[x265] [PATCH v2 06/10] AArch64: Optimise HBD interp_vert_ps_neon

Gerda Zsejke More gerdazsejke.more at arm.com
Tue Mar 11 19:47:11 UTC 2025


Optimise the HBD 4-tap and 8-tap Neon implementations of
interp_vert_ps_neon and extend these functions to support all CHROMA
and LUMA block sizes respectively.

For the 12-bit implementation the new 4-tap filter implementation is
up to 54% faster when coeffIdx==4 and up to 28% faster for the other
filter values compared to the existing Neon implementation. The new
8-tap filter implementation is up to 55% faster when coeffIdx==1,
64% when it is 2, and 58% when it is 3; compared to the existing Neon
implementation.

For the 10-bit implementation the new 4-tap filter implementation is
up to 58% faster when coeffIdx==4 and up to 49% faster for the other
filter values compared to the new 12-bit Neon implementation. The
8-tap filter implementation is up to 23% faster when coeffIdx is 1 or
18% when it is 3 compared to the new 12-bit Neon implementation.
---
 source/common/aarch64/filter-prim.cpp | 425 ++++++++++++++++++++++----
 source/common/aarch64/mem-neon.h      |  11 +
 2 files changed, 376 insertions(+), 60 deletions(-)

diff --git a/source/common/aarch64/filter-prim.cpp b/source/common/aarch64/filter-prim.cpp
index feb4460f2..082ff8c51 100644
--- a/source/common/aarch64/filter-prim.cpp
+++ b/source/common/aarch64/filter-prim.cpp
@@ -3650,6 +3650,341 @@ void inline interp8_vert_pp_neon(const pixel *src, intptr_t srcStride, pixel *ds
     }
 }
 
+template<bool coeff4, int width, int height>
+void inline interp4_vert_ps_neon(const pixel *src, intptr_t srcStride, int16_t *dst,
+                                 intptr_t dstStride, const int16_t coeffIdx)
+{
+    const int N_TAPS = 4;
+    uint16x4_t filter = vreinterpret_u16_s16(
+        vabs_s16(vld1_s16(X265_NS::g_chromaFilter[coeffIdx])));
+    uint32_t offset_u32;
+
+    if (coeff4)
+    {
+        // The -2 is needed because we will divide the filter values by 4.
+        offset_u32 = (unsigned)-IF_INTERNAL_OFFS << (SHIFT_INTERP_PS - 2);
+    }
+    else
+    {
+        offset_u32 = (unsigned)-IF_INTERNAL_OFFS << SHIFT_INTERP_PS;
+    }
+#if X265_DEPTH == 10
+    if (!coeff4)
+    {
+        // All filter values are even, halve them to avoid needing to widen to
+        // 32-bit elements in filter kernels.
+        filter = vshr_n_u16(filter, 1);
+        offset_u32 >>= 1;
+    }
+
+    const uint16x8_t offset = vdupq_n_u16((uint16_t)offset_u32);
+#else
+    const uint32x4_t offset = vdupq_n_u32(offset_u32);
+#endif // X265_DEPTH == 10
+
+    src -= (N_TAPS / 2 - 1) * srcStride;
+
+    if (width % 8 != 0)
+    {
+        if (width == 12 || width == 6)
+        {
+            const int n_store = width == 12 ? 8 : 6;
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in0[7];
+            load_u16x8xn<3>(s, srcStride, in0);
+            s += 3 * srcStride;
+
+            for (int row = 0; row + 4 <= height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in0 + 3);
+
+                int16x8_t res[4];
+                filter4_ps_u16x8<coeff4>(in0 + 0, filter, offset, res[0]);
+                filter4_ps_u16x8<coeff4>(in0 + 1, filter, offset, res[1]);
+                filter4_ps_u16x8<coeff4>(in0 + 2, filter, offset, res[2]);
+                filter4_ps_u16x8<coeff4>(in0 + 3, filter, offset, res[3]);
+
+                store_s16xnxm<n_store, 4>(res, d, dstStride);
+
+                in0[0] = in0[4];
+                in0[1] = in0[5];
+                in0[2] = in0[6];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            if (width == 6)
+            {
+                return;
+            }
+
+            src += 8;
+            dst += 8;
+        }
+
+        const int n_store = width > 4 ? 4 : width;
+
+        uint16x4_t in1[7];
+        load_u16x4xn<3>(src, srcStride, in1);
+        src += 3 * srcStride;
+
+        for (int row = 0; row + 4 <= height; row += 4)
+        {
+            load_u16x4xn<4>(src, srcStride, in1 + 3);
+
+            int16x4_t res[4];
+            filter4_ps_u16x4<coeff4>(in1 + 0, filter, offset, res[0]);
+            filter4_ps_u16x4<coeff4>(in1 + 1, filter, offset, res[1]);
+            filter4_ps_u16x4<coeff4>(in1 + 2, filter, offset, res[2]);
+            filter4_ps_u16x4<coeff4>(in1 + 3, filter, offset, res[3]);
+
+            store_s16xnxm<n_store, 4>(res, dst, dstStride);
+
+            in1[0] = in1[4];
+            in1[1] = in1[5];
+            in1[2] = in1[6];
+
+            src += 4 * srcStride;
+            dst += 4 * dstStride;
+        }
+
+        if (height & 2)
+        {
+            load_u16x4xn<2>(src, srcStride, in1 + 3);
+
+            int16x4_t res[2];
+            filter4_ps_u16x4<coeff4>(in1 + 0, filter, offset, res[0]);
+            filter4_ps_u16x4<coeff4>(in1 + 1, filter, offset, res[1]);
+
+            store_s16xnxm<n_store, 2>(res, dst, dstStride);
+        }
+    }
+    else
+    {
+        for (int col = 0; col < width; col += 8)
+        {
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in[7];
+            load_u16x8xn<3>(s, srcStride, in);
+            s += 3 * srcStride;
+
+            for (int row = 0; row + 4 <= height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in + 3);
+
+                int16x8_t res[4];
+                filter4_ps_u16x8<coeff4>(in + 0, filter, offset, res[0]);
+                filter4_ps_u16x8<coeff4>(in + 1, filter, offset, res[1]);
+                filter4_ps_u16x8<coeff4>(in + 2, filter, offset, res[2]);
+                filter4_ps_u16x8<coeff4>(in + 3, filter, offset, res[3]);
+
+                store_s16x8xn<4>(d, dstStride, res);
+
+                in[0] = in[4];
+                in[1] = in[5];
+                in[2] = in[6];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            if (height & 2)
+            {
+                load_u16x8xn<2>(s, srcStride, in + 3);
+
+                int16x8_t res[2];
+                filter4_ps_u16x8<coeff4>(in + 0, filter, offset, res[0]);
+                filter4_ps_u16x8<coeff4>(in + 1, filter, offset, res[1]);
+
+                store_s16x8xn<2>(d, dstStride, res);
+            }
+
+            src += 8;
+            dst += 8;
+        }
+    }
+}
+
+template<int coeffIdx, int width, int height>
+void inline interp8_vert_ps_neon(const pixel *src, intptr_t srcStride, int16_t *dst,
+                                 intptr_t dstStride)
+{
+    const int N_TAPS = 8;
+    const uint16x8_t filter =
+        vreinterpretq_u16_s16(vabsq_s16(vld1q_s16(X265_NS::g_lumaFilter[coeffIdx])));
+    uint32x4_t offset =
+        vdupq_n_u32((unsigned)-IF_INTERNAL_OFFS << SHIFT_INTERP_PS);
+
+    src -= (N_TAPS / 2 - 1) * srcStride;
+
+    if (width % 8 != 0)
+    {
+        const uint16_t *s = src;
+        int16_t *d = dst;
+
+        if (width == 12)
+        {
+            uint16x8_t in[11];
+            load_u16x8xn<7>(s, srcStride, in);
+            s += 7 * srcStride;
+
+            for (int row = 0; row < height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in + 7);
+
+                int16x8_t res[4];
+                filter8_ps_u16x8<coeffIdx>(in + 0, res[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 1, res[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 2, res[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 3, res[3], offset, filter);
+
+                store_s16x8xn<4>(d, dstStride, res);
+
+                in[0] = in[4];
+                in[1] = in[5];
+                in[2] = in[6];
+                in[3] = in[7];
+                in[4] = in[8];
+                in[5] = in[9];
+                in[6] = in[10];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            s = src + 8;
+            d = dst + 8;
+        }
+
+        uint16x4_t in[11];
+        load_u16x4xn<7>(s, srcStride, in);
+        s += 7 * srcStride;
+
+        for (int row = 0; row < height; row += 4)
+        {
+            load_u16x4xn<4>(s, srcStride, in + 7);
+
+            int16x4_t res[4];
+            filter8_ps_u16x4<coeffIdx>(in + 0, res[0], offset, filter);
+            filter8_ps_u16x4<coeffIdx>(in + 1, res[1], offset, filter);
+            filter8_ps_u16x4<coeffIdx>(in + 2, res[2], offset, filter);
+            filter8_ps_u16x4<coeffIdx>(in + 3, res[3], offset, filter);
+
+            store_s16x4xn<4>(d, dstStride, res);
+
+            in[0] = in[4];
+            in[1] = in[5];
+            in[2] = in[6];
+            in[3] = in[7];
+            in[4] = in[8];
+            in[5] = in[9];
+            in[6] = in[10];
+
+            s += 4 * srcStride;
+            d += 4 * dstStride;
+        }
+    }
+    else if (width % 16 != 0)
+    {
+        for (int col = 0; col < width; col += 8)
+        {
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in[11];
+            load_u16x8xn<7>(s, srcStride, in);
+            s += 7 * srcStride;
+
+            for (int row = 0; row < height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in + 7);
+
+                int16x8_t res[4];
+                filter8_ps_u16x8<coeffIdx>(in + 0, res[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 1, res[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 2, res[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 3, res[3], offset, filter);
+
+                store_s16x8xn<4>(d, dstStride, res);
+
+                in[0] = in[4];
+                in[1] = in[5];
+                in[2] = in[6];
+                in[3] = in[7];
+                in[4] = in[8];
+                in[5] = in[9];
+                in[6] = in[10];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            src += 8;
+            dst += 8;
+        }
+    }
+    else
+    {
+        for (int col = 0; col < width; col += 16)
+        {
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in0[11], in1[11];
+            load_u16x8xn<7>(s + 0, srcStride, in0);
+            load_u16x8xn<7>(s + 8, srcStride, in1);
+            s += 7 * srcStride;
+
+            for (int row = 0; row < height; row += 4)
+            {
+                load_u16x8xn<4>(s + 0, srcStride, in0 + 7);
+                load_u16x8xn<4>(s + 8, srcStride, in1 + 7);
+
+                int16x8_t res0[4], res1[4];
+                filter8_ps_u16x8<coeffIdx>(in0 + 0, res0[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in0 + 1, res0[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in0 + 2, res0[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in0 + 3, res0[3], offset, filter);
+
+                filter8_ps_u16x8<coeffIdx>(in1 + 0, res1[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in1 + 1, res1[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in1 + 2, res1[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in1 + 3, res1[3], offset, filter);
+
+                store_s16x8xn<4>(d + 0, dstStride, res0);
+                store_s16x8xn<4>(d + 8, dstStride, res1);
+
+                in0[0] = in0[4];
+                in0[1] = in0[5];
+                in0[2] = in0[6];
+                in0[3] = in0[7];
+                in0[4] = in0[8];
+                in0[5] = in0[9];
+                in0[6] = in0[10];
+
+                in1[0] = in1[4];
+                in1[1] = in1[5];
+                in1[2] = in1[6];
+                in1[3] = in1[7];
+                in1[4] = in1[8];
+                in1[5] = in1[9];
+                in1[6] = in1[10];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            src += 16;
+            dst += 16;
+        }
+    }
+}
+
 #endif // !HIGH_BIT_DEPTH
 }
 
@@ -3788,63 +4123,9 @@ void interp_vert_pp_neon(const pixel *src, intptr_t srcStride, pixel *dst,
     }
 }
 
-#if HIGH_BIT_DEPTH
-
 template<int N, int width, int height>
-void interp_vert_ps_neon(const uint16_t *src, intptr_t srcStride, int16_t *dst, intptr_t dstStride, int coeffIdx)
-{
-    const int16_t *c = (N == 4) ? g_chromaFilter[coeffIdx] : g_lumaFilter[coeffIdx];
-    int offset = (unsigned) - IF_INTERNAL_OFFS << SHIFT_INTERP_PS;
-    src -= (N / 2 - 1) * srcStride;
-
-    int16x8_t vc = vld1q_s16(c);
-    int32x4_t low_vc = vmovl_s16(vget_low_s16(vc));
-    int32x4_t high_vc = vmovl_s16(vget_high_s16(vc));
-
-    const int32x4_t voffset = vdupq_n_s32(offset);
-
-    int row, col;
-    for (row = 0; row < height; row++)
-    {
-        for (col = 0; col < width; col += 4)
-        {
-            int32x4_t vsum;
-
-            int32x4_t input[N];
-
-            for (int i = 0; i < N; i++)
-            {
-                uint16x4_t in_tmp = vld1_u16(src + col + i * srcStride);
-                input[i] = vreinterpretq_s32_u32(vmovl_u16(in_tmp));
-            }
-            vsum = voffset;
-
-            vsum = vmlaq_laneq_s32(vsum, input[0], low_vc, 0);
-            vsum = vmlaq_laneq_s32(vsum, input[1], low_vc, 1);
-            vsum = vmlaq_laneq_s32(vsum, input[2], low_vc, 2);
-            vsum = vmlaq_laneq_s32(vsum, input[3], low_vc, 3);
-
-            if (N == 8)
-            {
-                int32x4_t vsum1 = vmulq_laneq_s32(input[4], high_vc, 0);
-                vsum1 = vmlaq_laneq_s32(vsum1, input[5], high_vc, 1);
-                vsum1 = vmlaq_laneq_s32(vsum1, input[6], high_vc, 2);
-                vsum1 = vmlaq_laneq_s32(vsum1, input[7], high_vc, 3);
-                vsum = vaddq_s32(vsum, vsum1);
-            }
-
-            vst1_s16(dst + col, vshrn_n_s32(vsum, SHIFT_INTERP_PS));
-        }
-
-        src += srcStride;
-        dst += dstStride;
-    }
-}
-
-#else
-
-template<int N, int width, int height>
-void interp_vert_ps_neon(const uint8_t *src, intptr_t srcStride, int16_t *dst, intptr_t dstStride, int coeffIdx)
+void interp_vert_ps_neon(const pixel *src, intptr_t srcStride, int16_t *dst,
+                         intptr_t dstStride, int coeffIdx)
 {
     if (N == 8)
     {
@@ -3877,10 +4158,6 @@ void interp_vert_ps_neon(const uint8_t *src, intptr_t srcStride, int16_t *dst, i
     }
 }
 
-#endif
-
-
-
 #if HIGH_BIT_DEPTH
 template<int N, int width, int height>
 void interp_vert_sp_neon(const int16_t *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)
@@ -4293,6 +4570,34 @@ void setupFilterPrimitives_neon(EncoderPrimitives &p)
     p.pu[LUMA_4x8].luma_vpp                                 = interp_vert_pp_neon<8, 4, 8>;
     p.pu[LUMA_4x16].luma_vpp                                = interp_vert_pp_neon<8, 4, 16>;
     p.pu[LUMA_12x16].luma_vpp                               = interp_vert_pp_neon<8, 12, 16>;
+
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_2x4].filter_vps   = interp_vert_ps_neon<4, 2, 4>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_2x8].filter_vps   = interp_vert_ps_neon<4, 2, 8>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x2].filter_vps   = interp_vert_ps_neon<4, 4, 2>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].filter_vps   = interp_vert_ps_neon<4, 4, 4>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x8].filter_vps   = interp_vert_ps_neon<4, 4, 8>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x16].filter_vps  = interp_vert_ps_neon<4, 4, 16>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_6x8].filter_vps   = interp_vert_ps_neon<4, 6, 8>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_12x16].filter_vps = interp_vert_ps_neon<4, 12, 16>;
+
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_2x8].filter_vps   = interp_vert_ps_neon<4, 2, 8>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_2x16].filter_vps  = interp_vert_ps_neon<4, 2, 16>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x4].filter_vps   = interp_vert_ps_neon<4, 4, 4>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x8].filter_vps   = interp_vert_ps_neon<4, 4, 8>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x16].filter_vps  = interp_vert_ps_neon<4, 4, 16>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x32].filter_vps  = interp_vert_ps_neon<4, 4, 32>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_6x16].filter_vps  = interp_vert_ps_neon<4, 6, 16>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_12x32].filter_vps = interp_vert_ps_neon<4, 12, 32>;
+
+    p.chroma[X265_CSP_I444].pu[LUMA_4x4].filter_vps         = interp_vert_ps_neon<4, 4, 4>;
+    p.chroma[X265_CSP_I444].pu[LUMA_4x8].filter_vps         = interp_vert_ps_neon<4, 4, 8>;
+    p.chroma[X265_CSP_I444].pu[LUMA_4x16].filter_vps        = interp_vert_ps_neon<4, 4, 16>;
+    p.chroma[X265_CSP_I444].pu[LUMA_12x16].filter_vps       = interp_vert_ps_neon<4, 12, 16>;
+
+    p.pu[LUMA_4x4].luma_vps                                 = interp_vert_ps_neon<8, 4, 4>;
+    p.pu[LUMA_4x8].luma_vps                                 = interp_vert_ps_neon<8, 4, 8>;
+    p.pu[LUMA_4x16].luma_vps                                = interp_vert_ps_neon<8, 4, 16>;
+    p.pu[LUMA_12x16].luma_vps                               = interp_vert_ps_neon<8, 12, 16>;
 #endif // HIGH_BIT_DEPTH
 }
 
diff --git a/source/common/aarch64/mem-neon.h b/source/common/aarch64/mem-neon.h
index 7386cd736..5ebe8d5b0 100644
--- a/source/common/aarch64/mem-neon.h
+++ b/source/common/aarch64/mem-neon.h
@@ -355,6 +355,17 @@ static void inline store_s16xnxm(const int16x8_t *src, int16_t *dst,
     }
 }
 
+template<int N, int M>
+static void inline store_s16xnxm(const int16x4_t *src, int16_t *dst,
+                                 intptr_t dst_stride)
+{
+    switch (N)
+    {
+    case 2: return store_s16x2xn<M>(dst, dst_stride, src);
+    case 4: return store_s16x4xn<M>(dst, dst_stride, src);
+    }
+}
+
 template<int N, int M>
 static void inline store_u16xnxm(uint16_t *dst, intptr_t dst_stride,
                                  const uint16x8_t *src)
-- 
2.39.5 (Apple Git-154)

-------------- next part --------------
>From 76f9017e6475e447274b5ffac24b561ef6ef82d8 Mon Sep 17 00:00:00 2001
Message-Id: <76f9017e6475e447274b5ffac24b561ef6ef82d8.1741721714.git.gerdazsejke.more at arm.com>
In-Reply-To: <cover.1741721714.git.gerdazsejke.more at arm.com>
References: <cover.1741721714.git.gerdazsejke.more at arm.com>
From: Gerda Zsejke More <gerdazsejke.more at arm.com>
Date: Mon, 3 Feb 2025 10:27:51 +0100
Subject: [PATCH v2 06/10] AArch64: Optimise HBD interp_vert_ps_neon

Optimise the HBD 4-tap and 8-tap Neon implementations of
interp_vert_ps_neon and extend these functions to support all CHROMA
and LUMA block sizes respectively.

For the 12-bit implementation the new 4-tap filter implementation is
up to 54% faster when coeffIdx==4 and up to 28% faster for the other
filter values compared to the existing Neon implementation. The new
8-tap filter implementation is up to 55% faster when coeffIdx==1,
64% when it is 2, and 58% when it is 3; compared to the existing Neon
implementation.

For the 10-bit implementation the new 4-tap filter implementation is
up to 58% faster when coeffIdx==4 and up to 49% faster for the other
filter values compared to the new 12-bit Neon implementation. The
8-tap filter implementation is up to 23% faster when coeffIdx is 1 or
18% when it is 3 compared to the new 12-bit Neon implementation.
---
 source/common/aarch64/filter-prim.cpp | 425 ++++++++++++++++++++++----
 source/common/aarch64/mem-neon.h      |  11 +
 2 files changed, 376 insertions(+), 60 deletions(-)

diff --git a/source/common/aarch64/filter-prim.cpp b/source/common/aarch64/filter-prim.cpp
index feb4460f2..082ff8c51 100644
--- a/source/common/aarch64/filter-prim.cpp
+++ b/source/common/aarch64/filter-prim.cpp
@@ -3650,6 +3650,341 @@ void inline interp8_vert_pp_neon(const pixel *src, intptr_t srcStride, pixel *ds
     }
 }
 
+template<bool coeff4, int width, int height>
+void inline interp4_vert_ps_neon(const pixel *src, intptr_t srcStride, int16_t *dst,
+                                 intptr_t dstStride, const int16_t coeffIdx)
+{
+    const int N_TAPS = 4;
+    uint16x4_t filter = vreinterpret_u16_s16(
+        vabs_s16(vld1_s16(X265_NS::g_chromaFilter[coeffIdx])));
+    uint32_t offset_u32;
+
+    if (coeff4)
+    {
+        // The -2 is needed because we will divide the filter values by 4.
+        offset_u32 = (unsigned)-IF_INTERNAL_OFFS << (SHIFT_INTERP_PS - 2);
+    }
+    else
+    {
+        offset_u32 = (unsigned)-IF_INTERNAL_OFFS << SHIFT_INTERP_PS;
+    }
+#if X265_DEPTH == 10
+    if (!coeff4)
+    {
+        // All filter values are even, halve them to avoid needing to widen to
+        // 32-bit elements in filter kernels.
+        filter = vshr_n_u16(filter, 1);
+        offset_u32 >>= 1;
+    }
+
+    const uint16x8_t offset = vdupq_n_u16((uint16_t)offset_u32);
+#else
+    const uint32x4_t offset = vdupq_n_u32(offset_u32);
+#endif // X265_DEPTH == 10
+
+    src -= (N_TAPS / 2 - 1) * srcStride;
+
+    if (width % 8 != 0)
+    {
+        if (width == 12 || width == 6)
+        {
+            const int n_store = width == 12 ? 8 : 6;
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in0[7];
+            load_u16x8xn<3>(s, srcStride, in0);
+            s += 3 * srcStride;
+
+            for (int row = 0; row + 4 <= height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in0 + 3);
+
+                int16x8_t res[4];
+                filter4_ps_u16x8<coeff4>(in0 + 0, filter, offset, res[0]);
+                filter4_ps_u16x8<coeff4>(in0 + 1, filter, offset, res[1]);
+                filter4_ps_u16x8<coeff4>(in0 + 2, filter, offset, res[2]);
+                filter4_ps_u16x8<coeff4>(in0 + 3, filter, offset, res[3]);
+
+                store_s16xnxm<n_store, 4>(res, d, dstStride);
+
+                in0[0] = in0[4];
+                in0[1] = in0[5];
+                in0[2] = in0[6];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            if (width == 6)
+            {
+                return;
+            }
+
+            src += 8;
+            dst += 8;
+        }
+
+        const int n_store = width > 4 ? 4 : width;
+
+        uint16x4_t in1[7];
+        load_u16x4xn<3>(src, srcStride, in1);
+        src += 3 * srcStride;
+
+        for (int row = 0; row + 4 <= height; row += 4)
+        {
+            load_u16x4xn<4>(src, srcStride, in1 + 3);
+
+            int16x4_t res[4];
+            filter4_ps_u16x4<coeff4>(in1 + 0, filter, offset, res[0]);
+            filter4_ps_u16x4<coeff4>(in1 + 1, filter, offset, res[1]);
+            filter4_ps_u16x4<coeff4>(in1 + 2, filter, offset, res[2]);
+            filter4_ps_u16x4<coeff4>(in1 + 3, filter, offset, res[3]);
+
+            store_s16xnxm<n_store, 4>(res, dst, dstStride);
+
+            in1[0] = in1[4];
+            in1[1] = in1[5];
+            in1[2] = in1[6];
+
+            src += 4 * srcStride;
+            dst += 4 * dstStride;
+        }
+
+        if (height & 2)
+        {
+            load_u16x4xn<2>(src, srcStride, in1 + 3);
+
+            int16x4_t res[2];
+            filter4_ps_u16x4<coeff4>(in1 + 0, filter, offset, res[0]);
+            filter4_ps_u16x4<coeff4>(in1 + 1, filter, offset, res[1]);
+
+            store_s16xnxm<n_store, 2>(res, dst, dstStride);
+        }
+    }
+    else
+    {
+        for (int col = 0; col < width; col += 8)
+        {
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in[7];
+            load_u16x8xn<3>(s, srcStride, in);
+            s += 3 * srcStride;
+
+            for (int row = 0; row + 4 <= height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in + 3);
+
+                int16x8_t res[4];
+                filter4_ps_u16x8<coeff4>(in + 0, filter, offset, res[0]);
+                filter4_ps_u16x8<coeff4>(in + 1, filter, offset, res[1]);
+                filter4_ps_u16x8<coeff4>(in + 2, filter, offset, res[2]);
+                filter4_ps_u16x8<coeff4>(in + 3, filter, offset, res[3]);
+
+                store_s16x8xn<4>(d, dstStride, res);
+
+                in[0] = in[4];
+                in[1] = in[5];
+                in[2] = in[6];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            if (height & 2)
+            {
+                load_u16x8xn<2>(s, srcStride, in + 3);
+
+                int16x8_t res[2];
+                filter4_ps_u16x8<coeff4>(in + 0, filter, offset, res[0]);
+                filter4_ps_u16x8<coeff4>(in + 1, filter, offset, res[1]);
+
+                store_s16x8xn<2>(d, dstStride, res);
+            }
+
+            src += 8;
+            dst += 8;
+        }
+    }
+}
+
+template<int coeffIdx, int width, int height>
+void inline interp8_vert_ps_neon(const pixel *src, intptr_t srcStride, int16_t *dst,
+                                 intptr_t dstStride)
+{
+    const int N_TAPS = 8;
+    const uint16x8_t filter =
+        vreinterpretq_u16_s16(vabsq_s16(vld1q_s16(X265_NS::g_lumaFilter[coeffIdx])));
+    uint32x4_t offset =
+        vdupq_n_u32((unsigned)-IF_INTERNAL_OFFS << SHIFT_INTERP_PS);
+
+    src -= (N_TAPS / 2 - 1) * srcStride;
+
+    if (width % 8 != 0)
+    {
+        const uint16_t *s = src;
+        int16_t *d = dst;
+
+        if (width == 12)
+        {
+            uint16x8_t in[11];
+            load_u16x8xn<7>(s, srcStride, in);
+            s += 7 * srcStride;
+
+            for (int row = 0; row < height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in + 7);
+
+                int16x8_t res[4];
+                filter8_ps_u16x8<coeffIdx>(in + 0, res[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 1, res[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 2, res[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 3, res[3], offset, filter);
+
+                store_s16x8xn<4>(d, dstStride, res);
+
+                in[0] = in[4];
+                in[1] = in[5];
+                in[2] = in[6];
+                in[3] = in[7];
+                in[4] = in[8];
+                in[5] = in[9];
+                in[6] = in[10];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            s = src + 8;
+            d = dst + 8;
+        }
+
+        uint16x4_t in[11];
+        load_u16x4xn<7>(s, srcStride, in);
+        s += 7 * srcStride;
+
+        for (int row = 0; row < height; row += 4)
+        {
+            load_u16x4xn<4>(s, srcStride, in + 7);
+
+            int16x4_t res[4];
+            filter8_ps_u16x4<coeffIdx>(in + 0, res[0], offset, filter);
+            filter8_ps_u16x4<coeffIdx>(in + 1, res[1], offset, filter);
+            filter8_ps_u16x4<coeffIdx>(in + 2, res[2], offset, filter);
+            filter8_ps_u16x4<coeffIdx>(in + 3, res[3], offset, filter);
+
+            store_s16x4xn<4>(d, dstStride, res);
+
+            in[0] = in[4];
+            in[1] = in[5];
+            in[2] = in[6];
+            in[3] = in[7];
+            in[4] = in[8];
+            in[5] = in[9];
+            in[6] = in[10];
+
+            s += 4 * srcStride;
+            d += 4 * dstStride;
+        }
+    }
+    else if (width % 16 != 0)
+    {
+        for (int col = 0; col < width; col += 8)
+        {
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in[11];
+            load_u16x8xn<7>(s, srcStride, in);
+            s += 7 * srcStride;
+
+            for (int row = 0; row < height; row += 4)
+            {
+                load_u16x8xn<4>(s, srcStride, in + 7);
+
+                int16x8_t res[4];
+                filter8_ps_u16x8<coeffIdx>(in + 0, res[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 1, res[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 2, res[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in + 3, res[3], offset, filter);
+
+                store_s16x8xn<4>(d, dstStride, res);
+
+                in[0] = in[4];
+                in[1] = in[5];
+                in[2] = in[6];
+                in[3] = in[7];
+                in[4] = in[8];
+                in[5] = in[9];
+                in[6] = in[10];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            src += 8;
+            dst += 8;
+        }
+    }
+    else
+    {
+        for (int col = 0; col < width; col += 16)
+        {
+            const uint16_t *s = src;
+            int16_t *d = dst;
+
+            uint16x8_t in0[11], in1[11];
+            load_u16x8xn<7>(s + 0, srcStride, in0);
+            load_u16x8xn<7>(s + 8, srcStride, in1);
+            s += 7 * srcStride;
+
+            for (int row = 0; row < height; row += 4)
+            {
+                load_u16x8xn<4>(s + 0, srcStride, in0 + 7);
+                load_u16x8xn<4>(s + 8, srcStride, in1 + 7);
+
+                int16x8_t res0[4], res1[4];
+                filter8_ps_u16x8<coeffIdx>(in0 + 0, res0[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in0 + 1, res0[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in0 + 2, res0[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in0 + 3, res0[3], offset, filter);
+
+                filter8_ps_u16x8<coeffIdx>(in1 + 0, res1[0], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in1 + 1, res1[1], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in1 + 2, res1[2], offset, filter);
+                filter8_ps_u16x8<coeffIdx>(in1 + 3, res1[3], offset, filter);
+
+                store_s16x8xn<4>(d + 0, dstStride, res0);
+                store_s16x8xn<4>(d + 8, dstStride, res1);
+
+                in0[0] = in0[4];
+                in0[1] = in0[5];
+                in0[2] = in0[6];
+                in0[3] = in0[7];
+                in0[4] = in0[8];
+                in0[5] = in0[9];
+                in0[6] = in0[10];
+
+                in1[0] = in1[4];
+                in1[1] = in1[5];
+                in1[2] = in1[6];
+                in1[3] = in1[7];
+                in1[4] = in1[8];
+                in1[5] = in1[9];
+                in1[6] = in1[10];
+
+                s += 4 * srcStride;
+                d += 4 * dstStride;
+            }
+
+            src += 16;
+            dst += 16;
+        }
+    }
+}
+
 #endif // !HIGH_BIT_DEPTH
 }
 
@@ -3788,63 +4123,9 @@ void interp_vert_pp_neon(const pixel *src, intptr_t srcStride, pixel *dst,
     }
 }
 
-#if HIGH_BIT_DEPTH
-
 template<int N, int width, int height>
-void interp_vert_ps_neon(const uint16_t *src, intptr_t srcStride, int16_t *dst, intptr_t dstStride, int coeffIdx)
-{
-    const int16_t *c = (N == 4) ? g_chromaFilter[coeffIdx] : g_lumaFilter[coeffIdx];
-    int offset = (unsigned) - IF_INTERNAL_OFFS << SHIFT_INTERP_PS;
-    src -= (N / 2 - 1) * srcStride;
-
-    int16x8_t vc = vld1q_s16(c);
-    int32x4_t low_vc = vmovl_s16(vget_low_s16(vc));
-    int32x4_t high_vc = vmovl_s16(vget_high_s16(vc));
-
-    const int32x4_t voffset = vdupq_n_s32(offset);
-
-    int row, col;
-    for (row = 0; row < height; row++)
-    {
-        for (col = 0; col < width; col += 4)
-        {
-            int32x4_t vsum;
-
-            int32x4_t input[N];
-
-            for (int i = 0; i < N; i++)
-            {
-                uint16x4_t in_tmp = vld1_u16(src + col + i * srcStride);
-                input[i] = vreinterpretq_s32_u32(vmovl_u16(in_tmp));
-            }
-            vsum = voffset;
-
-            vsum = vmlaq_laneq_s32(vsum, input[0], low_vc, 0);
-            vsum = vmlaq_laneq_s32(vsum, input[1], low_vc, 1);
-            vsum = vmlaq_laneq_s32(vsum, input[2], low_vc, 2);
-            vsum = vmlaq_laneq_s32(vsum, input[3], low_vc, 3);
-
-            if (N == 8)
-            {
-                int32x4_t vsum1 = vmulq_laneq_s32(input[4], high_vc, 0);
-                vsum1 = vmlaq_laneq_s32(vsum1, input[5], high_vc, 1);
-                vsum1 = vmlaq_laneq_s32(vsum1, input[6], high_vc, 2);
-                vsum1 = vmlaq_laneq_s32(vsum1, input[7], high_vc, 3);
-                vsum = vaddq_s32(vsum, vsum1);
-            }
-
-            vst1_s16(dst + col, vshrn_n_s32(vsum, SHIFT_INTERP_PS));
-        }
-
-        src += srcStride;
-        dst += dstStride;
-    }
-}
-
-#else
-
-template<int N, int width, int height>
-void interp_vert_ps_neon(const uint8_t *src, intptr_t srcStride, int16_t *dst, intptr_t dstStride, int coeffIdx)
+void interp_vert_ps_neon(const pixel *src, intptr_t srcStride, int16_t *dst,
+                         intptr_t dstStride, int coeffIdx)
 {
     if (N == 8)
     {
@@ -3877,10 +4158,6 @@ void interp_vert_ps_neon(const uint8_t *src, intptr_t srcStride, int16_t *dst, i
     }
 }
 
-#endif
-
-
-
 #if HIGH_BIT_DEPTH
 template<int N, int width, int height>
 void interp_vert_sp_neon(const int16_t *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)
@@ -4293,6 +4570,34 @@ void setupFilterPrimitives_neon(EncoderPrimitives &p)
     p.pu[LUMA_4x8].luma_vpp                                 = interp_vert_pp_neon<8, 4, 8>;
     p.pu[LUMA_4x16].luma_vpp                                = interp_vert_pp_neon<8, 4, 16>;
     p.pu[LUMA_12x16].luma_vpp                               = interp_vert_pp_neon<8, 12, 16>;
+
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_2x4].filter_vps   = interp_vert_ps_neon<4, 2, 4>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_2x8].filter_vps   = interp_vert_ps_neon<4, 2, 8>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x2].filter_vps   = interp_vert_ps_neon<4, 4, 2>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].filter_vps   = interp_vert_ps_neon<4, 4, 4>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x8].filter_vps   = interp_vert_ps_neon<4, 4, 8>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_4x16].filter_vps  = interp_vert_ps_neon<4, 4, 16>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_6x8].filter_vps   = interp_vert_ps_neon<4, 6, 8>;
+    p.chroma[X265_CSP_I420].pu[CHROMA_420_12x16].filter_vps = interp_vert_ps_neon<4, 12, 16>;
+
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_2x8].filter_vps   = interp_vert_ps_neon<4, 2, 8>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_2x16].filter_vps  = interp_vert_ps_neon<4, 2, 16>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x4].filter_vps   = interp_vert_ps_neon<4, 4, 4>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x8].filter_vps   = interp_vert_ps_neon<4, 4, 8>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x16].filter_vps  = interp_vert_ps_neon<4, 4, 16>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_4x32].filter_vps  = interp_vert_ps_neon<4, 4, 32>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_6x16].filter_vps  = interp_vert_ps_neon<4, 6, 16>;
+    p.chroma[X265_CSP_I422].pu[CHROMA_422_12x32].filter_vps = interp_vert_ps_neon<4, 12, 32>;
+
+    p.chroma[X265_CSP_I444].pu[LUMA_4x4].filter_vps         = interp_vert_ps_neon<4, 4, 4>;
+    p.chroma[X265_CSP_I444].pu[LUMA_4x8].filter_vps         = interp_vert_ps_neon<4, 4, 8>;
+    p.chroma[X265_CSP_I444].pu[LUMA_4x16].filter_vps        = interp_vert_ps_neon<4, 4, 16>;
+    p.chroma[X265_CSP_I444].pu[LUMA_12x16].filter_vps       = interp_vert_ps_neon<4, 12, 16>;
+
+    p.pu[LUMA_4x4].luma_vps                                 = interp_vert_ps_neon<8, 4, 4>;
+    p.pu[LUMA_4x8].luma_vps                                 = interp_vert_ps_neon<8, 4, 8>;
+    p.pu[LUMA_4x16].luma_vps                                = interp_vert_ps_neon<8, 4, 16>;
+    p.pu[LUMA_12x16].luma_vps                               = interp_vert_ps_neon<8, 12, 16>;
 #endif // HIGH_BIT_DEPTH
 }
 
diff --git a/source/common/aarch64/mem-neon.h b/source/common/aarch64/mem-neon.h
index 7386cd736..5ebe8d5b0 100644
--- a/source/common/aarch64/mem-neon.h
+++ b/source/common/aarch64/mem-neon.h
@@ -355,6 +355,17 @@ static void inline store_s16xnxm(const int16x8_t *src, int16_t *dst,
     }
 }
 
+template<int N, int M>
+static void inline store_s16xnxm(const int16x4_t *src, int16_t *dst,
+                                 intptr_t dst_stride)
+{
+    switch (N)
+    {
+    case 2: return store_s16x2xn<M>(dst, dst_stride, src);
+    case 4: return store_s16x4xn<M>(dst, dst_stride, src);
+    }
+}
+
 template<int N, int M>
 static void inline store_u16xnxm(uint16_t *dst, intptr_t dst_stride,
                                  const uint16x8_t *src)
-- 
2.39.5 (Apple Git-154)



More information about the x265-devel mailing list