Merge pull request #5566 from lioncash/mov

[SVE256] Handle SSE insertions for MOVH(PD, PD, LPS) and MOVL(PD, PS, HPS)
This commit is contained in:
Ryan Houdek authored and GitHub committed 2026-06-17 18:55:22 -07:00
commit 3929d25dcc
7 files changed
+164 -4

No files matched your search

@@ -102,14 +102,14 @@ void OpDispatchBuilder::MOVHPDOp(OpcodeArgs) {
Ref Src = LoadSourceGPR(Op, Op->Src[0], Op->Flags);
Ref Dest = LoadSourceFPR_WithOpSize(Op, Op->Dest, OpSize::i128Bit, Op->Flags);
auto Result = _VInsElement(OpSize::i128Bit, OpSize::i64Bit, 1, 0, Dest, Src);
StoreResultFPR(Op, Result);
StoreResult_WithAVXInsert(VectorOpType::SSE, RegClass::FPR, Op, Result);
} else {
// If the destination is a GPR then the source is memory
// xmm1[127:64] = src
Ref Src = MakeSegmentAddress(Op, Op->Src[0]);
Ref Dest = LoadSourceFPR_WithOpSize(Op, Op->Dest, OpSize::i128Bit, Op->Flags);
auto Result = _VLoadVectorElement(OpSize::i128Bit, OpSize::i64Bit, Dest, 1, Src);
StoreResultFPR(Op, Result);
StoreResult_WithAVXInsert(VectorOpType::SSE, RegClass::FPR, Op, Result);
}
} else {
// In this case memory is the destination and the high bits of the XMM are source
@@ -141,13 +141,13 @@ void OpDispatchBuilder::MOVLPOp(OpcodeArgs) {
Ref Src = LoadSourceFPR(Op, Op->Src[0], Op->Flags, {.Align = OpSize::i128Bit});
Ref Dest = LoadSourceFPR(Op, Op->Dest, Op->Flags, {.Align = OpSize::i128Bit});
auto Result = _VInsElement(OpSize::i128Bit, OpSize::i64Bit, 0, 1, Dest, Src);
StoreResultFPR_WithOpSize(Op, Op->Dest, Result, OpSize::i128Bit, OpSize::i128Bit);
StoreResult_WithAVXInsert(VectorOpType::SSE, RegClass::FPR, Op, Result, OpSize::i128Bit);
} else {
const auto DstSize = OpSizeFromDst(Op);
Ref Src = MakeSegmentAddress(Op, Op->Src[0]);
Ref Dest = LoadSourceFPR_WithOpSize(Op, Op->Dest, DstSize, Op->Flags);
auto Result = _VLoadVectorElement(OpSize::i128Bit, OpSize::i64Bit, Dest, 0, Src);
StoreResultFPR(Op, Result);
StoreResult_WithAVXInsert(VectorOpType::SSE, RegClass::FPR, Op, Result);
}
} else {
Ref Src = LoadSourceFPR(Op, Op->Src[0], Op->Flags, {.Align = OpSize::i64Bit});
@@ -0,0 +1,28 @@
%ifdef CONFIG
{
"HostFeatures": ["AVX"],
"RegData": {
"XMM0": ["0xbd7eb46a1278f793", "0x7d7ccd8836d09fc2", "0xccdbcfc31f3ff0f3", "0x108390defebac4be"],
"XMM1": ["0x4549bd7eb46a1278", "0x4549bd7eb46a1278", "0", "0"],
"XMM2": ["0xc4be43cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"],
"XMM3": ["0x000043cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"]
}
}
%endif
vmovapd ymm0, [rel .data]
vmovapd ymm1, [rel .data + (1 * 32)]
vmovapd ymm2, [rel .data + (2 * 32)]
vmovapd ymm3, [rel .data + (3 * 32)]
movhlps xmm0, xmm1
vmovhlps xmm1, xmm2, xmm3
hlt
align 32
.data:
dq 0xfdecd28fab3fa4a5, 0x7d7ccd8836d09fc2, 0xccdbcfc31f3ff0f3, 0x108390defebac4be
dq 0x43cc1ad6970b4549, 0xbd7eb46a1278f793, 0xef673dac6e4cbb7b, 0x5b3d85d342718be9
dq 0xc4be43cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
dq 0x000043cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
@@ -0,0 +1,26 @@
%ifdef CONFIG
{
"HostFeatures": ["AVX"],
"RegData": {
"XMM0": ["0xfdecd28fab3fa4a5", "0xfdecd28fab3fa4a5", "0xccdbcfc31f3ff0f3", "0x108390defebac4be"],
"XMM1": ["0xc4be43cc1ad6970b", "0xfdecd28fab3fa4a5", "0", "0"],
"XMM2": ["0xc4be43cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"]
}
}
%endif
vmovapd ymm0, [rel .data]
vmovapd ymm1, [rel .data + (1 * 32)]
vmovapd ymm2, [rel .data + (2 * 32)]
movhpd xmm0, [rel .data]
vmovhpd xmm1, xmm2, [rel .data]
hlt
align 32
.data:
dq 0xfdecd28fab3fa4a5, 0x7d7ccd8836d09fc2, 0xccdbcfc31f3ff0f3, 0x108390defebac4be
dq 0x43cc1ad6970b4549, 0xbd7eb46a1278f793, 0xef673dac6e4cbb7b, 0x5b3d85d342718be9
dq 0xc4be43cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
dq 0x000043cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
@@ -0,0 +1,26 @@
%ifdef CONFIG
{
"HostFeatures": ["AVX"],
"RegData": {
"XMM0": ["0xfdecd28fab3fa4a5", "0xfdecd28fab3fa4a5", "0xccdbcfc31f3ff0f3", "0x108390defebac4be"],
"XMM1": ["0xc4be43cc1ad6970b", "0xfdecd28fab3fa4a5", "0", "0"],
"XMM2": ["0xc4be43cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"]
}
}
%endif
vmovapd ymm0, [rel .data]
vmovapd ymm1, [rel .data + (1 * 32)]
vmovapd ymm2, [rel .data + (2 * 32)]
movhps xmm0, [rel .data]
vmovhps xmm1, xmm2, [rel .data]
hlt
align 32
.data:
dq 0xfdecd28fab3fa4a5, 0x7d7ccd8836d09fc2, 0xccdbcfc31f3ff0f3, 0x108390defebac4be
dq 0x43cc1ad6970b4549, 0xbd7eb46a1278f793, 0xef673dac6e4cbb7b, 0x5b3d85d342718be9
dq 0xc4be43cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
dq 0x000043cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
@@ -0,0 +1,28 @@
%ifdef CONFIG
{
"HostFeatures": ["AVX"],
"RegData": {
"XMM0": ["0xfdecd28fab3fa4a5", "0x43cc1ad6970b4549", "0xccdbcfc31f3ff0f3", "0x108390defebac4be"],
"XMM1": ["0xc4be43cc1ad6970b", "0x000043cc1ad6970b", "0", "0"],
"XMM2": ["0xc4be43cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"],
"XMM3": ["0x000043cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"]
}
}
%endif
vmovapd ymm0, [rel .data]
vmovapd ymm1, [rel .data + (1 * 32)]
vmovapd ymm2, [rel .data + (2 * 32)]
vmovapd ymm3, [rel .data + (3 * 32)]
movlhps xmm0, xmm1
vmovlhps xmm1, xmm2, xmm3
hlt
align 32
.data:
dq 0xfdecd28fab3fa4a5, 0x7d7ccd8836d09fc2, 0xccdbcfc31f3ff0f3, 0x108390defebac4be
dq 0x43cc1ad6970b4549, 0xbd7eb46a1278f793, 0xef673dac6e4cbb7b, 0x5b3d85d342718be9
dq 0xc4be43cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
dq 0x000043cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
@@ -0,0 +1,26 @@
%ifdef CONFIG
{
"HostFeatures": ["AVX"],
"RegData": {
"XMM0": ["0xfdecd28fab3fa4a5", "0x7d7ccd8836d09fc2", "0xccdbcfc31f3ff0f3", "0x108390defebac4be"],
"XMM1": ["0xfdecd28fab3fa4a5", "0x4549bd7eb46a1278", "0", "0"],
"XMM2": ["0xc4be43cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"]
}
}
%endif
vmovapd ymm0, [rel .data]
vmovapd ymm1, [rel .data + (1 * 32)]
vmovapd ymm2, [rel .data + (2 * 32)]
movlpd xmm0, [rel .data]
vmovlpd xmm1, xmm2, [rel .data]
hlt
align 32
.data:
dq 0xfdecd28fab3fa4a5, 0x7d7ccd8836d09fc2, 0xccdbcfc31f3ff0f3, 0x108390defebac4be
dq 0x43cc1ad6970b4549, 0xbd7eb46a1278f793, 0xef673dac6e4cbb7b, 0x5b3d85d342718be9
dq 0xc4be43cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
dq 0x000043cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
@@ -0,0 +1,26 @@
%ifdef CONFIG
{
"HostFeatures": ["AVX"],
"RegData": {
"XMM0": ["0xfdecd28fab3fa4a5", "0x7d7ccd8836d09fc2", "0xccdbcfc31f3ff0f3", "0x108390defebac4be"],
"XMM1": ["0xfdecd28fab3fa4a5", "0x4549bd7eb46a1278", "0", "0"],
"XMM2": ["0xc4be43cc1ad6970b", "0x4549bd7eb46a1278", "0xf793ef673dac6e4c", "0xbb7b5b3d85d34271"]
}
}
%endif
vmovapd ymm0, [rel .data]
vmovapd ymm1, [rel .data + (1 * 32)]
vmovapd ymm2, [rel .data + (2 * 32)]
movlps xmm0, [rel .data]
vmovlps xmm1, xmm2, [rel .data]
hlt
align 32
.data:
dq 0xfdecd28fab3fa4a5, 0x7d7ccd8836d09fc2, 0xccdbcfc31f3ff0f3, 0x108390defebac4be
dq 0x43cc1ad6970b4549, 0xbd7eb46a1278f793, 0xef673dac6e4cbb7b, 0x5b3d85d342718be9
dq 0xc4be43cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271
dq 0x000043cc1ad6970b, 0x4549bd7eb46a1278, 0xf793ef673dac6e4c, 0xbb7b5b3d85d34271