aboutsummaryrefslogtreecommitdiff
path: root/src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp
diff options
context:
space:
mode:
Diffstat (limited to 'src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp')
-rw-r--r--src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp18
1 files changed, 9 insertions, 9 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp b/src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp
index 587bec366a..b50c240a3a 100644
--- a/src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp
+++ b/src/core/NEON/kernels/arm_gemm/transforms/a32_transpose_interleave_8way_32bit.hpp
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2017-2018 ARM Limited.
+ * Copyright (c) 2017-2021 Arm Limited.
*
* SPDX-License-Identifier: MIT
*
@@ -30,22 +30,22 @@
// Generic unblocked transposed 8x32-bit sized specialisation
template <>
template <typename T>
-inline void TransformImpl<8, 1, true, 4, 4, false>::Transform(
+void TransformImpl<8, 1, true, 4, 4, VLType::None>::Transform(
T* out, const T* const in, const int stride,
const int x0, const int xmax, const int k0, const int kmax
) {
// Redirect to a 16x uint16_t specialisation
- TransformImpl<16, 1, true, 2, 2, false>::Transform(
+ TransformImpl<16, 1, true, 2, 2, VLType::None>::Transform(
reinterpret_cast<uint16_t *>(out),
reinterpret_cast<const uint16_t *>(in),
stride*2, x0*2, xmax*2, k0, kmax
);
}
-// Generic 12x16-bit sized specialisation
+// Generic 16x16-bit sized specialisation
template <>
template <typename T>
-inline void TransformImpl<16, 1, true, 2, 2, false>::Transform(
+void TransformImpl<16, 1, true, 2, 2, VLType::None>::Transform(
T* out, const T* const in, const int stride,
const int x0, const int xmax, const int k0, const int kmax
) {
@@ -59,7 +59,7 @@ inline void TransformImpl<16, 1, true, 2, 2, false>::Transform(
// Specialised 16 x uint16_t version
template <>
-inline void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x1(const uint16_t *&in0, uint16_t *out) {
+void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x1(const uint16_t *&in0, uint16_t *out) {
__asm volatile (
"VLD1.32 {d0-d3}, [%[in0]]!\n"
"VST1.32 {d0-d3}, [%[out]]\n"
@@ -72,7 +72,7 @@ inline void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x1(con
}
template <>
-inline void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x2(const uint16_t *&in0, const uint16_t *&in1, uint16_t *out) {
+void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x2(const uint16_t *&in0, const uint16_t *&in1, uint16_t *out) {
__asm volatile (
"VLD1.32 {d0-d3}, [%[in0]]!\n"
"VST1.32 {d0-d3}, [%[out]]!\n"
@@ -90,7 +90,7 @@ inline void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x2(con
}
template <>
-inline void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x4(const uint16_t *&in0, const uint16_t *&in1, const uint16_t *&in2, const uint16_t *&in3, uint16_t *out) {
+void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x4(const uint16_t *&in0, const uint16_t *&in1, const uint16_t *&in2, const uint16_t *&in3, uint16_t *out) {
__asm __volatile (
"VLD1.32 {d0-d3}, [%[in0]]!\n"
"VST1.32 {d0-d3}, [%[out]]!\n"
@@ -117,7 +117,7 @@ inline void TransposeInterleaveCommon<16, uint16_t, uint16_t>::moveblock_1x4(con
template <>
template <>
-inline void TransformImpl<16, 1, true, 2, 2, false>::Transform(
+void TransformImpl<16, 1, true, 2, 2, VLType::None>::Transform(
uint16_t* out, const uint16_t* const in, const int stride,
const int x0, const int xmax, const int k0, const int kmax
) {