Facilitate eob support in inverse transform for neon Add implementation of partial inv txfm based on eob for neon version functions. For decoder, profiling results shows the cpu time of inv txfm drops around 46%. Commit ID: 48073d93c20dd2d319c0d45291d43e0dfb76c9a5 Sequence: 150 frames of night_720p30 Config: Default (compare neon with eob and existing c without eob) Change-Id: Icbe9ea2c398a0388d46d3f958c12476e42290793
diff --git a/av1/av1.cmake b/av1/av1.cmake index 45301d1..02c7908 100644 --- a/av1/av1.cmake +++ b/av1/av1.cmake
@@ -300,6 +300,8 @@ "${AOM_ROOT}/av1/common/arm/reconinter_neon.c" "${AOM_ROOT}/av1/common/arm/wiener_convolve_neon.c" "${AOM_ROOT}/av1/common/arm/selfguided_neon.c" + "${AOM_ROOT}/av1/common/arm/av1_inv_txfm_neon.c" + "${AOM_ROOT}/av1/common/arm/av1_inv_txfm_neon.h" "${AOM_ROOT}/av1/common/cdef_block_neon.c") list(APPEND AOM_AV1_ENCODER_INTRIN_SSE4_2
diff --git a/av1/common/arm/av1_inv_txfm_neon.c b/av1/common/arm/av1_inv_txfm_neon.c new file mode 100644 index 0000000..51c9914 --- /dev/null +++ b/av1/common/arm/av1_inv_txfm_neon.c
@@ -0,0 +1,844 @@ +/* + * Copyright (c) 2018, Alliance for Open Media. All rights reserved + * + * This source code is subject to the terms of the BSD 2 Clause License and + * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License + * was not distributed with this source code in the LICENSE file, you can + * obtain it at www.aomedia.org/license/software. If the Alliance for Open + * Media Patent License 1.0 was not distributed with this source code in the + * PATENTS file, you can obtain it at www.aomedia.org/license/patent. + */ + +#include "config/aom_config.h" +#include "config/aom_dsp_rtcd.h" +#include "config/av1_rtcd.h" + +#include "av1/common/av1_inv_txfm1d.h" +#include "av1/common/av1_inv_txfm1d_cfg.h" +#include "av1/common/av1_txfm.h" +#include "av1/common/enums.h" +#include "av1/common/idct.h" +#include "av1/common/arm/av1_inv_txfm_neon.h" + +static INLINE TxSetType find_TxSetType(TX_SIZE tx_size) { + const TX_SIZE tx_size_sqr_up = txsize_sqr_up_map[tx_size]; + TxSetType tx_set_type; + if (tx_size_sqr_up > TX_32X32) { + tx_set_type = EXT_TX_SET_DCTONLY; + } else if (tx_size_sqr_up == TX_32X32) { + tx_set_type = EXT_TX_SET_DCT_IDTX; + } else { + tx_set_type = EXT_TX_SET_ALL16; + } + return tx_set_type; +} + +// 1D itx types +typedef enum ATTRIBUTE_PACKED { + IDCT_1D, + IADST_1D, + IFLIPADST_1D = IADST_1D, + IIDENTITY_1D, + ITX_TYPES_1D, +} ITX_TYPE_1D; + +static const ITX_TYPE_1D vitx_1d_tab[TX_TYPES] = { + IDCT_1D, IADST_1D, IDCT_1D, IADST_1D, + IFLIPADST_1D, IDCT_1D, IFLIPADST_1D, IADST_1D, + IFLIPADST_1D, IIDENTITY_1D, IDCT_1D, IIDENTITY_1D, + IADST_1D, IIDENTITY_1D, IFLIPADST_1D, IIDENTITY_1D, +}; + +static const ITX_TYPE_1D hitx_1d_tab[TX_TYPES] = { + IDCT_1D, IDCT_1D, IADST_1D, IADST_1D, + IDCT_1D, IFLIPADST_1D, IFLIPADST_1D, IFLIPADST_1D, + IADST_1D, IIDENTITY_1D, IIDENTITY_1D, IDCT_1D, + IIDENTITY_1D, IADST_1D, IIDENTITY_1D, IFLIPADST_1D, +}; + +// 1D functions +static const transform_1d_neon lowbd_txfm_all_1d_arr[TX_SIZES][ITX_TYPES_1D] = { + { av1_idct4_new, av1_iadst4_new, av1_iidentity4_c }, + { av1_idct8_new, av1_iadst8_new, av1_iidentity8_c }, + { av1_idct16_new, av1_iadst16_new, av1_iidentity16_c }, + { av1_idct32_new, NULL, NULL }, + { av1_idct64_new, NULL, NULL }, +}; + +// Functions for blocks with eob at DC and within +// topleft 8x8, 16x16, 32x32 corner +static const transform_1d_neon + lowbd_txfm_all_1d_zeros_w8_arr[TX_SIZES][ITX_TYPES_1D][4] = { + { + { av1_idct4_new, av1_idct4_new, NULL, NULL }, + { av1_iadst4_new, av1_iadst4_new, NULL, NULL }, + { av1_iidentity4_c, av1_iidentity4_c, NULL, NULL }, + }, + { { av1_idct8_new, av1_idct8_new, NULL, NULL }, + { av1_iadst8_new, av1_iadst8_new, NULL, NULL }, + { av1_iidentity8_c, av1_iidentity8_c, NULL, NULL } }, + { + { av1_idct16_new, av1_idct16_new, av1_idct16_new, NULL }, + { av1_iadst16_new, av1_iadst16_new, av1_iadst16_new, NULL }, + { av1_iidentity16_c, av1_iidentity16_c, av1_iidentity16_c, NULL }, + }, + { { av1_idct32_new, av1_idct32_new, av1_idct32_new, av1_idct32_new }, + { NULL, NULL, NULL, NULL }, + { av1_iidentity32_c, av1_iidentity32_c, av1_iidentity32_c, + av1_iidentity32_c } }, + { { av1_idct64_new, av1_idct64_new, av1_idct64_new, av1_idct64_new }, + { NULL, NULL, NULL, NULL }, + { NULL, NULL, NULL, NULL } } + }; +static INLINE void lowbd_inv_txfm2d_add_idtx_neon(const int32_t *input, + uint8_t *output, int stride, + TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + DECLARE_ALIGNED(32, int, txfm_buf[32 * 32 + 32 + 32]); + int32_t *temp_in = txfm_buf; + + int eobx, eoby; + get_eobx_eoby_scan_default(&eobx, &eoby, tx_size, eob); + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_size_nonzero_h_div8 = (eoby + 8) >> 3; + + const int rect_type = get_rect_tx_log_ratio(txfm_size_col, txfm_size_row); + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + + const int fun_idx_x = lowbd_txfm_all_1d_zeros_idx[eobx]; + const int fun_idx_y = lowbd_txfm_all_1d_zeros_idx[eoby]; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txw_idx][hitx_1d_tab[tx_type]][fun_idx_x]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txh_idx][vitx_1d_tab[tx_type]][fun_idx_y]; + + assert(col_txfm != NULL); + assert(row_txfm != NULL); + + // row tx + int row_start = (buf_size_nonzero_h_div8 * 8); + for (int i = 0; i < row_start; i++) { + if (abs(rect_type) == 1) { + for (int j = 0; j < txfm_size_col; j++) + temp_in[j] = round_shift((int64_t)input[j] * NewInvSqrt2, NewSqrt2Bits); + row_txfm(temp_in, buf_ptr, cos_bit_row, stage_range); + } else { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + } + av1_round_shift_array(buf_ptr, txfm_size_col, -shift[0]); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + // Doing memset for the rows which are not processed in row transform. + memset(buf_ptr, 0, + sizeof(int32_t) * txfm_size_col * (txfm_size_row - row_start)); + + // col tx + for (int c = 0; c < txfm_size_col; c++) { + for (r = 0; r < txfm_size_row; ++r) temp_in[r] = buf[r * txfm_size_col + c]; + + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } +} + +static INLINE void lowbd_inv_txfm2d_add_v_identity_neon( + const int32_t *input, uint8_t *output, int stride, TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + DECLARE_ALIGNED(32, int, txfm_buf[32 * 32 + 32 + 32]); + int32_t *temp_in = txfm_buf; + + int eobx, eoby; + get_eobx_eoby_scan_v_identity(&eobx, &eoby, tx_size, eob); + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_size_nonzero_h_div8 = (eoby + 8) >> 3; + + const int rect_type = get_rect_tx_log_ratio(txfm_size_col, txfm_size_row); + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + + const int fun_idx_x = lowbd_txfm_all_1d_zeros_idx[eobx]; + const int fun_idx_y = lowbd_txfm_all_1d_zeros_idx[eoby]; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txw_idx][hitx_1d_tab[tx_type]][fun_idx_x]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txh_idx][vitx_1d_tab[tx_type]][fun_idx_y]; + + assert(col_txfm != NULL); + assert(row_txfm != NULL); + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + // row tx + int row_start = (buf_size_nonzero_h_div8 * 8); + for (int i = 0; i < row_start; i++) { + if (abs(rect_type) == 1) { + for (int j = 0; j < txfm_size_col; j++) + temp_in[j] = round_shift((int64_t)input[j] * NewInvSqrt2, NewSqrt2Bits); + row_txfm(temp_in, buf_ptr, cos_bit_row, stage_range); + } else { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + } + av1_round_shift_array(buf_ptr, txfm_size_col, -shift[0]); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + // Doing memset for the rows which are not processed in row transform. + memset(buf_ptr, 0, + sizeof(int32_t) * txfm_size_col * (txfm_size_row - row_start)); + + // col tx + for (int c = 0; c < txfm_size_col; c++) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +static INLINE void lowbd_inv_txfm2d_add_h_identity_neon( + const int32_t *input, uint8_t *output, int stride, TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + DECLARE_ALIGNED(32, int, txfm_buf[32 * 32 + 32 + 32]); + int32_t *temp_in = txfm_buf; + + int eobx, eoby; + get_eobx_eoby_scan_h_identity(&eobx, &eoby, tx_size, eob); + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_size_nonzero_h_div8 = (eoby + 8) >> 3; + + const int rect_type = get_rect_tx_log_ratio(txfm_size_col, txfm_size_row); + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + + const int fun_idx_x = lowbd_txfm_all_1d_zeros_idx[eobx]; + const int fun_idx_y = lowbd_txfm_all_1d_zeros_idx[eoby]; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txw_idx][hitx_1d_tab[tx_type]][fun_idx_x]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txh_idx][vitx_1d_tab[tx_type]][fun_idx_y]; + + assert(col_txfm != NULL); + assert(row_txfm != NULL); + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + // row tx + int row_start = (buf_size_nonzero_h_div8 * 8); + for (int i = 0; i < row_start; i++) { + if (abs(rect_type) == 1) { + for (int j = 0; j < txfm_size_col; j++) + temp_in[j] = round_shift((int64_t)input[j] * NewInvSqrt2, NewSqrt2Bits); + row_txfm(temp_in, buf_ptr, cos_bit_row, stage_range); + } else { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + } + av1_round_shift_array(buf_ptr, txfm_size_col, -shift[0]); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + // Doing memset for the rows which are not processed in row transform. + memset(buf_ptr, 0, + sizeof(int32_t) * txfm_size_col * (txfm_size_row - row_start)); + + // col tx + for (int c = 0; c < txfm_size_col; c++) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +static INLINE void lowbd_inv_txfm2d_add_4x4_neon(const int32_t *input, + uint8_t *output, int stride, + TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + (void)eob; + DECLARE_ALIGNED(32, int, txfm_buf[4 * 4 + 8 + 8]); + int32_t *temp_in = txfm_buf; + + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_arr[txw_idx][hitx_1d_tab[tx_type]]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_arr[txh_idx][vitx_1d_tab[tx_type]]; + + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + for (int i = 0; i < txfm_size_row; i++) { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + for (int c = 0; c < txfm_size_col; ++c) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +void lowbd_inv_txfm2d_add_4x8_neon(const int32_t *input, uint8_t *output, + int stride, TX_TYPE tx_type, TX_SIZE tx_size, + int eob) { + (void)eob; + DECLARE_ALIGNED(32, int, txfm_buf[4 * 8 + 8 + 8]); + int32_t *temp_in = txfm_buf; + + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_arr[txw_idx][hitx_1d_tab[tx_type]]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_arr[txh_idx][vitx_1d_tab[tx_type]]; + + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + for (int i = 0; i < txfm_size_row; i++) { + for (int j = 0; j < txfm_size_col; j++) + temp_in[j] = round_shift((int64_t)input[j] * NewInvSqrt2, NewSqrt2Bits); + + row_txfm(temp_in, buf_ptr, cos_bit_row, stage_range); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + for (int c = 0; c < txfm_size_col; ++c) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +void lowbd_inv_txfm2d_add_8x4_neon(const int32_t *input, uint8_t *output, + int stride, TX_TYPE tx_type, TX_SIZE tx_size, + int eob) { + (void)eob; + DECLARE_ALIGNED(32, int, txfm_buf[8 * 4 + 8 + 8]); + int32_t *temp_in = txfm_buf; + + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_arr[txw_idx][hitx_1d_tab[tx_type]]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_arr[txh_idx][vitx_1d_tab[tx_type]]; + + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + for (int i = 0; i < txfm_size_row; i++) { + for (int j = 0; j < txfm_size_col; j++) + temp_in[j] = round_shift((int64_t)input[j] * NewInvSqrt2, NewSqrt2Bits); + + row_txfm(temp_in, buf_ptr, cos_bit_row, stage_range); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + for (int c = 0; c < txfm_size_col; ++c) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +void lowbd_inv_txfm2d_add_4x16_neon(const int32_t *input, uint8_t *output, + int stride, TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + (void)eob; + DECLARE_ALIGNED(32, int, txfm_buf[4 * 16 + 16 + 16]); + int32_t *temp_in = txfm_buf; + + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_arr[txw_idx][hitx_1d_tab[tx_type]]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_arr[txh_idx][vitx_1d_tab[tx_type]]; + + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + for (int i = 0; i < txfm_size_row; i++) { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + av1_round_shift_array(buf_ptr, txfm_size_col, -shift[0]); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + for (int c = 0; c < txfm_size_col; ++c) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +void lowbd_inv_txfm2d_add_16x4_neon(const int32_t *input, uint8_t *output, + int stride, TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + (void)eob; + + DECLARE_ALIGNED(32, int, txfm_buf[16 * 4 + 16 + 16]); + int32_t *temp_in = txfm_buf; + + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + int r, bd = 8; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_arr[txw_idx][hitx_1d_tab[tx_type]]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_arr[txh_idx][vitx_1d_tab[tx_type]]; + + int ud_flip, lr_flip; + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + + for (int i = 0; i < txfm_size_row; i++) { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + av1_round_shift_array(buf_ptr, txfm_size_col, -shift[0]); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + for (int c = 0; c < txfm_size_col; ++c) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +static INLINE void lowbd_inv_txfm2d_add_no_identity_neon( + const int32_t *input, uint8_t *output, int stride, TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + DECLARE_ALIGNED(32, int, txfm_buf[64 * 64 + 64 + 64]); + int32_t *temp_in = txfm_buf; + + int eobx, eoby, ud_flip, lr_flip, row_start; + get_eobx_eoby_scan_default(&eobx, &eoby, tx_size, eob); + const int8_t *shift = inv_txfm_shift_ls[tx_size]; + const int txw_idx = get_txw_idx(tx_size); + const int txh_idx = get_txh_idx(tx_size); + const int cos_bit_col = inv_cos_bit_col[txw_idx][txh_idx]; + const int cos_bit_row = inv_cos_bit_row[txw_idx][txh_idx]; + const int txfm_size_col = tx_size_wide[tx_size]; + const int txfm_size_row = tx_size_high[tx_size]; + const int buf_size_nonzero_h_div8 = (eoby + 8) >> 3; + const int rect_type = get_rect_tx_log_ratio(txfm_size_col, txfm_size_row); + const int buf_offset = AOMMAX(txfm_size_row, txfm_size_col); + + int32_t *temp_out = temp_in + buf_offset; + int32_t *buf = temp_out + buf_offset; + int32_t *buf_ptr = buf; + const int8_t stage_range[MAX_TXFM_STAGE_NUM] = { 16 }; + const int bd = 8; + int r; + + const int fun_idx_x = lowbd_txfm_all_1d_zeros_idx[eobx]; + const int fun_idx_y = lowbd_txfm_all_1d_zeros_idx[eoby]; + const transform_1d_neon row_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txw_idx][hitx_1d_tab[tx_type]][fun_idx_x]; + const transform_1d_neon col_txfm = + lowbd_txfm_all_1d_zeros_w8_arr[txh_idx][vitx_1d_tab[tx_type]][fun_idx_y]; + + assert(col_txfm != NULL); + assert(row_txfm != NULL); + + get_flip_cfg(tx_type, &ud_flip, &lr_flip); + row_start = (buf_size_nonzero_h_div8 << 3); + + for (int i = 0; i < row_start; i++) { + if (abs(rect_type) == 1) { + for (int j = 0; j < txfm_size_col; j++) + temp_in[j] = round_shift((int64_t)input[j] * NewInvSqrt2, NewSqrt2Bits); + row_txfm(temp_in, buf_ptr, cos_bit_row, stage_range); + } else { + row_txfm(input, buf_ptr, cos_bit_row, stage_range); + } + av1_round_shift_array(buf_ptr, txfm_size_col, -shift[0]); + input += txfm_size_col; + buf_ptr += txfm_size_col; + } + + // Doing memset for the rows which are not processed in row transform. + memset(buf_ptr, 0, + sizeof(int32_t) * txfm_size_col * (txfm_size_row - row_start)); + + for (int c = 0; c < txfm_size_col; c++) { + if (lr_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + c]; + } else { + // flip left right + for (r = 0; r < txfm_size_row; ++r) + temp_in[r] = buf[r * txfm_size_col + (txfm_size_col - c - 1)]; + } + col_txfm(temp_in, temp_out, cos_bit_col, stage_range); + av1_round_shift_array(temp_out, txfm_size_row, -shift[1]); + + if (ud_flip == 0) { + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = + highbd_clip_pixel_add(output[r * stride + c], temp_out[r], bd); + } + } else { + // flip upside down + for (r = 0; r < txfm_size_row; ++r) { + output[r * stride + c] = highbd_clip_pixel_add( + output[r * stride + c], temp_out[txfm_size_row - r - 1], bd); + } + } + } +} + +static INLINE void lowbd_inv_txfm2d_add_universe_neon( + const int32_t *input, uint8_t *output, int stride, TX_TYPE tx_type, + TX_SIZE tx_size, int eob) { + switch (tx_type) { + case IDTX: + lowbd_inv_txfm2d_add_idtx_neon(input, output, stride, tx_type, tx_size, + eob); + break; + + case H_DCT: + case H_ADST: + case H_FLIPADST: + lowbd_inv_txfm2d_add_v_identity_neon(input, output, stride, tx_type, + tx_size, eob); + break; + + case V_DCT: + case V_ADST: + case V_FLIPADST: + lowbd_inv_txfm2d_add_h_identity_neon(input, output, stride, tx_type, + tx_size, eob); + break; + + default: + lowbd_inv_txfm2d_add_no_identity_neon(input, output, stride, tx_type, + tx_size, eob); + break; + } +} +void av1_lowbd_inv_txfm2d_add_neon(const int32_t *input, uint8_t *output, + int stride, TX_TYPE tx_type, TX_SIZE tx_size, + int eob) { + int row; + switch (tx_size) { + case TX_4X4: + lowbd_inv_txfm2d_add_4x4_neon(input, output, stride, tx_type, tx_size, + eob); + break; + + case TX_4X8: + lowbd_inv_txfm2d_add_4x8_neon(input, output, stride, tx_type, tx_size, + eob); + break; + + case TX_8X4: + lowbd_inv_txfm2d_add_8x4_neon(input, output, stride, tx_type, tx_size, + eob); + break; + + case TX_4X16: + lowbd_inv_txfm2d_add_4x16_neon(input, output, stride, tx_type, tx_size, + eob); + break; + + case TX_16X4: + lowbd_inv_txfm2d_add_16x4_neon(input, output, stride, tx_type, tx_size, + eob); + break; + + case TX_16X64: { + lowbd_inv_txfm2d_add_universe_neon(input, output, stride, tx_type, + tx_size, eob); + } break; + + case TX_64X16: { + int32_t mod_input[64 * 16]; + for (row = 0; row < 16; ++row) { + memcpy(mod_input + row * 64, input + row * 32, 32 * sizeof(*mod_input)); + memset(mod_input + row * 64 + 32, 0, 32 * sizeof(*mod_input)); + } + lowbd_inv_txfm2d_add_universe_neon(mod_input, output, stride, tx_type, + tx_size, eob); + } break; + + case TX_32X64: { + lowbd_inv_txfm2d_add_universe_neon(input, output, stride, tx_type, + tx_size, eob); + } break; + + case TX_64X32: { + int32_t mod_input[64 * 32]; + for (row = 0; row < 32; ++row) { + memcpy(mod_input + row * 64, input + row * 32, 32 * sizeof(*mod_input)); + memset(mod_input + row * 64 + 32, 0, 32 * sizeof(*mod_input)); + } + lowbd_inv_txfm2d_add_universe_neon(mod_input, output, stride, tx_type, + tx_size, eob); + } break; + + case TX_64X64: { + int32_t mod_input[64 * 64]; + for (row = 0; row < 32; ++row) { + memcpy(mod_input + row * 64, input + row * 32, 32 * sizeof(*mod_input)); + memset(mod_input + row * 64 + 32, 0, 32 * sizeof(*mod_input)); + } + lowbd_inv_txfm2d_add_universe_neon(mod_input, output, stride, tx_type, + tx_size, eob); + } break; + + default: + lowbd_inv_txfm2d_add_universe_neon(input, output, stride, tx_type, + tx_size, eob); + break; + } +} +void av1_inv_txfm_add_neon(const tran_low_t *dqcoeff, uint8_t *dst, int stride, + const TxfmParam *txfm_param) { + const TX_TYPE tx_type = txfm_param->tx_type; + if (!txfm_param->lossless) { + av1_lowbd_inv_txfm2d_add_neon(dqcoeff, dst, stride, tx_type, + txfm_param->tx_size, txfm_param->eob); + } else { + av1_inv_txfm_add_c(dqcoeff, dst, stride, txfm_param); + } +}
diff --git a/av1/common/arm/av1_inv_txfm_neon.h b/av1/common/arm/av1_inv_txfm_neon.h new file mode 100644 index 0000000..6af2d61 --- /dev/null +++ b/av1/common/arm/av1_inv_txfm_neon.h
@@ -0,0 +1,152 @@ +/* + * Copyright (c) 2018, Alliance for Open Media. All rights reserved + * + * This source code is subject to the terms of the BSD 2 Clause License and + * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License + * was not distributed with this source code in the LICENSE file, you can + * obtain it at www.aomedia.org/license/software. If the Alliance for Open + * Media Patent License 1.0 was not distributed with this source code in the + * PATENTS file, you can obtain it at www.aomedia.org/license/patent. + */ +#ifndef AV1_COMMON_ARM_AV1_INV_TXFM_NEON_H_ +#define AV1_COMMON_ARM_AV1_INV_TXFM_NEON_H_ + +#include "config/aom_config.h" +#include "config/av1_rtcd.h" + +#include "aom/aom_integer.h" +#include "av1/common/enums.h" +#include "av1/common/av1_inv_txfm1d.h" +#include "av1/common/av1_inv_txfm1d_cfg.h" +#include "av1/common/av1_txfm.h" + +typedef void (*transform_1d_neon)(const int32_t *input, int32_t *output, + const int8_t cos_bit, + const int8_t *stage_ptr); + +DECLARE_ALIGNED(16, static const int16_t, av1_eob_to_eobxy_8x8_default[8]) = { + 0x0707, 0x0707, 0x0707, 0x0707, 0x0707, 0x0707, 0x0707, 0x0707, +}; + +DECLARE_ALIGNED(16, static const int16_t, + av1_eob_to_eobxy_16x16_default[16]) = { + 0x0707, 0x0707, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, + 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, +}; + +DECLARE_ALIGNED(16, static const int16_t, + av1_eob_to_eobxy_32x32_default[32]) = { + 0x0707, 0x0f0f, 0x0f0f, 0x0f0f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, + 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, + 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, + 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, 0x1f1f, +}; + +DECLARE_ALIGNED(16, static const int16_t, av1_eob_to_eobxy_8x16_default[16]) = { + 0x0707, 0x0707, 0x0707, 0x0707, 0x0707, 0x0f07, 0x0f07, 0x0f07, + 0x0f07, 0x0f07, 0x0f07, 0x0f07, 0x0f07, 0x0f07, 0x0f07, 0x0f07, +}; + +DECLARE_ALIGNED(16, static const int16_t, av1_eob_to_eobxy_16x8_default[8]) = { + 0x0707, 0x0707, 0x070f, 0x070f, 0x070f, 0x070f, 0x070f, 0x070f, +}; + +DECLARE_ALIGNED(16, static const int16_t, + av1_eob_to_eobxy_16x32_default[32]) = { + 0x0707, 0x0707, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f0f, + 0x0f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, + 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, + 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, 0x1f0f, +}; + +DECLARE_ALIGNED(16, static const int16_t, + av1_eob_to_eobxy_32x16_default[16]) = { + 0x0707, 0x0f0f, 0x0f0f, 0x0f0f, 0x0f1f, 0x0f1f, 0x0f1f, 0x0f1f, + 0x0f1f, 0x0f1f, 0x0f1f, 0x0f1f, 0x0f1f, 0x0f1f, 0x0f1f, 0x0f1f, +}; + +DECLARE_ALIGNED(16, static const int16_t, av1_eob_to_eobxy_8x32_default[32]) = { + 0x0707, 0x0707, 0x0707, 0x0707, 0x0707, 0x0f07, 0x0f07, 0x0f07, + 0x0f07, 0x0f07, 0x0f07, 0x0f07, 0x0f07, 0x1f07, 0x1f07, 0x1f07, + 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, + 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, 0x1f07, +}; + +DECLARE_ALIGNED(16, static const int16_t, av1_eob_to_eobxy_32x8_default[8]) = { + 0x0707, 0x070f, 0x070f, 0x071f, 0x071f, 0x071f, 0x071f, 0x071f, +}; + +DECLARE_ALIGNED(16, static const int16_t *, + av1_eob_to_eobxy_default[TX_SIZES_ALL]) = { + NULL, + av1_eob_to_eobxy_8x8_default, + av1_eob_to_eobxy_16x16_default, + av1_eob_to_eobxy_32x32_default, + av1_eob_to_eobxy_32x32_default, + NULL, + NULL, + av1_eob_to_eobxy_8x16_default, + av1_eob_to_eobxy_16x8_default, + av1_eob_to_eobxy_16x32_default, + av1_eob_to_eobxy_32x16_default, + av1_eob_to_eobxy_32x32_default, + av1_eob_to_eobxy_32x32_default, + NULL, + NULL, + av1_eob_to_eobxy_8x32_default, + av1_eob_to_eobxy_32x8_default, + av1_eob_to_eobxy_16x32_default, + av1_eob_to_eobxy_32x16_default, +}; + +static const int lowbd_txfm_all_1d_zeros_idx[32] = { + 0, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, +}; + +// Transform block width in log2 for eob (size of 64 map to 32) +static const int tx_size_wide_log2_eob[TX_SIZES_ALL] = { + 2, 3, 4, 5, 5, 2, 3, 3, 4, 4, 5, 5, 5, 2, 4, 3, 5, 4, 5, +}; + +static int eob_fill[32] = { + 0, 7, 7, 7, 7, 7, 7, 7, 15, 15, 15, 15, 15, 15, 15, 15, + 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, +}; + +static INLINE void get_eobx_eoby_scan_default(int *eobx, int *eoby, + TX_SIZE tx_size, int eob) { + if (eob == 1) { + *eobx = 0; + *eoby = 0; + return; + } + + const int tx_w_log2 = tx_size_wide_log2_eob[tx_size]; + const int eob_row = (eob - 1) >> tx_w_log2; + const int eobxy = av1_eob_to_eobxy_default[tx_size][eob_row]; + *eobx = eobxy & 0xFF; + *eoby = eobxy >> 8; +} + +static INLINE void get_eobx_eoby_scan_v_identity(int *eobx, int *eoby, + TX_SIZE tx_size, int eob) { + eob -= 1; + const int txfm_size_row = tx_size_high[tx_size]; + const int eoby_max = AOMMIN(32, txfm_size_row) - 1; + *eobx = eob / (eoby_max + 1); + *eoby = (eob >= eoby_max) ? eoby_max : eob_fill[eob]; +} + +static INLINE void get_eobx_eoby_scan_h_identity(int *eobx, int *eoby, + TX_SIZE tx_size, int eob) { + eob -= 1; + const int txfm_size_col = tx_size_wide[tx_size]; + const int eobx_max = AOMMIN(32, txfm_size_col) - 1; + *eobx = (eob >= eobx_max) ? eobx_max : eob_fill[eob]; + const int temp_eoby = eob / (eobx_max + 1); + assert(temp_eoby < 32); + *eoby = eob_fill[temp_eoby]; +} + +#endif // AV1_COMMON_ARM_AV1_INV_TXFM_NEON_H_
diff --git a/av1/common/av1_rtcd_defs.pl b/av1/common/av1_rtcd_defs.pl index 7ec808d..5284b36 100755 --- a/av1/common/av1_rtcd_defs.pl +++ b/av1/common/av1_rtcd_defs.pl
@@ -106,7 +106,7 @@ #inv txfm add_proto qw/void av1_inv_txfm_add/, "const tran_low_t *dqcoeff, uint8_t *dst, int stride, const TxfmParam *txfm_param"; -specialize qw/av1_inv_txfm_add ssse3 avx2/; +specialize qw/av1_inv_txfm_add ssse3 avx2 neon/; add_proto qw/void av1_highbd_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd"; add_proto qw/void av1_highbd_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
diff --git a/test/av1_inv_txfm2d_test.cc b/test/av1_inv_txfm2d_test.cc index 461e7eb..11e231b 100644 --- a/test/av1_inv_txfm2d_test.cc +++ b/test/av1_inv_txfm2d_test.cc
@@ -364,4 +364,15 @@ ::testing::Values(av1_lowbd_inv_txfm2d_add_avx2)); #endif // HAVE_AVX2 +#if HAVE_NEON + +extern "C" void av1_lowbd_inv_txfm2d_add_neon(const int32_t *input, + uint8_t *output, int stride, + TX_TYPE tx_type, TX_SIZE tx_size, + int eob); + +INSTANTIATE_TEST_CASE_P(NEON, AV1LbdInvTxfm2d, + ::testing::Values(av1_lowbd_inv_txfm2d_add_neon)); +#endif // HAVE_NEON + } // namespace