maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[x86] Implement a faster vector population count based on the PSHUFB 

in-register LUT technique.

Summary:
A description of this technique can be found here:
http://wm.ite.pl/articles/sse-popcount.html

The core of the idea is to use an in-register lookup table and the
PSHUFB instruction to compute the population count for the low and high
nibbles of each byte, and then to use horizontal sums to aggregate these
into vector population counts with wider element types.

On x86 there is an instruction that will directly compute the horizontal
sum for the low 8 and high 8 bytes, giving vNi64 popcount very easily.
Various tricks are used to get vNi32 and vNi16 from the vNi8 that the
LUT computes.

The base implemantion of this, and most of the work, was done by Bruno
in a follow up to D6531. See Bruno's detailed post there for lots of
timing information about these changes.

I have extended Bruno's patch in the following ways:

0) I committed the new tests with baseline sequences so this shows
   a diff, and regenerated the tests using the update scripts.

1) Bruno had noticed and mentioned in IRC a redundant mask that
   I removed.

2) I introduced a particular optimization for the i32 vector cases where
   we use PSHL + PSADBW to compute the the low i32 popcounts, and PSHUFD
   + PSADBW to compute doubled high i32 popcounts. This takes advantage
   of the fact that to line up the high i32 popcounts we have to shift
   them anyways, and we can shift them by one fewer bit to effectively
   divide the count by two. While the PSHUFD based horizontal add is no
   faster, it doesn't require registers or load traffic the way a mask
   would, and provides more ILP as it happens on different ports with
   high throughput.

3) I did some code cleanups throughout to simplify the implementation
   logic.

4) I refactored it to continue to use the parallel bitmath lowering when
   SSSE3 is not available to preserve the performance of that version on
   SSE2 targets where it is still much better than scalarizing as we'll
   still do a bitmath implementation of popcount even in scalar code
   there.

With #1 and #2 above, I analyzed the result in IACA for sandybridge,
ivybridge, and haswell. In every case I measured, the throughput is the
same or better using the LUT lowering, even v2i64 and v4i64, and even
compared with using the native popcnt instruction! The latency of the
LUT lowering is often higher than the latency of the scalarized popcnt
instruction sequence, but I think those latency measurements are deeply
misleading. Keeping the operation fully in the vector unit and having
many chances for increased throughput seems much more likely to win.

With this, we can lower every integer vector popcount implementation
using the LUT strategy if we have SSSE3 or better (and thus have
PSHUFB). I've updated the operation lowering to reflect this. This also
fixes an issue where we were scalarizing horribly some AVX lowerings.

Finally, there are some remaining cleanups. There is duplication between
the two techniques in how they perform the horizontal sum once the byte
population count is computed. I'm going to factor and merge those two in
a separate follow-up commit.

Differential Revision: http://reviews.llvm.org/D10084

llvm-svn: 238636
This commit is contained in:
Chandler Carruth 2015-05-30 03:20:59 +00:00
parent c2e400de83
commit 6ba9730a4e
6 changed files with 696 additions and 2052 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

204
llvm/lib/Target/X86/X86ISelLowering.cpp
View File

@ -842,15 +842,10 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
setOperationAction(ISD::INSERT_VECTOR_ELT, MVT::v4i32, Custom);
setOperationAction(ISD::INSERT_VECTOR_ELT, MVT::v4f32, Custom);
// Only provide customized ctpop vector bit twiddling for vector types we
// know to perform better than using the popcnt instructions on each vector
// element. If popcnt isn't supported, always provide the custom version.
if (!Subtarget->hasPOPCNT()) {
setOperationAction(ISD::CTPOP, MVT::v2i64, Custom);
setOperationAction(ISD::CTPOP, MVT::v4i32, Custom);
setOperationAction(ISD::CTPOP, MVT::v8i16, Custom);
setOperationAction(ISD::CTPOP, MVT::v16i8, Custom);
}
setOperationAction(ISD::CTPOP, MVT::v16i8, Custom);
setOperationAction(ISD::CTPOP, MVT::v8i16, Custom);
setOperationAction(ISD::CTPOP, MVT::v4i32, Custom);
setOperationAction(ISD::CTPOP, MVT::v2i64, Custom);
// Custom lower build_vector, vector_shuffle, and extract_vector_elt.
for (int i = MVT::v16i8; i != MVT::v2i64; ++i) {
@ -1115,6 +1110,11 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
setOperationAction(ISD::TRUNCATE, MVT::v8i16, Custom);
setOperationAction(ISD::TRUNCATE, MVT::v4i32, Custom);
setOperationAction(ISD::CTPOP, MVT::v32i8, Custom);
setOperationAction(ISD::CTPOP, MVT::v16i16, Custom);
setOperationAction(ISD::CTPOP, MVT::v8i32, Custom);
setOperationAction(ISD::CTPOP, MVT::v4i64, Custom);
if (Subtarget->hasFMA() || Subtarget->hasFMA4()) {
setOperationAction(ISD::FMA, MVT::v8f32, Legal);
setOperationAction(ISD::FMA, MVT::v4f64, Legal);
@ -1149,16 +1149,6 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
// when we have a 256bit-wide blend with immediate.
setOperationAction(ISD::UINT_TO_FP, MVT::v8i32, Custom);
// Only provide customized ctpop vector bit twiddling for vector types we
// know to perform better than using the popcnt instructions on each
// vector element. If popcnt isn't supported, always provide the custom
// version.
if (!Subtarget->hasPOPCNT())
setOperationAction(ISD::CTPOP, MVT::v4i64, Custom);
// Custom CTPOP always performs better on natively supported v8i32
setOperationAction(ISD::CTPOP, MVT::v8i32, Custom);
// AVX2 also has wider vector sign/zero extending loads, VPMOV[SZ]X
setLoadExtAction(ISD::SEXTLOAD, MVT::v16i16, MVT::v16i8, Legal);
setLoadExtAction(ISD::SEXTLOAD, MVT::v8i32, MVT::v8i8, Legal);
@ -17329,12 +17319,164 @@ static SDValue LowerBITCAST(SDValue Op, const X86Subtarget *Subtarget,
return SDValue();
}
static SDValue LowerVectorCTPOPInRegLUT(SDValue Op, SDLoc DL,
const X86Subtarget *Subtarget,
SelectionDAG &DAG) {
EVT VT = Op.getValueType();
MVT EltVT = VT.getVectorElementType().getSimpleVT();
unsigned VecSize = VT.getSizeInBits();
// Implement a lookup table in register by using an algorithm based on:
// http://wm.ite.pl/articles/sse-popcount.html
//
// The general idea is that every lower byte nibble in the input vector is an
// index into a in-register pre-computed pop count table. We then split up the
// input vector in two new ones: (1) a vector with only the shifted-right
// higher nibbles for each byte and (2) a vector with the lower nibbles (and
// masked out higher ones) for each byte. PSHUB is used separately with both
// to index the in-register table. Next, both are added and the result is a
// i8 vector where each element contains the pop count for input byte.
//
// To obtain the pop count for elements != i8, we follow up with the same
// approach and use additional tricks as described below.
//
const int LUT[16] = {/* 0 */ 0, /* 1 */ 1, /* 2 */ 1, /* 3 */ 2,
/* 4 */ 1, /* 5 */ 2, /* 6 */ 2, /* 7 */ 3,
/* 8 */ 1, /* 9 */ 2, /* a */ 2, /* b */ 3,
/* c */ 2, /* d */ 3, /* e */ 3, /* f */ 4};
int NumByteElts = VecSize / 8;
MVT ByteVecVT = MVT::getVectorVT(MVT::i8, NumByteElts);
SDValue In = DAG.getNode(ISD::BITCAST, DL, ByteVecVT, Op);
SmallVector<SDValue, 16> LUTVec;
for (int i = 0; i < NumByteElts; ++i)
LUTVec.push_back(DAG.getConstant(LUT[i % 16], DL, MVT::i8));
SDValue InRegLUT = DAG.getNode(ISD::BUILD_VECTOR, DL, ByteVecVT, LUTVec);
SmallVector<SDValue, 16> Mask0F(NumByteElts,
DAG.getConstant(0x0F, DL, MVT::i8));
SDValue M0F = DAG.getNode(ISD::BUILD_VECTOR, DL, ByteVecVT, Mask0F);
// High nibbles
SmallVector<SDValue, 16> Four(NumByteElts, DAG.getConstant(4, DL, MVT::i8));
SDValue FourV = DAG.getNode(ISD::BUILD_VECTOR, DL, ByteVecVT, Four);
SDValue HighNibbles = DAG.getNode(ISD::SRL, DL, ByteVecVT, In, FourV);
// Low nibbles
SDValue LowNibbles = DAG.getNode(ISD::AND, DL, ByteVecVT, In, M0F);
// The input vector is used as the shuffle mask that index elements into the
// LUT. After counting low and high nibbles, add the vector to obtain the
// final pop count per i8 element.
SDValue HighPopCnt =
DAG.getNode(X86ISD::PSHUFB, DL, ByteVecVT, InRegLUT, HighNibbles);
SDValue LowPopCnt =
DAG.getNode(X86ISD::PSHUFB, DL, ByteVecVT, InRegLUT, LowNibbles);
SDValue PopCnt = DAG.getNode(ISD::ADD, DL, ByteVecVT, HighPopCnt, LowPopCnt);
if (EltVT == MVT::i8)
return PopCnt;
// PSADBW instruction horizontally add all bytes and leave the result in i64
// chunks, thus directly computes the pop count for v2i64 and v4i64.
if (EltVT == MVT::i64) {
SDValue Zeros = getZeroVector(ByteVecVT, Subtarget, DAG, DL);
PopCnt = DAG.getNode(X86ISD::PSADBW, DL, ByteVecVT, PopCnt, Zeros);
return DAG.getNode(ISD::BITCAST, DL, VT, PopCnt);
}
int NumI64Elts = VecSize / 64;
MVT VecI64VT = MVT::getVectorVT(MVT::i64, NumI64Elts);
if (EltVT == MVT::i32) {
// We unpack the low half and high half into i32s interleaved with zeros so
// that we can use PSADBW to horizontally sum them. The most useful part of
// this is that it lines up the results of two PSADBW instructions to be
// two v2i64 vectors which concatenated are the 4 population counts. We can
// then use PACKUSWB to shrink and concatenate them into a v4i32 again.
SDValue Zeros = getZeroVector(VT, Subtarget, DAG, DL);
SDValue Low = DAG.getNode(X86ISD::UNPCKL, DL, VT, PopCnt, Zeros);
SDValue High = DAG.getNode(X86ISD::UNPCKH, DL, VT, PopCnt, Zeros);
// Do the horizontal sums into two v2i64s.
Zeros = getZeroVector(ByteVecVT, Subtarget, DAG, DL);
Low = DAG.getNode(X86ISD::PSADBW, DL, ByteVecVT,
DAG.getNode(ISD::BITCAST, DL, ByteVecVT, Low), Zeros);
High = DAG.getNode(X86ISD::PSADBW, DL, ByteVecVT,
DAG.getNode(ISD::BITCAST, DL, ByteVecVT, High), Zeros);
// Merge them together.
MVT ShortVecVT = MVT::getVectorVT(MVT::i16, VecSize / 16);
PopCnt = DAG.getNode(X86ISD::PACKUS, DL, ByteVecVT,
DAG.getNode(ISD::BITCAST, DL, ShortVecVT, Low),
DAG.getNode(ISD::BITCAST, DL, ShortVecVT, High));
return DAG.getNode(ISD::BITCAST, DL, VT, PopCnt);
}
// To obtain pop count for each i16 element, shuffle the byte pop count to get
// even and odd elements into distinct vectors, add them and zero-extend each
// i8 elemento into i16, i.e.:
//
// B -> pop count per i8
// W -> pop count per i16
//
// Y = shuffle B, undef <0, 2, ...>
// Z = shuffle B, undef <1, 3, ...>
// W = zext <... x i8> to <... x i16> (Y + Z)
//
// Use a byte shuffle mask that matches PSHUFB.
//
assert(EltVT == MVT::i16 && "Unknown how to handle type");
SDValue Undef = DAG.getUNDEF(ByteVecVT);
SmallVector<int, 32> MaskA, MaskB;
// We can't use PSHUFB across lanes, so do the shuffle and sum inside each
// 128-bit lane, and then collapse the result.
int NumLanes = NumByteElts / 16;
assert(NumByteElts % 16 == 0 && "Must have 16-byte multiple vectors!");
for (int i = 0; i < NumLanes; ++i) {
for (int j = 0; j < 8; ++j) {
MaskA.push_back(i * 16 + j * 2);
MaskB.push_back(i * 16 + (j * 2) + 1);
}
MaskA.append((size_t)8, -1);
MaskB.append((size_t)8, -1);
}
SDValue ShuffA = DAG.getVectorShuffle(ByteVecVT, DL, PopCnt, Undef, MaskA);
SDValue ShuffB = DAG.getVectorShuffle(ByteVecVT, DL, PopCnt, Undef, MaskB);
PopCnt = DAG.getNode(ISD::ADD, DL, ByteVecVT, ShuffA, ShuffB);
SmallVector<int, 4> Mask;
for (int i = 0; i < NumLanes; ++i)
Mask.push_back(2 * i);
Mask.append((size_t)NumLanes, -1);
PopCnt = DAG.getNode(ISD::BITCAST, DL, VecI64VT, PopCnt);
PopCnt =
DAG.getVectorShuffle(VecI64VT, DL, PopCnt, DAG.getUNDEF(VecI64VT), Mask);
PopCnt = DAG.getNode(ISD::BITCAST, DL, ByteVecVT, PopCnt);
// Zero extend i8s into i16 elts
SmallVector<int, 16> ZExtInRegMask;
for (int i = 0; i < NumByteElts / 2; ++i) {
ZExtInRegMask.push_back(i);
ZExtInRegMask.push_back(NumByteElts);
}
return DAG.getNode(
ISD::BITCAST, DL, VT,
DAG.getVectorShuffle(ByteVecVT, DL, PopCnt,
getZeroVector(ByteVecVT, Subtarget, DAG, DL),
ZExtInRegMask));
}
static SDValue LowerVectorCTPOPBitmath(SDValue Op, SDLoc DL,
const X86Subtarget *Subtarget,
SelectionDAG &DAG) {
MVT VT = Op.getSimpleValueType();
assert((VT.is128BitVector() || VT.is256BitVector()) &&
"CTPOP lowering only implemented for 128/256-bit wide vector types");
assert(VT.is128BitVector() &&
"Only 128-bit vector bitmath lowering supported.");
int VecSize = VT.getSizeInBits();
int NumElts = VT.getVectorNumElements();
@ -17344,9 +17486,9 @@ static SDValue LowerVectorCTPOPBitmath(SDValue Op, SDLoc DL,
// This is the vectorized version of the "best" algorithm from
// http://graphics.stanford.edu/~seander/bithacks.html#CountBitsSetParallel
// with a minor tweak to use a series of adds + shifts instead of vector
// multiplications. Implemented for all integer vector types.
//
// FIXME: Use strategies from http://wm.ite.pl/articles/sse-popcount.html
// multiplications. Implemented for all integer vector types. We only use
// this when we don't have SSSE3 which allows a LUT-based lowering that is
// much faster, even faster than using native popcnt instructions.
SDValue Cst55 = DAG.getConstant(APInt::getSplat(Len, APInt(8, 0x55)), DL,
EltVT);
@ -17424,7 +17566,6 @@ static SDValue LowerVectorCTPOPBitmath(SDValue Op, SDLoc DL,
return V;
}
static SDValue LowerVectorCTPOP(SDValue Op, const X86Subtarget *Subtarget,
SelectionDAG &DAG) {
MVT VT = Op.getSimpleValueType();
@ -17434,6 +17575,12 @@ static SDValue LowerVectorCTPOP(SDValue Op, const X86Subtarget *Subtarget,
SDLoc DL(Op.getNode());
SDValue Op0 = Op.getOperand(0);
if (!Subtarget->hasSSSE3()) {
// We can't use the fast LUT approach, so fall back on vectorized bitmath.
assert(VT.is128BitVector() && "Only 128-bit vectors supported in SSE!");
return LowerVectorCTPOPBitmath(Op0, DL, Subtarget, DAG);
}
if (VT.is256BitVector() && !Subtarget->hasInt256()) {
unsigned NumElems = VT.getVectorNumElements();
@ -17442,11 +17589,11 @@ static SDValue LowerVectorCTPOP(SDValue Op, const X86Subtarget *Subtarget,
SDValue RHS = Extract128BitVector(Op0, NumElems/2, DAG, DL);
return DAG.getNode(ISD::CONCAT_VECTORS, DL, VT,
LowerVectorCTPOPBitmath(LHS, DL, Subtarget, DAG),
LowerVectorCTPOPBitmath(RHS, DL, Subtarget, DAG));
LowerVectorCTPOPInRegLUT(LHS, DL, Subtarget, DAG),
LowerVectorCTPOPInRegLUT(RHS, DL, Subtarget, DAG));
}
return LowerVectorCTPOPBitmath(Op0, DL, Subtarget, DAG);
return LowerVectorCTPOPInRegLUT(Op0, DL, Subtarget, DAG);
}
static SDValue LowerCTPOP(SDValue Op, const X86Subtarget *Subtarget,
@ -18149,6 +18296,7 @@ const char *X86TargetLowering::getTargetNodeName(unsigned Opcode) const {
case X86ISD::VPERMI: return "X86ISD::VPERMI";
case X86ISD::PMULUDQ: return "X86ISD::PMULUDQ";
case X86ISD::PMULDQ: return "X86ISD::PMULDQ";
case X86ISD::PSADBW: return "X86ISD::PSADBW";
case X86ISD::VASTART_SAVE_XMM_REGS: return "X86ISD::VASTART_SAVE_XMM_REGS";
case X86ISD::VAARG_64: return "X86ISD::VAARG_64";
case X86ISD::WIN_ALLOCA: return "X86ISD::WIN_ALLOCA";

3
llvm/lib/Target/X86/X86ISelLowering.h
View File

@ -184,6 +184,9 @@ namespace llvm {
/// Shuffle 16 8-bit values within a vector.
PSHUFB,
/// Compute Sum of Absolute Differences.
PSADBW,
/// Bitwise Logical AND NOT of Packed FP values.
ANDNP,

3
llvm/lib/Target/X86/X86InstrFragmentsSIMD.td
View File

@ -78,6 +78,9 @@ def X86cmps : SDNode<"X86ISD::FSETCC", SDTX86Cmps>;
def X86pshufb : SDNode<"X86ISD::PSHUFB",
SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisSameAs<0,1>,
SDTCisSameAs<0,2>]>>;
def X86psadbw : SDNode<"X86ISD::PSADBW",
SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisSameAs<0,1>,
SDTCisSameAs<0,2>]>>;
def X86andnp : SDNode<"X86ISD::ANDNP",
SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisSameAs<0,1>,
SDTCisSameAs<0,2>]>>;

14
llvm/lib/Target/X86/X86InstrSSE.td
View File

@ -4053,6 +4053,20 @@ defm PAVGW : PDI_binop_all_int<0xE3, "pavgw", int_x86_sse2_pavg_w,
defm PSADBW : PDI_binop_all_int<0xF6, "psadbw", int_x86_sse2_psad_bw,
int_x86_avx2_psad_bw, SSE_PMADD, 1>;
let Predicates = [HasAVX2] in
def : Pat<(v32i8 (X86psadbw (v32i8 VR256:$src1),
(v32i8 VR256:$src2))),
(VPSADBWYrr VR256:$src2, VR256:$src1)>;
let Predicates = [HasAVX] in
def : Pat<(v16i8 (X86psadbw (v16i8 VR128:$src1),
(v16i8 VR128:$src2))),
(VPSADBWrr VR128:$src2, VR128:$src1)>;
def : Pat<(v16i8 (X86psadbw (v16i8 VR128:$src1),
(v16i8 VR128:$src2))),
(PSADBWrr VR128:$src2, VR128:$src1)>;
let Predicates = [HasAVX] in
defm VPMULUDQ : PDI_binop_rm2<0xF4, "vpmuludq", X86pmuludq, v2i64, v4i32, VR128,
loadv2i64, i128mem, SSE_INTMUL_ITINS_P, 1, 0>,

551
llvm/test/CodeGen/X86/vector-popcnt-128.ll
View File

@ -8,215 +8,418 @@
target triple = "x86_64-unknown-unknown"
define <2 x i64> @testv2i64(<2 x i64> %in) {
; SSE-LABEL: testv2i64:
; SSE: # BB#0:
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrlq $1, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: psubq %xmm1, %xmm0
; SSE-NEXT: movdqa {{.*#+}} xmm1 = [3689348814741910323,3689348814741910323]
; SSE-NEXT: movdqa %xmm0, %xmm2
; SSE-NEXT: pand %xmm1, %xmm2
; SSE-NEXT: psrlq $2, %xmm0
; SSE-NEXT: pand %xmm1, %xmm0
; SSE-NEXT: paddq %xmm2, %xmm0
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrlq $4, %xmm1
; SSE-NEXT: paddq %xmm0, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: psllq $32, %xmm0
; SSE-NEXT: paddb %xmm1, %xmm0
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psllq $16, %xmm1
; SSE-NEXT: paddb %xmm0, %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: psllq $8, %xmm0
; SSE-NEXT: paddb %xmm1, %xmm0
; SSE-NEXT: psrlq $56, %xmm0
; SSE-NEXT: retq
; SSE2-LABEL: testv2i64:
; SSE2: # BB#0:
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrlq $1, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: psubq %xmm1, %xmm0
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [3689348814741910323,3689348814741910323]
; SSE2-NEXT: movdqa %xmm0, %xmm2
; SSE2-NEXT: pand %xmm1, %xmm2
; SSE2-NEXT: psrlq $2, %xmm0
; SSE2-NEXT: pand %xmm1, %xmm0
; SSE2-NEXT: paddq %xmm2, %xmm0
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrlq $4, %xmm1
; SSE2-NEXT: paddq %xmm0, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: movdqa %xmm1, %xmm0
; SSE2-NEXT: psllq $32, %xmm0
; SSE2-NEXT: paddb %xmm1, %xmm0
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psllq $16, %xmm1
; SSE2-NEXT: paddb %xmm0, %xmm1
; SSE2-NEXT: movdqa %xmm1, %xmm0
; SSE2-NEXT: psllq $8, %xmm0
; SSE2-NEXT: paddb %xmm1, %xmm0
; SSE2-NEXT: psrlq $56, %xmm0
; SSE2-NEXT: retq
;
; SSE3-LABEL: testv2i64:
; SSE3: # BB#0:
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrlq $1, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: psubq %xmm1, %xmm0
; SSE3-NEXT: movdqa {{.*#+}} xmm1 = [3689348814741910323,3689348814741910323]
; SSE3-NEXT: movdqa %xmm0, %xmm2
; SSE3-NEXT: pand %xmm1, %xmm2
; SSE3-NEXT: psrlq $2, %xmm0
; SSE3-NEXT: pand %xmm1, %xmm0
; SSE3-NEXT: paddq %xmm2, %xmm0
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrlq $4, %xmm1
; SSE3-NEXT: paddq %xmm0, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: movdqa %xmm1, %xmm0
; SSE3-NEXT: psllq $32, %xmm0
; SSE3-NEXT: paddb %xmm1, %xmm0
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psllq $16, %xmm1
; SSE3-NEXT: paddb %xmm0, %xmm1
; SSE3-NEXT: movdqa %xmm1, %xmm0
; SSE3-NEXT: psllq $8, %xmm0
; SSE3-NEXT: paddb %xmm1, %xmm0
; SSE3-NEXT: psrlq $56, %xmm0
; SSE3-NEXT: retq
;
; SSSE3-LABEL: testv2i64:
; SSSE3: # BB#0:
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSSE3-NEXT: movdqa %xmm0, %xmm2
; SSSE3-NEXT: pand %xmm1, %xmm2
; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSSE3-NEXT: movdqa %xmm3, %xmm4
; SSSE3-NEXT: pshufb %xmm2, %xmm4
; SSSE3-NEXT: psrlw $4, %xmm0
; SSSE3-NEXT: pand %xmm1, %xmm0
; SSSE3-NEXT: pshufb %xmm0, %xmm3
; SSSE3-NEXT: paddb %xmm4, %xmm3
; SSSE3-NEXT: pxor %xmm0, %xmm0
; SSSE3-NEXT: psadbw %xmm3, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: testv2i64:
; SSE41: # BB#0:
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE41-NEXT: movdqa %xmm0, %xmm2
; SSE41-NEXT: pand %xmm1, %xmm2
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSE41-NEXT: movdqa %xmm3, %xmm4
; SSE41-NEXT: pshufb %xmm2, %xmm4
; SSE41-NEXT: psrlw $4, %xmm0
; SSE41-NEXT: pand %xmm1, %xmm0
; SSE41-NEXT: pshufb %xmm0, %xmm3
; SSE41-NEXT: paddb %xmm4, %xmm3
; SSE41-NEXT: pxor %xmm0, %xmm0
; SSE41-NEXT: psadbw %xmm3, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: testv2i64:
; AVX: # BB#0:
; AVX-NEXT: vpsrlq $1, %xmm0, %xmm1
; AVX-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
; AVX-NEXT: vpsubq %xmm1, %xmm0, %xmm0
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [3689348814741910323,3689348814741910323]
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm2
; AVX-NEXT: vpsrlq $2, %xmm0, %xmm0
; AVX-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; AVX-NEXT: vpshufb %xmm2, %xmm3, %xmm2
; AVX-NEXT: vpsrlw $4, %xmm0, %xmm0
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpaddq %xmm0, %xmm2, %xmm0
; AVX-NEXT: vpsrlq $4, %xmm0, %xmm1
; AVX-NEXT: vpaddq %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
; AVX-NEXT: vpsllq $32, %xmm0, %xmm1
; AVX-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpsllq $16, %xmm0, %xmm1
; AVX-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpsllq $8, %xmm0, %xmm1
; AVX-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpsrlq $56, %xmm0, %xmm0
; AVX-NEXT: vpshufb %xmm0, %xmm3, %xmm0
; AVX-NEXT: vpaddb %xmm2, %xmm0, %xmm0
; AVX-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX-NEXT: vpsadbw %xmm0, %xmm1, %xmm0
; AVX-NEXT: retq
%out = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %in)
ret <2 x i64> %out
}
define <4 x i32> @testv4i32(<4 x i32> %in) {
; SSE-LABEL: testv4i32:
; SSE: # BB#0:
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrld $1, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: psubd %xmm1, %xmm0
; SSE-NEXT: movdqa {{.*#+}} xmm1 = [858993459,858993459,858993459,858993459]
; SSE-NEXT: movdqa %xmm0, %xmm2
; SSE-NEXT: pand %xmm1, %xmm2
; SSE-NEXT: psrld $2, %xmm0
; SSE-NEXT: pand %xmm1, %xmm0
; SSE-NEXT: paddd %xmm2, %xmm0
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrld $4, %xmm1
; SSE-NEXT: paddd %xmm0, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm2
; SSE-NEXT: psllq $16, %xmm2
; SSE-NEXT: paddb %xmm1, %xmm2
; SSE-NEXT: movdqa %xmm2, %xmm0
; SSE-NEXT: psllq $8, %xmm0
; SSE-NEXT: paddb %xmm2, %xmm0
; SSE-NEXT: psrld $24, %xmm0
; SSE-NEXT: retq
; SSE2-LABEL: testv4i32:
; SSE2: # BB#0:
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrld $1, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: psubd %xmm1, %xmm0
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [858993459,858993459,858993459,858993459]
; SSE2-NEXT: movdqa %xmm0, %xmm2
; SSE2-NEXT: pand %xmm1, %xmm2
; SSE2-NEXT: psrld $2, %xmm0
; SSE2-NEXT: pand %xmm1, %xmm0
; SSE2-NEXT: paddd %xmm2, %xmm0
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrld $4, %xmm1
; SSE2-NEXT: paddd %xmm0, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: movdqa %xmm1, %xmm2
; SSE2-NEXT: psllq $16, %xmm2
; SSE2-NEXT: paddb %xmm1, %xmm2
; SSE2-NEXT: movdqa %xmm2, %xmm0
; SSE2-NEXT: psllq $8, %xmm0
; SSE2-NEXT: paddb %xmm2, %xmm0
; SSE2-NEXT: psrld $24, %xmm0
; SSE2-NEXT: retq
;
; AVX1-LABEL: testv4i32:
; AVX1: # BB#0:
; AVX1-NEXT: vpsrld $1, %xmm0, %xmm1
; AVX1-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
; AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vmovdqa {{.*#+}} xmm1 = [858993459,858993459,858993459,858993459]
; AVX1-NEXT: vpand %xmm1, %xmm0, %xmm2
; AVX1-NEXT: vpsrld $2, %xmm0, %xmm0
; AVX1-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; AVX1-NEXT: vpsrld $4, %xmm0, %xmm1
; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
; AVX1-NEXT: vpsllq $16, %xmm0, %xmm1
; AVX1-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpsllq $8, %xmm0, %xmm1
; AVX1-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpsrld $24, %xmm0, %xmm0
; AVX1-NEXT: retq
; SSE3-LABEL: testv4i32:
; SSE3: # BB#0:
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrld $1, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: psubd %xmm1, %xmm0
; SSE3-NEXT: movdqa {{.*#+}} xmm1 = [858993459,858993459,858993459,858993459]
; SSE3-NEXT: movdqa %xmm0, %xmm2
; SSE3-NEXT: pand %xmm1, %xmm2
; SSE3-NEXT: psrld $2, %xmm0
; SSE3-NEXT: pand %xmm1, %xmm0
; SSE3-NEXT: paddd %xmm2, %xmm0
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrld $4, %xmm1
; SSE3-NEXT: paddd %xmm0, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: movdqa %xmm1, %xmm2
; SSE3-NEXT: psllq $16, %xmm2
; SSE3-NEXT: paddb %xmm1, %xmm2
; SSE3-NEXT: movdqa %xmm2, %xmm0
; SSE3-NEXT: psllq $8, %xmm0
; SSE3-NEXT: paddb %xmm2, %xmm0
; SSE3-NEXT: psrld $24, %xmm0
; SSE3-NEXT: retq
;
; AVX2-LABEL: testv4i32:
; AVX2: # BB#0:
; AVX2-NEXT: vpsrld $1, %xmm0, %xmm1
; AVX2-NEXT: vpbroadcastd {{.*}}(%rip), %xmm2
; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
; AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpbroadcastd {{.*}}(%rip), %xmm1
; AVX2-NEXT: vpand %xmm1, %xmm0, %xmm2
; AVX2-NEXT: vpsrld $2, %xmm0, %xmm0
; AVX2-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; AVX2-NEXT: vpsrld $4, %xmm0, %xmm1
; AVX2-NEXT: vpaddd %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpbroadcastd {{.*}}(%rip), %xmm1
; AVX2-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpsllq $16, %xmm0, %xmm1
; AVX2-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpsllq $8, %xmm0, %xmm1
; AVX2-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpsrld $24, %xmm0, %xmm0
; AVX2-NEXT: retq
; SSSE3-LABEL: testv4i32:
; SSSE3: # BB#0:
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSSE3-NEXT: movdqa %xmm0, %xmm3
; SSSE3-NEXT: pand %xmm2, %xmm3
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSSE3-NEXT: movdqa %xmm1, %xmm4
; SSSE3-NEXT: pshufb %xmm3, %xmm4
; SSSE3-NEXT: psrlw $4, %xmm0
; SSSE3-NEXT: pand %xmm2, %xmm0
; SSSE3-NEXT: pshufb %xmm0, %xmm1
; SSSE3-NEXT: paddb %xmm4, %xmm1
; SSSE3-NEXT: pxor %xmm0, %xmm0
; SSSE3-NEXT: movdqa %xmm1, %xmm2
; SSSE3-NEXT: punpckhdq {{.*#+}} xmm2 = xmm2[2],xmm0[2],xmm2[3],xmm0[3]
; SSSE3-NEXT: psadbw %xmm0, %xmm2
; SSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
; SSSE3-NEXT: psadbw %xmm0, %xmm1
; SSSE3-NEXT: packuswb %xmm2, %xmm1
; SSSE3-NEXT: movdqa %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: testv4i32:
; SSE41: # BB#0:
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE41-NEXT: movdqa %xmm0, %xmm3
; SSE41-NEXT: pand %xmm2, %xmm3
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSE41-NEXT: movdqa %xmm1, %xmm4
; SSE41-NEXT: pshufb %xmm3, %xmm4
; SSE41-NEXT: psrlw $4, %xmm0
; SSE41-NEXT: pand %xmm2, %xmm0
; SSE41-NEXT: pshufb %xmm0, %xmm1
; SSE41-NEXT: paddb %xmm4, %xmm1
; SSE41-NEXT: pxor %xmm0, %xmm0
; SSE41-NEXT: movdqa %xmm1, %xmm2
; SSE41-NEXT: punpckhdq {{.*#+}} xmm2 = xmm2[2],xmm0[2],xmm2[3],xmm0[3]
; SSE41-NEXT: psadbw %xmm0, %xmm2
; SSE41-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
; SSE41-NEXT: psadbw %xmm0, %xmm1
; SSE41-NEXT: packuswb %xmm2, %xmm1
; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: testv4i32:
; AVX: # BB#0:
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm2
; AVX-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; AVX-NEXT: vpshufb %xmm2, %xmm3, %xmm2
; AVX-NEXT: vpsrlw $4, %xmm0, %xmm0
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpshufb %xmm0, %xmm3, %xmm0
; AVX-NEXT: vpaddb %xmm2, %xmm0, %xmm0
; AVX-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX-NEXT: vpunpckhdq {{.*#+}} xmm2 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
; AVX-NEXT: vpsadbw %xmm2, %xmm1, %xmm2
; AVX-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; AVX-NEXT: vpsadbw %xmm0, %xmm1, %xmm0
; AVX-NEXT: vpackuswb %xmm2, %xmm0, %xmm0
; AVX-NEXT: retq
%out = call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> %in)
ret <4 x i32> %out
}
define <8 x i16> @testv8i16(<8 x i16> %in) {
; SSE-LABEL: testv8i16:
; SSE: # BB#0:
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrlw $1, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: psubw %xmm1, %xmm0
; SSE-NEXT: movdqa {{.*#+}} xmm1 = [13107,13107,13107,13107,13107,13107,13107,13107]
; SSE-NEXT: movdqa %xmm0, %xmm2
; SSE-NEXT: pand %xmm1, %xmm2
; SSE-NEXT: psrlw $2, %xmm0
; SSE-NEXT: pand %xmm1, %xmm0
; SSE-NEXT: paddw %xmm2, %xmm0
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrlw $4, %xmm1
; SSE-NEXT: paddw %xmm0, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: psllq $8, %xmm0
; SSE-NEXT: paddb %xmm1, %xmm0
; SSE-NEXT: psrlw $8, %xmm0
; SSE-NEXT: retq
; SSE2-LABEL: testv8i16:
; SSE2: # BB#0:
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrlw $1, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: psubw %xmm1, %xmm0
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [13107,13107,13107,13107,13107,13107,13107,13107]
; SSE2-NEXT: movdqa %xmm0, %xmm2
; SSE2-NEXT: pand %xmm1, %xmm2
; SSE2-NEXT: psrlw $2, %xmm0
; SSE2-NEXT: pand %xmm1, %xmm0
; SSE2-NEXT: paddw %xmm2, %xmm0
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrlw $4, %xmm1
; SSE2-NEXT: paddw %xmm0, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: movdqa %xmm1, %xmm0
; SSE2-NEXT: psllq $8, %xmm0
; SSE2-NEXT: paddb %xmm1, %xmm0
; SSE2-NEXT: psrlw $8, %xmm0
; SSE2-NEXT: retq
;
; SSE3-LABEL: testv8i16:
; SSE3: # BB#0:
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrlw $1, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: psubw %xmm1, %xmm0
; SSE3-NEXT: movdqa {{.*#+}} xmm1 = [13107,13107,13107,13107,13107,13107,13107,13107]
; SSE3-NEXT: movdqa %xmm0, %xmm2
; SSE3-NEXT: pand %xmm1, %xmm2
; SSE3-NEXT: psrlw $2, %xmm0
; SSE3-NEXT: pand %xmm1, %xmm0
; SSE3-NEXT: paddw %xmm2, %xmm0
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrlw $4, %xmm1
; SSE3-NEXT: paddw %xmm0, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: movdqa %xmm1, %xmm0
; SSE3-NEXT: psllq $8, %xmm0
; SSE3-NEXT: paddb %xmm1, %xmm0
; SSE3-NEXT: psrlw $8, %xmm0
; SSE3-NEXT: retq
;
; SSSE3-LABEL: testv8i16:
; SSSE3: # BB#0:
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSSE3-NEXT: movdqa %xmm0, %xmm3
; SSSE3-NEXT: pand %xmm2, %xmm3
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSSE3-NEXT: movdqa %xmm1, %xmm4
; SSSE3-NEXT: pshufb %xmm3, %xmm4
; SSSE3-NEXT: psrlw $4, %xmm0
; SSSE3-NEXT: pand %xmm2, %xmm0
; SSSE3-NEXT: pshufb %xmm0, %xmm1
; SSSE3-NEXT: paddb %xmm4, %xmm1
; SSSE3-NEXT: movdqa %xmm1, %xmm0
; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u]
; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; SSSE3-NEXT: paddb %xmm0, %xmm1
; SSSE3-NEXT: pxor %xmm0, %xmm0
; SSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
; SSSE3-NEXT: movdqa %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: testv8i16:
; SSE41: # BB#0:
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE41-NEXT: movdqa %xmm0, %xmm2
; SSE41-NEXT: pand %xmm1, %xmm2
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSE41-NEXT: movdqa %xmm3, %xmm4
; SSE41-NEXT: pshufb %xmm2, %xmm4
; SSE41-NEXT: psrlw $4, %xmm0
; SSE41-NEXT: pand %xmm1, %xmm0
; SSE41-NEXT: pshufb %xmm0, %xmm3
; SSE41-NEXT: paddb %xmm4, %xmm3
; SSE41-NEXT: movdqa %xmm3, %xmm0
; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u]
; SSE41-NEXT: pshufb {{.*#+}} xmm3 = xmm3[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; SSE41-NEXT: paddb %xmm0, %xmm3
; SSE41-NEXT: pmovzxbw {{.*#+}} xmm0 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
; SSE41-NEXT: retq
;
; AVX-LABEL: testv8i16:
; AVX: # BB#0:
; AVX-NEXT: vpsrlw $1, %xmm0, %xmm1
; AVX-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
; AVX-NEXT: vpsubw %xmm1, %xmm0, %xmm0
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [13107,13107,13107,13107,13107,13107,13107,13107]
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm2
; AVX-NEXT: vpsrlw $2, %xmm0, %xmm0
; AVX-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; AVX-NEXT: vpshufb %xmm2, %xmm3, %xmm2
; AVX-NEXT: vpsrlw $4, %xmm0, %xmm0
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpaddw %xmm0, %xmm2, %xmm0
; AVX-NEXT: vpsrlw $4, %xmm0, %xmm1
; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
; AVX-NEXT: vpsllq $8, %xmm0, %xmm1
; AVX-NEXT: vpshufb %xmm0, %xmm3, %xmm0
; AVX-NEXT: vpaddb %xmm2, %xmm0, %xmm0
; AVX-NEXT: vpshufb {{.*#+}} xmm1 = xmm0[1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u]
; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; AVX-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpsrlw $8, %xmm0, %xmm0
; AVX-NEXT: vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; AVX-NEXT: retq
%out = call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> %in)
ret <8 x i16> %out
}
define <16 x i8> @testv16i8(<16 x i8> %in) {
; SSE-LABEL: testv16i8:
; SSE: # BB#0:
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrlw $1, %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: pand {{.*}}(%rip), %xmm1
; SSE-NEXT: psubb %xmm1, %xmm0
; SSE-NEXT: movdqa {{.*#+}} xmm1 = [51,51,51,51,51,51,51,51,51,51,51,51,51,51,51,51]
; SSE-NEXT: movdqa %xmm0, %xmm2
; SSE-NEXT: pand %xmm1, %xmm2
; SSE-NEXT: psrlw $2, %xmm0
; SSE-NEXT: pand {{.*}}(%rip), %xmm0
; SSE-NEXT: pand %xmm1, %xmm0
; SSE-NEXT: paddb %xmm2, %xmm0
; SSE-NEXT: movdqa %xmm0, %xmm1
; SSE-NEXT: psrlw $4, %xmm1
; SSE-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE-NEXT: pand %xmm2, %xmm1
; SSE-NEXT: paddb %xmm0, %xmm1
; SSE-NEXT: pand %xmm2, %xmm1
; SSE-NEXT: movdqa %xmm1, %xmm0
; SSE-NEXT: retq
; SSE2-LABEL: testv16i8:
; SSE2: # BB#0:
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrlw $1, %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
; SSE2-NEXT: psubb %xmm1, %xmm0
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [51,51,51,51,51,51,51,51,51,51,51,51,51,51,51,51]
; SSE2-NEXT: movdqa %xmm0, %xmm2
; SSE2-NEXT: pand %xmm1, %xmm2
; SSE2-NEXT: psrlw $2, %xmm0
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
; SSE2-NEXT: pand %xmm1, %xmm0
; SSE2-NEXT: paddb %xmm2, %xmm0
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: psrlw $4, %xmm1
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE2-NEXT: pand %xmm2, %xmm1
; SSE2-NEXT: paddb %xmm0, %xmm1
; SSE2-NEXT: pand %xmm2, %xmm1
; SSE2-NEXT: movdqa %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSE3-LABEL: testv16i8:
; SSE3: # BB#0:
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrlw $1, %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: pand {{.*}}(%rip), %xmm1
; SSE3-NEXT: psubb %xmm1, %xmm0
; SSE3-NEXT: movdqa {{.*#+}} xmm1 = [51,51,51,51,51,51,51,51,51,51,51,51,51,51,51,51]
; SSE3-NEXT: movdqa %xmm0, %xmm2
; SSE3-NEXT: pand %xmm1, %xmm2
; SSE3-NEXT: psrlw $2, %xmm0
; SSE3-NEXT: pand {{.*}}(%rip), %xmm0
; SSE3-NEXT: pand %xmm1, %xmm0
; SSE3-NEXT: paddb %xmm2, %xmm0
; SSE3-NEXT: movdqa %xmm0, %xmm1
; SSE3-NEXT: psrlw $4, %xmm1
; SSE3-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE3-NEXT: pand %xmm2, %xmm1
; SSE3-NEXT: paddb %xmm0, %xmm1
; SSE3-NEXT: pand %xmm2, %xmm1
; SSE3-NEXT: movdqa %xmm1, %xmm0
; SSE3-NEXT: retq
;
; SSSE3-LABEL: testv16i8:
; SSSE3: # BB#0:
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSSE3-NEXT: movdqa %xmm0, %xmm3
; SSSE3-NEXT: pand %xmm2, %xmm3
; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSSE3-NEXT: movdqa %xmm1, %xmm4
; SSSE3-NEXT: pshufb %xmm3, %xmm4
; SSSE3-NEXT: psrlw $4, %xmm0
; SSSE3-NEXT: pand %xmm2, %xmm0
; SSSE3-NEXT: pshufb %xmm0, %xmm1
; SSSE3-NEXT: paddb %xmm4, %xmm1
; SSSE3-NEXT: movdqa %xmm1, %xmm0
; SSSE3-NEXT: retq
;
; SSE41-LABEL: testv16i8:
; SSE41: # BB#0:
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; SSE41-NEXT: movdqa %xmm0, %xmm3
; SSE41-NEXT: pand %xmm2, %xmm3
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; SSE41-NEXT: movdqa %xmm1, %xmm4
; SSE41-NEXT: pshufb %xmm3, %xmm4
; SSE41-NEXT: psrlw $4, %xmm0
; SSE41-NEXT: pand %xmm2, %xmm0
; SSE41-NEXT: pshufb %xmm0, %xmm1
; SSE41-NEXT: paddb %xmm4, %xmm1
; SSE41-NEXT: movdqa %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX-LABEL: testv16i8:
; AVX: # BB#0:
; AVX-NEXT: vpsrlw $1, %xmm0, %xmm1
; AVX-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
; AVX-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
; AVX-NEXT: vpsubb %xmm1, %xmm0, %xmm0
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [51,51,51,51,51,51,51,51,51,51,51,51,51,51,51,51]
; AVX-NEXT: vmovdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm2
; AVX-NEXT: vpsrlw $2, %xmm0, %xmm0
; AVX-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
; AVX-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4]
; AVX-NEXT: vpshufb %xmm2, %xmm3, %xmm2
; AVX-NEXT: vpsrlw $4, %xmm0, %xmm0
; AVX-NEXT: vpand %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpaddb %xmm0, %xmm2, %xmm0
; AVX-NEXT: vpsrlw $4, %xmm0, %xmm1
; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
; AVX-NEXT: vpand %xmm2, %xmm1, %xmm1
; AVX-NEXT: vpaddb %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpand %xmm2, %xmm0, %xmm0
; AVX-NEXT: vpshufb %xmm0, %xmm3, %xmm0
; AVX-NEXT: vpaddb %xmm2, %xmm0, %xmm0
; AVX-NEXT: retq
%out = call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> %in)
ret <16 x i8> %out

1973
llvm/test/CodeGen/X86/vector-popcnt-256.ll
View File

File diff suppressed because it is too large Load Diff