diff options
| -rw-r--r-- | llvm/lib/Target/X86/X86InstrSSE.td | 44 | ||||
| -rw-r--r-- | llvm/test/CodeGen/X86/avx-intrinsics-x86.ll | 1 |
2 files changed, 25 insertions, 20 deletions
diff --git a/llvm/lib/Target/X86/X86InstrSSE.td b/llvm/lib/Target/X86/X86InstrSSE.td index ee2331250c1..e91d9683bd1 100644 --- a/llvm/lib/Target/X86/X86InstrSSE.td +++ b/llvm/lib/Target/X86/X86InstrSSE.td @@ -6826,63 +6826,67 @@ multiclass SS48I_binop_rm2<bits<8> opc, string OpcodeStr, SDNode OpNode, } let Predicates = [HasAVX, NoVLX] in { - defm VPMINSB : SS48I_binop_rm<0x38, "vpminsb", smin, v16i8, VR128, - loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, - VEX_4V; defm VPMINSD : SS48I_binop_rm<0x39, "vpminsd", smin, v4i32, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; defm VPMINUD : SS48I_binop_rm<0x3B, "vpminud", umin, v4i32, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; - defm VPMINUW : SS48I_binop_rm<0x3A, "vpminuw", umin, v8i16, VR128, + defm VPMAXSD : SS48I_binop_rm<0x3D, "vpmaxsd", smax, v4i32, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; - defm VPMAXSB : SS48I_binop_rm<0x3C, "vpmaxsb", smax, v16i8, VR128, + defm VPMAXUD : SS48I_binop_rm<0x3F, "vpmaxud", umax, v4i32, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; - defm VPMAXSD : SS48I_binop_rm<0x3D, "vpmaxsd", smax, v4i32, VR128, + defm VPMULDQ : SS48I_binop_rm2<0x28, "vpmuldq", X86pmuldq, v2i64, v4i32, + VR128, loadv2i64, i128mem, + SSE_INTMUL_ITINS_P, 1, 0>, VEX_4V; +} +let Predicates = [HasAVX, NoVLX_Or_NoBWI] in { + defm VPMINSB : SS48I_binop_rm<0x38, "vpminsb", smin, v16i8, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; - defm VPMAXUD : SS48I_binop_rm<0x3F, "vpmaxud", umax, v4i32, VR128, + defm VPMINUW : SS48I_binop_rm<0x3A, "vpminuw", umin, v8i16, VR128, + loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, + VEX_4V; + defm VPMAXSB : SS48I_binop_rm<0x3C, "vpmaxsb", smax, v16i8, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; defm VPMAXUW : SS48I_binop_rm<0x3E, "vpmaxuw", umax, v8i16, VR128, loadv2i64, i128mem, 0, SSE_INTALU_ITINS_P>, VEX_4V; - defm VPMULDQ : SS48I_binop_rm2<0x28, "vpmuldq", X86pmuldq, v2i64, v4i32, - VR128, loadv2i64, i128mem, - SSE_INTMUL_ITINS_P, 1, 0>, VEX_4V; } let Predicates = [HasAVX2, NoVLX] in { - defm VPMINSBY : SS48I_binop_rm<0x38, "vpminsb", smin, v32i8, VR256, - loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, - VEX_4V, VEX_L; defm VPMINSDY : SS48I_binop_rm<0x39, "vpminsd", smin, v8i32, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; defm VPMINUDY : SS48I_binop_rm<0x3B, "vpminud", umin, v8i32, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; - defm VPMINUWY : SS48I_binop_rm<0x3A, "vpminuw", umin, v16i16, VR256, + defm VPMAXSDY : SS48I_binop_rm<0x3D, "vpmaxsd", smax, v8i32, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; - defm VPMAXSBY : SS48I_binop_rm<0x3C, "vpmaxsb", smax, v32i8, VR256, + defm VPMAXUDY : SS48I_binop_rm<0x3F, "vpmaxud", umax, v8i32, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; - defm VPMAXSDY : SS48I_binop_rm<0x3D, "vpmaxsd", smax, v8i32, VR256, + defm VPMULDQY : SS48I_binop_rm2<0x28, "vpmuldq", X86pmuldq, v4i64, v8i32, + VR256, loadv4i64, i256mem, + SSE_INTMUL_ITINS_P, 1, 0>, VEX_4V, VEX_L; +} +let Predicates = [HasAVX2, NoVLX] in { + defm VPMINSBY : SS48I_binop_rm<0x38, "vpminsb", smin, v32i8, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; - defm VPMAXUDY : SS48I_binop_rm<0x3F, "vpmaxud", umax, v8i32, VR256, + defm VPMINUWY : SS48I_binop_rm<0x3A, "vpminuw", umin, v16i16, VR256, + loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, + VEX_4V, VEX_L; + defm VPMAXSBY : SS48I_binop_rm<0x3C, "vpmaxsb", smax, v32i8, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; defm VPMAXUWY : SS48I_binop_rm<0x3E, "vpmaxuw", umax, v16i16, VR256, loadv4i64, i256mem, 0, SSE_INTALU_ITINS_P>, VEX_4V, VEX_L; - defm VPMULDQY : SS48I_binop_rm2<0x28, "vpmuldq", X86pmuldq, v4i64, v8i32, - VR256, loadv4i64, i256mem, - SSE_INTMUL_ITINS_P, 1, 0>, VEX_4V, VEX_L; } let Constraints = "$src1 = $dst" in { diff --git a/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll b/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll index f8d285866a7..b6205fa7586 100644 --- a/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll +++ b/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll @@ -1,5 +1,6 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=avx,aes,pclmul | FileCheck %s +; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=avx512vl,aes,pclmul | FileCheck %s define <2 x i64> @test_x86_aesni_aesdec(<2 x i64> %a0, <2 x i64> %a1) { ; CHECK-LABEL: test_x86_aesni_aesdec: |
