,0x5b,0xd0]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtqq2ps %ymm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfc,0x29,0x5b,0xc8]
; X86-NEXT:    vaddps %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc2]
; X86-NEXT:    vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_256:
; X64:       # %bb.0:
; X64-NEXT:    vcvtqq2ps %ymm0, %xmm2 # encoding: [0x62,0xf1,0xfc,0x28,0x5b,0xd0]
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtqq2ps %ymm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfc,0x29,0x5b,0xc8]
; X64-NEXT:    vaddps %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc2]
; X64-NEXT:    vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT:    retq # encoding: [0xc3]
  %cvt1 = sitofp <4 x i64> %x0 to <4 x float>
  %1 = bitcast i8 %x2 to <8 x i1>
  %extract = shufflevector <8 x i1> %1, <8 x i1> %1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %2 = select <4 x i1> %extract, <4 x float> %cvt1, <4 x float> %x1
  %cvt = sitofp <4 x i64> %x0 to <4 x float>
  %res2 = fadd <4 x float> %2, %cvt
  ret <4 x float> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvttpd2qq.128(<2 x double>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvtt_pd2qq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_pd2qq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttpd2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x7a,0xc8]
; X86-NEXT:    vcvttpd2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x7a,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_pd2qq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttpd2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x7a,0xc8]
; X64-NEXT:    vcvttpd2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x7a,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.cvttpd2qq.256(<4 x double>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvtt_pd2qq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_pd2qq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttpd2qq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x7a,0xc8]
; X86-NEXT:    vcvttpd2qq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x7a,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_pd2qq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttpd2qq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x7a,0xc8]
; X64-NEXT:    vcvttpd2qq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x7a,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.128(<2 x double>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvtt_pd2uqq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_pd2uqq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttpd2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x78,0xc8]
; X86-NEXT:    vcvttpd2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x78,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_pd2uqq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttpd2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x09,0x78,0xc8]
; X64-NEXT:    vcvttpd2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0xfd,0x08,0x78,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.256(<4 x double>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvtt_pd2uqq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_pd2uqq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttpd2uqq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x78,0xc8]
; X86-NEXT:    vcvttpd2uqq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x78,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_pd2uqq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttpd2uqq %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0xfd,0x29,0x78,0xc8]
; X64-NEXT:    vcvttpd2uqq %ymm0, %ymm0 # encoding: [0x62,0xf1,0xfd,0x28,0x78,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvtt_ps2qq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttps2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0xc8]
; X86-NEXT:    vcvttps2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttps2qq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0xc8]
; X64-NEXT:    vcvttps2qq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

define <2 x i64> @test_int_x86_avx512_cvtt_ps2qq_128_load(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvtt_ps2qq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvtt_ps2qq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvtt_ps2qq_128_load(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvtt_ps2qq_128_load(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvtt_ps2qq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvtt_ps2qq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}


define <2 x i64> @test_int_x86_avx512_cvtt_ps2qq_128_load_2(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvtt_ps2qq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvtt_ps2qq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvtt_ps2qq_128_load_2(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvtt_ps2qq_128_load_2(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvtt_ps2qq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvtt_ps2qq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_cvtt_ps2qq_128_load_3(<4 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvtt_ps2qq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvtt_ps2qq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvtt_ps2qq_128_load_3(<4 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvtt_ps2qq_128_load_3(<4 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvtt_ps2qq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2qq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7a,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvtt_ps2qq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2qq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x7a,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

declare <4 x i64> @llvm.x86.avx512.mask.cvttps2qq.256(<4 x float>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvtt_ps2qq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttps2qq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x7a,0xc8]
; X86-NEXT:    vcvttps2qq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x7a,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttps2qq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x7a,0xc8]
; X64-NEXT:    vcvttps2qq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x7a,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvttps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_uqq2ps_128(<2 x i64> %x0, <4 x float> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtuqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xff,0x09,0x7a,0xc8]
; X86-NEXT:    vcvtuqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xff,0x08,0x7a,0xc0]
; X86-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtuqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xff,0x09,0x7a,0xc8]
; X64-NEXT:    vcvtuqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xff,0x08,0x7a,0xc0]
; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

define <4 x float>@test_int_x86_avx512_mask_cvt_uqq2ps_128_zext(<2 x i64> %x0, <4 x float> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_128_zext:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtuqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xff,0x09,0x7a,0xc8]
; X86-NEXT:    vcvtuqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xff,0x08,0x7a,0xc0]
; X86-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_128_zext:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtuqq2ps %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xff,0x09,0x7a,0xc8]
; X64-NEXT:    vcvtuqq2ps %xmm0, %xmm0 # encoding: [0x62,0xf1,0xff,0x08,0x7a,0xc0]
; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 %x2)
  %res1 = shufflevector <4 x float> %res, <4 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
  %res2 = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 -1)
  %res3 = shufflevector <4 x float> %res2, <4 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
  %res4 = fadd <4 x float> %res1, %res3
  ret <4 x float> %res4
}

declare <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.256(<4 x i64>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_uqq2ps_256(<4 x i64> %x0, <4 x float> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_256:
; X86:       # %bb.0:
; X86-NEXT:    vcvtuqq2ps %ymm0, %xmm2 # encoding: [0x62,0xf1,0xff,0x28,0x7a,0xd0]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvtuqq2ps %ymm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xff,0x29,0x7a,0xc8]
; X86-NEXT:    vaddps %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc2]
; X86-NEXT:    vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_256:
; X64:       # %bb.0:
; X64-NEXT:    vcvtuqq2ps %ymm0, %xmm2 # encoding: [0x62,0xf1,0xff,0x28,0x7a,0xd0]
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvtuqq2ps %ymm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0xff,0x29,0x7a,0xc8]
; X64-NEXT:    vaddps %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc2]
; X64-NEXT:    vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT:    retq # encoding: [0xc3]
  %cvt1 = uitofp <4 x i64> %x0 to <4 x float>
  %1 = bitcast i8 %x2 to <8 x i1>
  %extract = shufflevector <8 x i1> %1, <8 x i1> %1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %2 = select <4 x i1> %extract, <4 x float> %cvt1, <4 x float> %x1
  %cvt = uitofp <4 x i64> %x0 to <4 x float>
  %res2 = fadd <4 x float> %2, %cvt
  ret <4 x float> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_cvtt_ps2uqq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttps2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0xc8]
; X86-NEXT:    vcvttps2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0xc0]
; X86-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttps2uqq %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0xc8]
; X64-NEXT:    vcvttps2uqq %xmm0, %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0xc0]
; X64-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

define <2 x i64> @test_int_x86_avx512_cvtt_ps2uqq_128_load(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvtt_ps2uqq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvtt_ps2uqq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvtt_ps2uqq_128_load(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}


define <2 x i64> @test_int_x86_avx512_cvtt_ps2uqq_128_load_2(<2 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvtt_ps2uqq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvtt_ps2uqq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0b, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvtt_ps2uqq_128_load_2(<2 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0b, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load_2(<2 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load_2:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load_2:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <2 x float>, <2 x float>* %p
  %x0b = shufflevector <2 x float> %x0, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_cvtt_ps2uqq_128_load_3(<4 x float>* %p) {
; X86-LABEL: test_int_x86_avx512_cvtt_ps2uqq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_cvtt_ps2uqq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 # encoding: [0x62,0xf1,0x7d,0x08,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> undef, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_mask_cvtt_ps2uqq_128_load_3(<4 x float>* %p, <2 x i64> %passthru, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> %passthru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load_3(<4 x float>* %p, i8 %mask) {
; X86-LABEL: test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load_3:
; X86:       # %bb.0:
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT:    vcvttps2uqq (%eax), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x78,0x00]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_cvtt_ps2uqq_128_load_3:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT:    vcvttps2uqq (%rdi), %xmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0x89,0x78,0x07]
; X64-NEXT:    retq # encoding: [0xc3]
  %x0 = load <4 x float>, <4 x float>* %p
  %res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

declare <4 x i64> @llvm.x86.avx512.mask.cvttps2uqq.256(<4 x float>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_cvtt_ps2uqq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vcvttps2uqq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x78,0xc8]
; X86-NEXT:    vcvttps2uqq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x78,0xc0]
; X86-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vcvttps2uqq %xmm0, %ymm1 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0x78,0xc8]
; X64-NEXT:    vcvttps2uqq %xmm0, %ymm0 # encoding: [0x62,0xf1,0x7d,0x28,0x78,0xc0]
; X64-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0xd4,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.cvttps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.reduce.pd.128(<2 x double>, i32, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_reduce_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_reduce_pd_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vreducepd $4, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x56,0xc8,0x04]
; X86-NEXT:    vreducepd $8, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x56,0xc0,0x08]
; X86-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_reduce_pd_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vreducepd $4, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x56,0xc8,0x04]
; X64-NEXT:    vreducepd $8, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x56,0xc0,0x08]
; X64-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.reduce.pd.128(<2 x double> %x0, i32 4, <2 x double> %x2, i8 %x3)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.reduce.pd.128(<2 x double> %x0, i32 8, <2 x double> %x2, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.reduce.pd.256(<4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_reduce_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_reduce_pd_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vreducepd $4, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x56,0xc8,0x04]
; X86-NEXT:    vreducepd $0, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x56,0xc0,0x00]
; X86-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_reduce_pd_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vreducepd $4, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x56,0xc8,0x04]
; X64-NEXT:    vreducepd $0, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x56,0xc0,0x00]
; X64-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.reduce.pd.256(<4 x double> %x0, i32 4, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.reduce.pd.256(<4 x double> %x0, i32 0, <4 x double> %x2, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.reduce.ps.128(<4 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_reduce_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_reduce_ps_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vreduceps $4, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x56,0xc8,0x04]
; X86-NEXT:    vreduceps $88, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x56,0xc0,0x58]
; X86-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_reduce_ps_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vreduceps $4, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x56,0xc8,0x04]
; X64-NEXT:    vreduceps $88, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x56,0xc0,0x58]
; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.reduce.ps.128(<4 x float> %x0, i32 4, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.reduce.ps.128(<4 x float> %x0, i32 88, <4 x float> %x2, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_reduce_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_reduce_ps_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vreduceps $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x56,0xc8,0x0b]
; X86-NEXT:    vreduceps $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0b]
; X86-NEXT:    vaddps %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf4,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_reduce_ps_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vreduceps $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x56,0xc8,0x0b]
; X64-NEXT:    vreduceps $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0b]
; X64-NEXT:    vaddps %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf4,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.range.pd.128(<2 x double>, <2 x double>, i32, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_range_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_range_pd_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vrangepd $4, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x50,0xd1,0x04]
; X86-NEXT:    vrangepd $8, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x50,0xc1,0x08]
; X86-NEXT:    vaddpd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_range_pd_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vrangepd $4, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x50,0xd1,0x04]
; X64-NEXT:    vrangepd $8, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x50,0xc1,0x08]
; X64-NEXT:    vaddpd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.range.pd.128(<2 x double> %x0, <2 x double> %x1, i32 4, <2 x double> %x3, i8 %x4)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.range.pd.128(<2 x double> %x0, <2 x double> %x1, i32 8, <2 x double> %x3, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.range.pd.256(<4 x double>, <4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_range_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_range_pd_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vrangepd $4, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x50,0xd1,0x04]
; X86-NEXT:    vrangepd $88, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x50,0xc1,0x58]
; X86-NEXT:    vaddpd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_range_pd_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vrangepd $4, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x50,0xd1,0x04]
; X64-NEXT:    vrangepd $88, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x50,0xc1,0x58]
; X64-NEXT:    vaddpd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.range.pd.256(<4 x double> %x0, <4 x double> %x1, i32 4, <4 x double> %x3, i8 %x4)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.range.pd.256(<4 x double> %x0, <4 x double> %x1, i32 88, <4 x double> %x3, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.range.ps.128(<4 x float>, <4 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_range_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_range_ps_128:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vrangeps $4, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x50,0xd1,0x04]
; X86-NEXT:    vrangeps $88, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x50,0xc1,0x58]
; X86-NEXT:    vaddps %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe8,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_range_ps_128:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vrangeps $4, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x50,0xd1,0x04]
; X64-NEXT:    vrangeps $88, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x50,0xc1,0x58]
; X64-NEXT:    vaddps %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe8,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.range.ps.128(<4 x float> %x0, <4 x float> %x1, i32 4, <4 x float> %x3, i8 %x4)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.range.ps.128(<4 x float> %x0, <4 x float> %x1, i32 88, <4 x float> %x3, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.range.ps.256(<8 x float>, <8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_range_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_range_ps_256:
; X86:       # %bb.0:
; X86-NEXT:    kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT:    vrangeps $4, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x50,0xd1,0x04]
; X86-NEXT:    vrangeps $88, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x50,0xc1,0x58]
; X86-NEXT:    vaddps %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xec,0x58,0xc0]
; X86-NEXT:    retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_range_ps_256:
; X64:       # %bb.0:
; X64-NEXT:    kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT:    vrangeps $4, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x50,0xd1,0x04]
; X64-NEXT:    vrangeps $88, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x50,0xc1,0x58]
; X64-NEXT:    vaddps %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xec,0x58,0xc0]
; X64-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.range.ps.256(<8 x float> %x0, <8 x float> %x1, i32 4, <8 x float> %x3, i8 %x4)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.range.ps.256(<8 x float> %x0, <8 x float> %x1, i32 88, <8 x float> %x3, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <4 x i1> @llvm.x86.avx512.fpclass.ps.128(<4 x float>, i32)

define i8 @test_int_x86_avx512_fpclass_ps_128(<4 x float> %x0) {
; CHECK-LABEL: test_int_x86_avx512_fpclass_ps_128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfpclassps $4, %xmm0, %k1 # encoding: [0x62,0xf3,0x7d,0x08,0x66,0xc8,0x04]
; CHECK-NEXT:    vfpclassps $2, %xmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x66,0xc0,0x02]
; CHECK-NEXT:    kmovw %k0, %eax # encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    # kill: def $al killed $al killed $eax
; CHECK-NEXT:    ret{{[l|q]}} # encoding: [0xc3]
  %res = call <4 x i1> @llvm.x86.avx512.fpclass.ps.128(<4 x float> %x0, i32 2)
  %res1 = call <4 x i1> @llvm.x86.avx512.fpclass.ps.128(<4 x float> %x0, i32 4)
  %1 = and <4 x i1> %res1, %res
  %2 = shufflevector <4 x i1> %1, <4 x i1> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %3 = bitcast <8 x i1> %2 to i8
  ret i8 %3
}

declare <8 x i1> @llvm.x86.avx512.fpclass.ps.256(<8 x float>, i32)

define i8 @test_int_x86_avx512_fpclass_ps_256(<8 x float> %x0) {
; CHECK-LABEL: test_int_x86_avx512_fpclass_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfpclassps $4, %ymm0, %k1 # encoding: [0x62,0xf3,0x7d,0x28,0x66,0xc8,0x04]
; CHECK-NEXT:    vfpclassps $2, %ymm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x66,0xc0,0x02]
; CHECK-NEXT:    kmovw %k0, %eax # encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    # kill: def $al killed $al killed $eax
; CHECK-NEXT:    vzeroupper # encoding: [0xc5,0xf8,0x77]
; CHECK-NEXT:    ret{{[l|q]}} # encoding: [0xc3]
  %res = call <8 x i1> @llvm.x86.avx512.fpclass.ps.256(<8 x float> %x0, i32 2)
  %res1 = call <8 x i1> @llvm.x86.avx512.fpclass.ps.256(<8 x float> %x0, i32 4)
  %1 = and <8 x i1> %res1, %res
  %2 = bitcast <8 x i1> %1 to i8
  ret i8 %2
}

declare <2 x i1> @llvm.x86.avx512.fpclass.pd.128(<2 x double>, i32)

define i8 @test_int_x86_avx512_fpclass_pd_128(<2 x double> %x0) {
; CHECK-LABEL: test_int_x86_avx512_fpclass_pd_128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfpclasspd $2, %xmm0, %k1 # encoding: [0x62,0xf3,0xfd,0x08,0x66,0xc8,0x02]
; CHECK-NEXT:    vfpclasspd $4, %xmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x66,0xc0,0x04]
; CHECK-NEXT:    kmovw %k0, %eax # encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    # kill: def $al killed $al killed $eax
; CHECK-NEXT:    ret{{[l|q]}} # encoding: [0xc3]
  %res = call <2 x i1> @llvm.x86.avx512.fpclass.pd.128(<2 x double> %x0, i32 4)
  %res1 = call <2 x i1> @llvm.x86.avx512.fpclass.pd.128(<2 x double> %x0, i32 2)
  %1 = and <2 x i1> %res1, %res
  %2 = shufflevector <2 x i1> %1, <2 x i1> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 2, i32 3, i32 2, i32 3>
  %3 = bitcast <8 x i1> %2 to i8
  ret i8 %3
}

declare <4 x i1> @llvm.x86.avx512.fpclass.pd.256(<4 x double>, i32)

define i8 @test_int_x86_avx512_fpclass_pd_256(<4 x double> %x0) {
; CHECK-LABEL: test_int_x86_avx512_fpclass_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfpclasspd $4, %ymm0, %k1 # encoding: [0x62,0xf3,0xfd,0x28,0x66,0xc8,0x04]
; CHECK-NEXT:    vfpclasspd $2, %ymm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x66,0xc0,0x02]
; CHECK-NEXT:    kmovw %k0, %eax # encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    # kill: def $al killed $al killed $eax
; CHECK-NEXT:    vzeroupper # encoding: [0xc5,0xf8,0x77]
; CHECK-NEXT:    ret{{[l|q]}} # encoding: [0xc3]
  %res = call <4 x i1> @llvm.x86.avx512.fpclass.pd.256(<4 x double> %x0, i32 2)
  %res1 = call <4 x i1> @llvm.x86.avx512.fpclass.pd.256(<4 x double> %x0, i32 4)
  %1 = and <4 x i1> %res1, %res
  %2 = shufflevector <4 x i1> %1, <4 x i1> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %3 = bitcast <8 x i1> %2 to i8
  ret i8 %3
}