From 74921eee924625426429044decefe3673561b174 Mon Sep 17 00:00:00 2001 From: Michael Tyler Date: Wed, 12 Apr 2023 17:43:17 +0100 Subject: Update CPU kernel implementations and guard directives Resolves COMPMID-6023 Change-Id: I868975d14c4f98af6716726feda22405a6a4c891 Signed-off-by: Michael Tyler Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9686 Tested-by: Arm Jenkins Reviewed-by: Viet-Hoa Do Comments-Addressed: Arm Jenkins Benchmark: Arm Jenkins --- .../NEON/kernels/arm_conv/depthwise/depthwise_fp32.cpp | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) (limited to 'src/core/NEON/kernels/arm_conv/depthwise/depthwise_fp32.cpp') diff --git a/src/core/NEON/kernels/arm_conv/depthwise/depthwise_fp32.cpp b/src/core/NEON/kernels/arm_conv/depthwise/depthwise_fp32.cpp index 09ee983907..382ccd3c62 100644 --- a/src/core/NEON/kernels/arm_conv/depthwise/depthwise_fp32.cpp +++ b/src/core/NEON/kernels/arm_conv/depthwise/depthwise_fp32.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2021-2022 Arm Limited. + * Copyright (c) 2021-2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -103,6 +103,18 @@ namespace ); } + template + unsigned int fast_mode_cycle_estimate(const DepthwiseArgs &args, const Nothing &) + { + // First-pass: compute the number of output pixels which will be computed. + return arm_gemm::roundup(args.output_rows, Strategy::output_rows) * + arm_gemm::roundup(args.output_cols, Strategy::output_cols) * + arm_gemm::iceildiv( + (long unsigned) args.input_channels * args.channel_multiplier, + arm_gemm::utils::get_vector_length(Strategy::vl_type) + ) * 2 / 3; + } + #if defined(__aarch64__) unsigned int not_preferred(const DepthwiseArgs &, const Nothing &) { -- cgit v1.2.1