Optimize predict_max_partition Removed unneeded softmax calculation. Change-Id: I2e5528f606961f4c222ac7f55ce7dc217d64a219
diff --git a/av1/encoder/partition_strategy.c b/av1/encoder/partition_strategy.c index c05c506..59e4a875 100644 --- a/av1/encoder/partition_strategy.c +++ b/av1/encoder/partition_strategy.c
@@ -638,6 +638,7 @@ MACROBLOCKD *xd = &x->e_mbd; const BLOCK_SIZE sb_size = cm->seq_params.sb_size; + // Currently this only allows 128X128 SB size. May extend it to 64X64 SB size. assert(sb_size == BLOCK_128X128); int f_idx = 0; @@ -701,14 +702,18 @@ if (log_sse > max_log_sse) max_log_sse = log_sse; } aom_clear_system_state(); - const float avg_mv_row = sum_mv_row / 64.0f; - const float var_mv_row = sum_mv_row_sq / 64.0f - avg_mv_row * avg_mv_row; + const int blks = mb_rows * mb_cols; + const float avg_mv_row = sum_mv_row / (float)blks; + const float var_mv_row = + sum_mv_row_sq / (float)blks - avg_mv_row * avg_mv_row; - const float avg_mv_col = sum_mv_col / 64.0f; - const float var_mv_col = sum_mv_col_sq / 64.0f - avg_mv_col * avg_mv_col; + const float avg_mv_col = sum_mv_col / (float)blks; + const float var_mv_col = + sum_mv_col_sq / (float)blks - avg_mv_col * avg_mv_col; - const float avg_log_sse = sum_log_sse / 64.0f; - const float var_log_sse = sum_log_sse_sq / 64.0f - avg_log_sse * avg_log_sse; + const float avg_log_sse = sum_log_sse / (float)blks; + const float var_log_sse = + sum_log_sse_sq / (float)blks - avg_log_sse * avg_log_sse; features[f_idx++] = avg_log_sse; features[f_idx++] = avg_mv_col; @@ -730,8 +735,7 @@ BLOCK_SIZE av1_predict_max_partition(const AV1_COMP *const cpi, const MACROBLOCK *const x, const float *features) { - float scores[MAX_NUM_CLASSES_MAX_MIN_PART_PRED] = { 0.0f }, - probs[MAX_NUM_CLASSES_MAX_MIN_PART_PRED] = { 0.0f }; + float scores[MAX_NUM_CLASSES_MAX_MIN_PART_PRED] = { 0.0f }; const NN_CONFIG *nn_config = &av1_max_part_pred_nn_config; assert(cpi->sf.part_sf.auto_max_partition_based_on_simple_motion != @@ -739,21 +743,26 @@ aom_clear_system_state(); av1_nn_predict(features, nn_config, 1, scores); - av1_nn_softmax(scores, probs, MAX_NUM_CLASSES_MAX_MIN_PART_PRED); int result = MAX_NUM_CLASSES_MAX_MIN_PART_PRED - 1; if (cpi->sf.part_sf.auto_max_partition_based_on_simple_motion == DIRECT_PRED) { result = 0; - float max_prob = probs[0]; + float max_score = scores[0]; for (int i = 1; i < MAX_NUM_CLASSES_MAX_MIN_PART_PRED; ++i) { - if (probs[i] > max_prob) { - max_prob = probs[i]; + if (scores[i] > max_score) { + max_score = scores[i]; result = i; } } - } else if (cpi->sf.part_sf.auto_max_partition_based_on_simple_motion == - RELAXED_PRED) { + return (BLOCK_SIZE)((result + 2) * 3); + } + + float probs[MAX_NUM_CLASSES_MAX_MIN_PART_PRED] = { 0.0f }; + av1_nn_softmax(scores, probs, MAX_NUM_CLASSES_MAX_MIN_PART_PRED); + + if (cpi->sf.part_sf.auto_max_partition_based_on_simple_motion == + RELAXED_PRED) { for (result = MAX_NUM_CLASSES_MAX_MIN_PART_PRED - 1; result >= 0; --result) { if (result < MAX_NUM_CLASSES_MAX_MIN_PART_PRED - 1) {