[SDNPHasChain, SDNPOutFlag]>;
def X86ucomi : SDNode<"X86ISD::UCOMI", SDTX86CmpTest,
[SDNPHasChain, SDNPOutFlag]>;
+def X86comi_new: SDNode<"X86ISD::COMI_NEW", SDTX86CmpTest,
+ [SDNPHasChain]>;
+def X86ucomi_new: SDNode<"X86ISD::UCOMI_NEW",SDTX86CmpTest>;
def X86s2vec : SDNode<"X86ISD::S2VEC", SDTypeProfile<1, 1, []>, []>;
def X86pextrw : SDNode<"X86ISD::PEXTRW", SDTypeProfile<1, 2, []>, []>;
def X86pinsrw : SDNode<"X86ISD::PINSRW", SDTypeProfile<1, 3, []>, []>;
Requires<[HasSSE1]>;
def IMPLICIT_DEF_FR32 : I<0, Pseudo, (outs FR32:$dst), (ins),
"#IMPLICIT_DEF $dst",
- [(set FR32:$dst, (undef))]>, Requires<[HasSSE2]>;
+ [(set FR32:$dst, (undef))]>, Requires<[HasSSE1]>;
def IMPLICIT_DEF_FR64 : I<0, Pseudo, (outs FR64:$dst), (ins),
"#IMPLICIT_DEF $dst",
[(set FR64:$dst, (undef))]>, Requires<[HasSSE2]>;
def memopv4i32 : PatFrag<(ops node:$ptr), (v4i32 (memop node:$ptr))>;
def memopv2i64 : PatFrag<(ops node:$ptr), (v2i64 (memop node:$ptr))>;
+// SSSE3 uses MMX registers for some instructions. They aren't aligned on a
+// 16-byte boundary.
+def memop64 : PatFrag<(ops node:$ptr), (ld node:$ptr), [{
+ if (LoadSDNode *LD = dyn_cast<LoadSDNode>(N))
+ return LD->getExtensionType() == ISD::NON_EXTLOAD &&
+ LD->getAddressingMode() == ISD::UNINDEXED &&
+ LD->getAlignment() >= 8;
+ return false;
+}]>;
+
+def memopv8i8 : PatFrag<(ops node:$ptr), (v8i8 (memop64 node:$ptr))>;
+def memopv16i8 : PatFrag<(ops node:$ptr), (v16i8 (memop64 node:$ptr))>;
+def memopv4i16 : PatFrag<(ops node:$ptr), (v4i16 (memop64 node:$ptr))>;
+def memopv8i16 : PatFrag<(ops node:$ptr), (v8i16 (memop64 node:$ptr))>;
+def memopv2i32 : PatFrag<(ops node:$ptr), (v2i32 (memop64 node:$ptr))>;
+
def bc_v4f32 : PatFrag<(ops node:$in), (v4f32 (bitconvert node:$in))>;
def bc_v2f64 : PatFrag<(ops node:$in), (v2f64 (bitconvert node:$in))>;
def bc_v16i8 : PatFrag<(ops node:$in), (v16i8 (bitconvert node:$in))>;
// CMOV* - Used to implement the SSE SELECT DAG operation. Expanded by the
// scheduler into a branch sequence.
-let usesCustomDAGSchedInserter = 1 in { // Expanded by the scheduler.
+// These are expanded by the scheduler.
+let Uses = [EFLAGS], usesCustomDAGSchedInserter = 1 in {
def CMOV_FR32 : I<0, Pseudo,
(outs FR32:$dst), (ins FR32:$t, FR32:$f, i8imm:$cond),
"#CMOV_FR32 PSEUDO!",
"#CMOV_V2I64 PSEUDO!",
[(set VR128:$dst,
(v2i64 (X86cmov VR128:$t, VR128:$f, imm:$cond)))]>;
+
+ def NEW_CMOV_FR32 : I<0, Pseudo,
+ (outs FR32:$dst), (ins FR32:$t, FR32:$f, i8imm:$cond),
+ "#CMOV_FR32 PSEUDO!",
+ [(set FR32:$dst, (X86cmov_new FR32:$t, FR32:$f, imm:$cond,
+ EFLAGS))]>;
+ def NEW_CMOV_FR64 : I<0, Pseudo,
+ (outs FR64:$dst), (ins FR64:$t, FR64:$f, i8imm:$cond),
+ "#CMOV_FR64 PSEUDO!",
+ [(set FR64:$dst, (X86cmov_new FR64:$t, FR64:$f, imm:$cond,
+ EFLAGS))]>;
+ def NEW_CMOV_V4F32 : I<0, Pseudo,
+ (outs VR128:$dst), (ins VR128:$t, VR128:$f, i8imm:$cond),
+ "#CMOV_V4F32 PSEUDO!",
+ [(set VR128:$dst,
+ (v4f32 (X86cmov_new VR128:$t, VR128:$f, imm:$cond,
+ EFLAGS)))]>;
+ def NEW_CMOV_V2F64 : I<0, Pseudo,
+ (outs VR128:$dst), (ins VR128:$t, VR128:$f, i8imm:$cond),
+ "#CMOV_V2F64 PSEUDO!",
+ [(set VR128:$dst,
+ (v2f64 (X86cmov_new VR128:$t, VR128:$f, imm:$cond,
+ EFLAGS)))]>;
+ def NEW_CMOV_V2I64 : I<0, Pseudo,
+ (outs VR128:$dst), (ins VR128:$t, VR128:$f, i8imm:$cond),
+ "#CMOV_V2I64 PSEUDO!",
+ [(set VR128:$dst,
+ (v2i64 (X86cmov_new VR128:$t, VR128:$f, imm:$cond,
+ EFLAGS)))]>;
}
//===----------------------------------------------------------------------===//
// Move Instructions
def MOVSSrr : SSI<0x10, MRMSrcReg, (outs FR32:$dst), (ins FR32:$src),
"movss\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1, isReMaterializable = 1 in
def MOVSSrm : SSI<0x10, MRMSrcMem, (outs FR32:$dst), (ins f32mem:$src),
"movss\t{$src, $dst|$dst, $src}",
[(set FR32:$dst, (loadf32 addr:$src))]>;
"cmp${cc}ss\t{$src, $dst|$dst, $src}", []>;
}
+let Defs = [EFLAGS] in {
def UCOMISSrr: PSI<0x2E, MRMSrcReg, (outs), (ins FR32:$src1, FR32:$src2),
"ucomiss\t{$src2, $src1|$src1, $src2}",
[(X86cmp FR32:$src1, FR32:$src2)]>;
"ucomiss\t{$src2, $src1|$src1, $src2}",
[(X86cmp FR32:$src1, (loadf32 addr:$src2))]>;
+def NEW_UCOMISSrr: PSI<0x2E, MRMSrcReg, (outs), (ins FR32:$src1, FR32:$src2),
+ "ucomiss\t{$src2, $src1|$src1, $src2}",
+ [(X86cmp_new FR32:$src1, FR32:$src2), (implicit EFLAGS)]>;
+def NEW_UCOMISSrm: PSI<0x2E, MRMSrcMem, (outs), (ins FR32:$src1, f32mem:$src2),
+ "ucomiss\t{$src2, $src1|$src1, $src2}",
+ [(X86cmp_new FR32:$src1, (loadf32 addr:$src2)),
+ (implicit EFLAGS)]>;
+} // Defs = [EFLAGS]
+
// Aliases to match intrinsics which expect XMM operand(s).
let isTwoAddress = 1 in {
def Int_CMPSSrr : SSI<0xC2, MRMSrcReg,
(load addr:$src), imm:$cc))]>;
}
+let Defs = [EFLAGS] in {
def Int_UCOMISSrr: PSI<0x2E, MRMSrcReg, (outs), (ins VR128:$src1, VR128:$src2),
"ucomiss\t{$src2, $src1|$src1, $src2}",
[(X86ucomi (v4f32 VR128:$src1), VR128:$src2)]>;
"comiss\t{$src2, $src1|$src1, $src2}",
[(X86comi (v4f32 VR128:$src1), (load addr:$src2))]>;
+def NEW_Int_UCOMISSrr: PSI<0x2E, MRMSrcReg, (outs),
+ (ins VR128:$src1, VR128:$src2),
+ "ucomiss\t{$src2, $src1|$src1, $src2}",
+ [(X86ucomi_new (v4f32 VR128:$src1), VR128:$src2),
+ (implicit EFLAGS)]>;
+def NEW_Int_UCOMISSrm: PSI<0x2E, MRMSrcMem, (outs),
+ (ins VR128:$src1, f128mem:$src2),
+ "ucomiss\t{$src2, $src1|$src1, $src2}",
+ [(X86ucomi_new (v4f32 VR128:$src1), (load addr:$src2)),
+ (implicit EFLAGS)]>;
+
+def NEW_Int_COMISSrr: PSI<0x2F, MRMSrcReg, (outs),
+ (ins VR128:$src1, VR128:$src2),
+ "comiss\t{$src2, $src1|$src1, $src2}",
+ [(X86comi_new (v4f32 VR128:$src1), VR128:$src2),
+ (implicit EFLAGS)]>;
+def NEW_Int_COMISSrm: PSI<0x2F, MRMSrcMem, (outs),
+ (ins VR128:$src1, f128mem:$src2),
+ "comiss\t{$src2, $src1|$src1, $src2}",
+ [(X86comi_new (v4f32 VR128:$src1), (load addr:$src2)),
+ (implicit EFLAGS)]>;
+} // Defs = [EFLAGS]
+
// Aliases of packed SSE1 instructions for scalar use. These all have names that
// start with 'Fs'.
// Alias instructions that map fld0 to pxor for sse.
+let isReMaterializable = 1 in
def FsFLD0SS : I<0xEF, MRMInitReg, (outs FR32:$dst), (ins),
"pxor\t$dst, $dst", [(set FR32:$dst, fp32imm0)]>,
Requires<[HasSSE1]>, TB, OpSize;
// Alias instruction to load FR32 from f128mem using movaps. Upper bits are
// disregarded.
+let isLoad = 1 in
def FsMOVAPSrm : PSI<0x28, MRMSrcMem, (outs FR32:$dst), (ins f128mem:$src),
"movaps\t{$src, $dst|$dst, $src}",
[(set FR32:$dst, (alignedloadfsf32 addr:$src))]>;
}
// Vector intrinsic operation, reg+mem.
- def PSrm_Int : PSI<opc, MRMSrcMem, (outs VR128:$dst), (ins VR128:$src1, f32mem:$src2),
+ def PSrm_Int : PSI<opc, MRMSrcMem, (outs VR128:$dst), (ins VR128:$src1, f128mem:$src2),
!strconcat(OpcodeStr, "ps\t{$src2, $dst|$dst, $src2}"),
[(set VR128:$dst, (V4F32Int VR128:$src1, (load addr:$src2)))]>;
}
// Move Instructions
def MOVAPSrr : PSI<0x28, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),
"movaps\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1, isReMaterializable = 1 in
def MOVAPSrm : PSI<0x28, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
"movaps\t{$src, $dst|$dst, $src}",
[(set VR128:$dst, (alignedloadv4f32 addr:$src))]>;
def MOVUPSrr : PSI<0x10, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),
"movups\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1 in
def MOVUPSrm : PSI<0x10, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
"movups\t{$src, $dst|$dst, $src}",
[(set VR128:$dst, (loadv4f32 addr:$src))]>;
[(store (v4f32 VR128:$src), addr:$dst)]>;
// Intrinsic forms of MOVUPS load and store
+let isLoad = 1 in
def MOVUPSrm_Int : PSI<0x10, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
"movups\t{$src, $dst|$dst, $src}",
[(set VR128:$dst, (int_x86_sse_loadu_ps addr:$src))]>;
}
// Vector intrinsic operation, mem
- def PSm_Int : PSI<opc, MRMSrcMem, (outs VR128:$dst), (ins f32mem:$src),
+ def PSm_Int : PSI<opc, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
!strconcat(OpcodeStr, "ps\t{$src, $dst|$dst, $src}"),
[(set VR128:$dst, (V4F32Int (load addr:$src)))]>;
}
"shufps\t{$src3, $src2, $dst|$dst, $src2, $src3}",
[(set VR128:$dst,
(v4f32 (vector_shuffle
- VR128:$src1, (load addr:$src2),
+ VR128:$src1, (memopv4f32 addr:$src2),
SHUFP_shuffle_mask:$src3)))]>;
let AddedComplexity = 10 in {
"unpckhps\t{$src2, $dst|$dst, $src2}",
[(set VR128:$dst,
(v4f32 (vector_shuffle
- VR128:$src1, (load addr:$src2),
+ VR128:$src1, (memopv4f32 addr:$src2),
UNPCKH_shuffle_mask)))]>;
def UNPCKLPSrr : PSI<0x14, MRMSrcReg,
"unpcklps\t{$src2, $dst|$dst, $src2}",
[(set VR128:$dst,
(v4f32 (vector_shuffle
- VR128:$src1, (load addr:$src2),
+ VR128:$src1, (memopv4f32 addr:$src2),
UNPCKL_shuffle_mask)))]>;
} // AddedComplexity
} // isTwoAddress
// Move Instructions
def MOVSDrr : SDI<0x10, MRMSrcReg, (outs FR64:$dst), (ins FR64:$src),
"movsd\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1, isReMaterializable = 1 in
def MOVSDrm : SDI<0x10, MRMSrcMem, (outs FR64:$dst), (ins f64mem:$src),
"movsd\t{$src, $dst|$dst, $src}",
[(set FR64:$dst, (loadf64 addr:$src))]>;
"cmp${cc}sd\t{$src, $dst|$dst, $src}", []>;
}
+let Defs = [EFLAGS] in {
def UCOMISDrr: PDI<0x2E, MRMSrcReg, (outs), (ins FR64:$src1, FR64:$src2),
"ucomisd\t{$src2, $src1|$src1, $src2}",
[(X86cmp FR64:$src1, FR64:$src2)]>;
"ucomisd\t{$src2, $src1|$src1, $src2}",
[(X86cmp FR64:$src1, (loadf64 addr:$src2))]>;
+def NEW_UCOMISDrr: PDI<0x2E, MRMSrcReg, (outs), (ins FR64:$src1, FR64:$src2),
+ "ucomisd\t{$src2, $src1|$src1, $src2}",
+ [(X86cmp_new FR64:$src1, FR64:$src2), (implicit EFLAGS)]>;
+def NEW_UCOMISDrm: PDI<0x2E, MRMSrcMem, (outs), (ins FR64:$src1, f64mem:$src2),
+ "ucomisd\t{$src2, $src1|$src1, $src2}",
+ [(X86cmp_new FR64:$src1, (loadf64 addr:$src2)),
+ (implicit EFLAGS)]>;
+}
+
// Aliases to match intrinsics which expect XMM operand(s).
let isTwoAddress = 1 in {
def Int_CMPSDrr : SDI<0xC2, MRMSrcReg,
(load addr:$src), imm:$cc))]>;
}
+let Defs = [EFLAGS] in {
def Int_UCOMISDrr: PDI<0x2E, MRMSrcReg, (outs), (ins VR128:$src1, VR128:$src2),
"ucomisd\t{$src2, $src1|$src1, $src2}",
[(X86ucomi (v2f64 VR128:$src1), (v2f64 VR128:$src2))]>;
"comisd\t{$src2, $src1|$src1, $src2}",
[(X86comi (v2f64 VR128:$src1), (load addr:$src2))]>;
+def NEW_Int_UCOMISDrr: PDI<0x2E, MRMSrcReg, (outs),
+ (ins VR128:$src1, VR128:$src2),
+ "ucomisd\t{$src2, $src1|$src1, $src2}",
+ [(X86ucomi_new (v2f64 VR128:$src1), (v2f64 VR128:$src2)),
+ (implicit EFLAGS)]>;
+def NEW_Int_UCOMISDrm: PDI<0x2E, MRMSrcMem, (outs),
+ (ins VR128:$src1, f128mem:$src2),
+ "ucomisd\t{$src2, $src1|$src1, $src2}",
+ [(X86ucomi_new (v2f64 VR128:$src1), (load addr:$src2)),
+ (implicit EFLAGS)]>;
+
+def NEW_Int_COMISDrr: PDI<0x2F, MRMSrcReg, (outs),
+ (ins VR128:$src1, VR128:$src2),
+ "comisd\t{$src2, $src1|$src1, $src2}",
+ [(X86comi_new (v2f64 VR128:$src1), (v2f64 VR128:$src2)),
+ (implicit EFLAGS)]>;
+def NEW_Int_COMISDrm: PDI<0x2F, MRMSrcMem, (outs),
+ (ins VR128:$src1, f128mem:$src2),
+ "comisd\t{$src2, $src1|$src1, $src2}",
+ [(X86comi_new (v2f64 VR128:$src1), (load addr:$src2)),
+ (implicit EFLAGS)]>;
+} // Defs = EFLAGS]
+
// Aliases of packed SSE2 instructions for scalar use. These all have names that
// start with 'Fs'.
// Alias instructions that map fld0 to pxor for sse.
+let isReMaterializable = 1 in
def FsFLD0SD : I<0xEF, MRMInitReg, (outs FR64:$dst), (ins),
"pxor\t$dst, $dst", [(set FR64:$dst, fpimm0)]>,
Requires<[HasSSE2]>, TB, OpSize;
// Alias instruction to load FR64 from f128mem using movapd. Upper bits are
// disregarded.
+let isLoad = 1 in
def FsMOVAPDrm : PDI<0x28, MRMSrcMem, (outs FR64:$dst), (ins f128mem:$src),
"movapd\t{$src, $dst|$dst, $src}",
[(set FR64:$dst, (alignedloadfsf64 addr:$src))]>;
}
// Vector intrinsic operation, reg+mem.
- def PDrm_Int : PDI<opc, MRMSrcMem, (outs VR128:$dst), (ins VR128:$src1, f64mem:$src2),
+ def PDrm_Int : PDI<opc, MRMSrcMem, (outs VR128:$dst), (ins VR128:$src1, f128mem:$src2),
!strconcat(OpcodeStr, "pd\t{$src2, $dst|$dst, $src2}"),
[(set VR128:$dst, (V2F64Int VR128:$src1, (load addr:$src2)))]>;
}
// Move Instructions
def MOVAPDrr : PDI<0x28, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),
"movapd\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1, isReMaterializable = 1 in
def MOVAPDrm : PDI<0x28, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
"movapd\t{$src, $dst|$dst, $src}",
[(set VR128:$dst, (alignedloadv2f64 addr:$src))]>;
def MOVUPDrr : PDI<0x10, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),
"movupd\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1 in
def MOVUPDrm : PDI<0x10, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
"movupd\t{$src, $dst|$dst, $src}",
[(set VR128:$dst, (loadv2f64 addr:$src))]>;
}
// Vector intrinsic operation, mem
- def PDm_Int : PDI<opc, MRMSrcMem, (outs VR128:$dst), (ins f64mem:$src),
+ def PDm_Int : PDI<opc, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src),
!strconcat(OpcodeStr, "pd\t{$src, $dst|$dst, $src}"),
[(set VR128:$dst, (V2F64Int (load addr:$src)))]>;
}
"shufpd\t{$src3, $src2, $dst|$dst, $src2, $src3}",
[(set VR128:$dst,
(v2f64 (vector_shuffle
- VR128:$src1, (load addr:$src2),
+ VR128:$src1, (memopv2f64 addr:$src2),
SHUFP_shuffle_mask:$src3)))]>;
let AddedComplexity = 10 in {
"unpckhpd\t{$src2, $dst|$dst, $src2}",
[(set VR128:$dst,
(v2f64 (vector_shuffle
- VR128:$src1, (load addr:$src2),
+ VR128:$src1, (memopv2f64 addr:$src2),
UNPCKH_shuffle_mask)))]>;
def UNPCKLPDrr : PDI<0x14, MRMSrcReg,
"unpcklpd\t{$src2, $dst|$dst, $src2}",
[(set VR128:$dst,
(v2f64 (vector_shuffle
- VR128:$src1, (load addr:$src2),
+ VR128:$src1, (memopv2f64 addr:$src2),
UNPCKL_shuffle_mask)))]>;
} // AddedComplexity
} // isTwoAddress
// Move Instructions
def MOVDQArr : PDI<0x6F, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),
"movdqa\t{$src, $dst|$dst, $src}", []>;
+let isLoad = 1 in
def MOVDQArm : PDI<0x6F, MRMSrcMem, (outs VR128:$dst), (ins i128mem:$src),
"movdqa\t{$src, $dst|$dst, $src}",
[/*(set VR128:$dst, (alignedloadv2i64 addr:$src))*/]>;
def MOVDQAmr : PDI<0x7F, MRMDestMem, (outs), (ins i128mem:$dst, VR128:$src),
"movdqa\t{$src, $dst|$dst, $src}",
[/*(alignedstore (v2i64 VR128:$src), addr:$dst)*/]>;
+let isLoad = 1 in
def MOVDQUrm : I<0x6F, MRMSrcMem, (outs VR128:$dst), (ins i128mem:$src),
"movdqu\t{$src, $dst|$dst, $src}",
[/*(set VR128:$dst, (loadv2i64 addr:$src))*/]>,
XS, Requires<[HasSSE2]>;
// Intrinsic forms of MOVDQU load and store
+let isLoad = 1 in
def MOVDQUrm_Int : I<0x6F, MRMSrcMem, (outs VR128:$dst), (ins i128mem:$src),
"movdqu\t{$src, $dst|$dst, $src}",
[(set VR128:$dst, (int_x86_sse2_loadu_dq addr:$src))]>,
(outs VR128:$dst), (ins VR128:$src1, i128mem:$src2),
"pandn\t{$src2, $dst|$dst, $src2}",
[(set VR128:$dst, (v2i64 (and (vnot VR128:$src1),
- (load addr:$src2))))]>;
+ (memopv2i64 addr:$src2))))]>;
}
// SSE2 Integer comparison
[(set GR32:$dst, (int_x86_sse2_pmovmskb_128 VR128:$src))]>;
// Conditional store
+let Uses = [EDI] in
def MASKMOVDQU : PDI<0xF7, MRMSrcReg, (outs), (ins VR128:$src, VR128:$mask),
"maskmovdqu\t{$mask, $src|$src, $mask}",
- [(int_x86_sse2_maskmov_dqu VR128:$src, VR128:$mask, EDI)]>,
- Imp<[EDI],[]>;
+ [(int_x86_sse2_maskmov_dqu VR128:$src, VR128:$mask, EDI)]>;
// Non-temporal stores
def MOVNTPDmr : PDI<0x2B, MRMDestMem, (outs), (ins i128mem:$dst, VR128:$src),
// SSSE3 Instructions
//===----------------------------------------------------------------------===//
-// SSE3 Instruction Templates:
+// SSSE3 Instruction Templates:
//
-// SS38I - SSSE3 instructions with T8 and OpSize prefixes.
-// SS3AI - SSSE3 instructions with TA and OpSize prefixes.
+// SS38I - SSSE3 instructions with T8 prefix.
+// SS3AI - SSSE3 instructions with TA prefix.
+//
+// Note: SSSE3 instructions have 64-bit and 128-bit versions. The 64-bit version
+// uses the MMX registers. We put those instructions here because they better
+// fit into the SSSE3 instruction category rather than the MMX category.
class SS38I<bits<8> o, Format F, dag outs, dag ins, string asm,
list<dag> pattern>
- : I<o, F, outs, ins, asm, pattern>, T8, OpSize, Requires<[HasSSSE3]>;
+ : I<o, F, outs, ins, asm, pattern>, T8, Requires<[HasSSSE3]>;
class SS3AI<bits<8> o, Format F, dag outs, dag ins, string asm,
list<dag> pattern>
- : I<o, F, outs, ins, asm, pattern>, TA, OpSize, Requires<[HasSSSE3]>;
+ : I<o, F, outs, ins, asm, pattern>, TA, Requires<[HasSSSE3]>;
+
+/// SS3I_unop_rm_int_8 - Simple SSSE3 unary operator whose type is v*i8.
+let isTwoAddress = 1 in {
+ multiclass SS3I_unop_rm_int_8<bits<8> opc, string OpcodeStr,
+ Intrinsic IntId64, Intrinsic IntId128,
+ bit Commutable = 0> {
+ def rr64 : SS38I<opc, MRMSrcReg, (outs VR64:$dst), (ins VR64:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR64:$dst, (IntId64 VR64:$src))]> {
+ let isCommutable = Commutable;
+ }
+ def rm64 : SS38I<opc, MRMSrcMem, (outs VR64:$dst), (ins i64mem:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR64:$dst,
+ (IntId64 (bitconvert (memopv8i8 addr:$src))))]>;
+
+ def rr128 : SS38I<opc, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR128:$dst, (IntId128 VR128:$src))]>,
+ OpSize {
+ let isCommutable = Commutable;
+ }
+ def rm128 : SS38I<opc, MRMSrcMem, (outs VR128:$dst),
+ (ins i128mem:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR128:$dst,
+ (IntId128
+ (bitconvert (memopv16i8 addr:$src))))]>, OpSize;
+ }
+}
-/// SS3I_binop_rm_int - Simple SSSE3 binary operatr whose type is v2i64.
+/// SS3I_unop_rm_int_16 - Simple SSSE3 unary operator whose type is v*i16.
let isTwoAddress = 1 in {
- multiclass SS3I_binop_rm_int<bits<8> opc, string OpcodeStr, Intrinsic IntId,
- bit Commutable = 0> {
- def rr : SS38I<opc, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src1, VR128:$src2),
- !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
- [(set VR128:$dst, (IntId VR128:$src1, VR128:$src2))]> {
+ multiclass SS3I_unop_rm_int_16<bits<8> opc, string OpcodeStr,
+ Intrinsic IntId64, Intrinsic IntId128,
+ bit Commutable = 0> {
+ def rr64 : SS38I<opc, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR64:$dst, (IntId64 VR64:$src))]> {
let isCommutable = Commutable;
}
- def rm : SS38I<opc, MRMSrcMem, (outs VR128:$dst), (ins VR128:$src1, i128mem:$src2),
- !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
- [(set VR128:$dst,
- (IntId VR128:$src1,
- (bitconvert (memopv2i64 addr:$src2))))]>;
+ def rm64 : SS38I<opc, MRMSrcMem, (outs VR64:$dst),
+ (ins i64mem:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR64:$dst,
+ (IntId64
+ (bitconvert (memopv4i16 addr:$src))))]>;
+
+ def rr128 : SS38I<opc, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR128:$dst, (IntId128 VR128:$src))]>,
+ OpSize {
+ let isCommutable = Commutable;
+ }
+ def rm128 : SS38I<opc, MRMSrcMem, (outs VR128:$dst),
+ (ins i128mem:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR128:$dst,
+ (IntId128
+ (bitconvert (memopv8i16 addr:$src))))]>, OpSize;
}
}
-defm PMULHRSW128 : SS3I_binop_rm_int<0x0B, "pmulhrsw",
- int_x86_ssse3_pmulhrsw_128, 1>;
+/// SS3I_unop_rm_int_32 - Simple SSSE3 unary operator whose type is v*i32.
+let isTwoAddress = 1 in {
+ multiclass SS3I_unop_rm_int_32<bits<8> opc, string OpcodeStr,
+ Intrinsic IntId64, Intrinsic IntId128,
+ bit Commutable = 0> {
+ def rr64 : SS38I<opc, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR64:$dst, (IntId64 VR64:$src))]> {
+ let isCommutable = Commutable;
+ }
+ def rm64 : SS38I<opc, MRMSrcMem, (outs VR64:$dst),
+ (ins i64mem:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR64:$dst,
+ (IntId64
+ (bitconvert (memopv2i32 addr:$src))))]>;
+
+ def rr128 : SS38I<opc, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR128:$dst, (IntId128 VR128:$src))]>,
+ OpSize {
+ let isCommutable = Commutable;
+ }
+ def rm128 : SS38I<opc, MRMSrcMem, (outs VR128:$dst),
+ (ins i128mem:$src),
+ !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
+ [(set VR128:$dst,
+ (IntId128
+ (bitconvert (memopv4i32 addr:$src))))]>, OpSize;
+ }
+}
+
+defm PABSB : SS3I_unop_rm_int_8 <0x1C, "pabsb",
+ int_x86_ssse3_pabs_b,
+ int_x86_ssse3_pabs_b_128>;
+defm PABSW : SS3I_unop_rm_int_16<0x1D, "pabsw",
+ int_x86_ssse3_pabs_w,
+ int_x86_ssse3_pabs_w_128>;
+defm PABSD : SS3I_unop_rm_int_32<0x1E, "pabsd",
+ int_x86_ssse3_pabs_d,
+ int_x86_ssse3_pabs_d_128>;
+
+/// SS3I_binop_rm_int_8 - Simple SSSE3 binary operator whose type is v*i8.
+let isTwoAddress = 1 in {
+ multiclass SS3I_binop_rm_int_8<bits<8> opc, string OpcodeStr,
+ Intrinsic IntId64, Intrinsic IntId128,
+ bit Commutable = 0> {
+ def rr64 : SS38I<opc, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src1, VR64:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR64:$dst, (IntId64 VR64:$src1, VR64:$src2))]> {
+ let isCommutable = Commutable;
+ }
+ def rm64 : SS38I<opc, MRMSrcMem, (outs VR64:$dst),
+ (ins VR64:$src1, i64mem:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR64:$dst,
+ (IntId64 VR64:$src1,
+ (bitconvert (memopv8i8 addr:$src2))))]>;
+
+ def rr128 : SS38I<opc, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src1, VR128:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR128:$dst, (IntId128 VR128:$src1, VR128:$src2))]>,
+ OpSize {
+ let isCommutable = Commutable;
+ }
+ def rm128 : SS38I<opc, MRMSrcMem, (outs VR128:$dst),
+ (ins VR128:$src1, i128mem:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR128:$dst,
+ (IntId128 VR128:$src1,
+ (bitconvert (memopv16i8 addr:$src2))))]>, OpSize;
+ }
+}
+
+/// SS3I_binop_rm_int_16 - Simple SSSE3 binary operator whose type is v*i16.
+let isTwoAddress = 1 in {
+ multiclass SS3I_binop_rm_int_16<bits<8> opc, string OpcodeStr,
+ Intrinsic IntId64, Intrinsic IntId128,
+ bit Commutable = 0> {
+ def rr64 : SS38I<opc, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src1, VR64:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR64:$dst, (IntId64 VR64:$src1, VR64:$src2))]> {
+ let isCommutable = Commutable;
+ }
+ def rm64 : SS38I<opc, MRMSrcMem, (outs VR64:$dst),
+ (ins VR64:$src1, i64mem:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR64:$dst,
+ (IntId64 VR64:$src1,
+ (bitconvert (memopv4i16 addr:$src2))))]>;
+
+ def rr128 : SS38I<opc, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src1, VR128:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR128:$dst, (IntId128 VR128:$src1, VR128:$src2))]>,
+ OpSize {
+ let isCommutable = Commutable;
+ }
+ def rm128 : SS38I<opc, MRMSrcMem, (outs VR128:$dst),
+ (ins VR128:$src1, i128mem:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR128:$dst,
+ (IntId128 VR128:$src1,
+ (bitconvert (memopv8i16 addr:$src2))))]>, OpSize;
+ }
+}
+
+/// SS3I_binop_rm_int_32 - Simple SSSE3 binary operator whose type is v*i32.
+let isTwoAddress = 1 in {
+ multiclass SS3I_binop_rm_int_32<bits<8> opc, string OpcodeStr,
+ Intrinsic IntId64, Intrinsic IntId128,
+ bit Commutable = 0> {
+ def rr64 : SS38I<opc, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src1, VR64:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR64:$dst, (IntId64 VR64:$src1, VR64:$src2))]> {
+ let isCommutable = Commutable;
+ }
+ def rm64 : SS38I<opc, MRMSrcMem, (outs VR64:$dst),
+ (ins VR64:$src1, i64mem:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR64:$dst,
+ (IntId64 VR64:$src1,
+ (bitconvert (memopv2i32 addr:$src2))))]>;
+
+ def rr128 : SS38I<opc, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src1, VR128:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR128:$dst, (IntId128 VR128:$src1, VR128:$src2))]>,
+ OpSize {
+ let isCommutable = Commutable;
+ }
+ def rm128 : SS38I<opc, MRMSrcMem, (outs VR128:$dst),
+ (ins VR128:$src1, i128mem:$src2),
+ !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
+ [(set VR128:$dst,
+ (IntId128 VR128:$src1,
+ (bitconvert (memopv4i32 addr:$src2))))]>, OpSize;
+ }
+}
+
+defm PHADDW : SS3I_binop_rm_int_16<0x01, "phaddw",
+ int_x86_ssse3_phadd_w,
+ int_x86_ssse3_phadd_w_128, 1>;
+defm PHADDD : SS3I_binop_rm_int_32<0x02, "phaddd",
+ int_x86_ssse3_phadd_d,
+ int_x86_ssse3_phadd_d_128, 1>;
+defm PHADDSW : SS3I_binop_rm_int_16<0x03, "phaddsw",
+ int_x86_ssse3_phadd_sw,
+ int_x86_ssse3_phadd_sw_128, 1>;
+defm PHSUBW : SS3I_binop_rm_int_16<0x05, "phsubw",
+ int_x86_ssse3_phsub_w,
+ int_x86_ssse3_phsub_w_128>;
+defm PHSUBD : SS3I_binop_rm_int_32<0x06, "phsubd",
+ int_x86_ssse3_phsub_d,
+ int_x86_ssse3_phsub_d_128>;
+defm PHSUBSW : SS3I_binop_rm_int_16<0x07, "phsubsw",
+ int_x86_ssse3_phsub_sw,
+ int_x86_ssse3_phsub_sw_128>;
+defm PMADDUBSW : SS3I_binop_rm_int_8 <0x04, "pmaddubsw",
+ int_x86_ssse3_pmadd_ub_sw,
+ int_x86_ssse3_pmadd_ub_sw_128, 1>;
+defm PMULHRSW : SS3I_binop_rm_int_16<0x0B, "pmulhrsw",
+ int_x86_ssse3_pmul_hr_sw,
+ int_x86_ssse3_pmul_hr_sw_128, 1>;
+defm PSHUFB : SS3I_binop_rm_int_8 <0x00, "pshufb",
+ int_x86_ssse3_pshuf_b,
+ int_x86_ssse3_pshuf_b_128>;
+defm PSIGNB : SS3I_binop_rm_int_8 <0x08, "psignb",
+ int_x86_ssse3_psign_b,
+ int_x86_ssse3_psign_b_128>;
+defm PSIGNW : SS3I_binop_rm_int_16<0x09, "psignw",
+ int_x86_ssse3_psign_w,
+ int_x86_ssse3_psign_w_128>;
+defm PSIGND : SS3I_binop_rm_int_32<0x09, "psignd",
+ int_x86_ssse3_psign_d,
+ int_x86_ssse3_psign_d_128>;
+
+let isTwoAddress = 1 in {
+ def PALIGNR64rr : SS3AI<0x0F, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src1, VR64:$src2, i16imm:$src3),
+ "palignr\t{$src2, $dst|$dst, $src2}",
+ [(set VR64:$dst,
+ (int_x86_ssse3_palign_r
+ VR64:$src1, VR64:$src2,
+ imm:$src3))]>;
+ def PALIGNR64rm : SS3AI<0x0F, MRMSrcReg, (outs VR64:$dst),
+ (ins VR64:$src1, i64mem:$src2, i16imm:$src3),
+ "palignr\t{$src2, $dst|$dst, $src2}",
+ [(set VR64:$dst,
+ (int_x86_ssse3_palign_r
+ VR64:$src1,
+ (bitconvert (memopv2i32 addr:$src2)),
+ imm:$src3))]>;
+
+ def PALIGNR128rr : SS3AI<0x0F, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src1, VR128:$src2, i32imm:$src3),
+ "palignr\t{$src2, $dst|$dst, $src2}",
+ [(set VR128:$dst,
+ (int_x86_ssse3_palign_r_128
+ VR128:$src1, VR128:$src2,
+ imm:$src3))]>, OpSize;
+ def PALIGNR128rm : SS3AI<0x0F, MRMSrcReg, (outs VR128:$dst),
+ (ins VR128:$src1, i128mem:$src2, i32imm:$src3),
+ "palignr\t{$src2, $dst|$dst, $src2}",
+ [(set VR128:$dst,
+ (int_x86_ssse3_palign_r_128
+ VR128:$src1,
+ (bitconvert (memopv4i32 addr:$src2)),
+ imm:$src3))]>, OpSize;
+}
//===----------------------------------------------------------------------===//
// Non-Instruction Patterns
//===----------------------------------------------------------------------===//
// 128-bit vector undef's.
+def : Pat<(v4f32 (undef)), (IMPLICIT_DEF_VR128)>, Requires<[HasSSE2]>;
def : Pat<(v2f64 (undef)), (IMPLICIT_DEF_VR128)>, Requires<[HasSSE2]>;
def : Pat<(v16i8 (undef)), (IMPLICIT_DEF_VR128)>, Requires<[HasSSE2]>;
def : Pat<(v8i16 (undef)), (IMPLICIT_DEF_VR128)>, Requires<[HasSSE2]>;
SHUFP_unary_shuffle_mask:$sm),
(SHUFPSrri VR128:$src1, VR128:$src1, SHUFP_unary_shuffle_mask:$sm)>,
Requires<[HasSSE1]>;
+// Special unary SHUFPDrri case.
+def : Pat<(vector_shuffle (v2f64 VR128:$src1), (undef),
+ SHUFP_unary_shuffle_mask:$sm),
+ (SHUFPDrri VR128:$src1, VR128:$src1, SHUFP_unary_shuffle_mask:$sm)>,
+ Requires<[HasSSE2]>;
// Unary v4f32 shuffle with PSHUF* in order to fold a load.
def : Pat<(vector_shuffle (memopv4f32 addr:$src1), (undef),
SHUFP_unary_shuffle_mask:$sm),
(PANDNrr VR128:$src1, VR128:$src2)>, Requires<[HasSSE2]>;
def : Pat<(v2i64 (and (xor VR128:$src1, (bc_v2i64 (v4i32 immAllOnesV))),
- (load addr:$src2))),
+ (memopv2i64 addr:$src2))),
(PANDNrm VR128:$src1, addr:$src2)>, Requires<[HasSSE2]>;
def : Pat<(v2i64 (and (xor VR128:$src1, (bc_v2i64 (v8i16 immAllOnesV))),
- (load addr:$src2))),
+ (memopv2i64 addr:$src2))),
(PANDNrm VR128:$src1, addr:$src2)>, Requires<[HasSSE2]>;
def : Pat<(v2i64 (and (xor VR128:$src1, (bc_v2i64 (v16i8 immAllOnesV))),
- (load addr:$src2))),
+ (memopv2i64 addr:$src2))),
(PANDNrm VR128:$src1, addr:$src2)>, Requires<[HasSSE2]>;
// Use movaps / movups for SSE integer load / store (one byte shorter).
// (vextract (v4i32 bc (v4f32 s2v (f32 load $addr))), 0) -> (i32 load $addr)
def : Pat<(vector_extract
(bc_v4i32 (v4f32 (scalar_to_vector (loadf32 addr:$src)))), (iPTR 0)),
- (MOV32rm addr:$src)>;
+ (MOV32rm addr:$src)>, Requires<[HasSSE2]>;
def : Pat<(vector_extract
(bc_v2i64 (v2f64 (scalar_to_vector (loadf64 addr:$src)))), (iPTR 0)),
- (MOV64rm addr:$src)>;
+ (MOV64rm addr:$src)>, Requires<[HasSSE2, In64BitMode]>;