From 0a5a57a3f794de851408bae1c63b1660b4c5cbe7 Mon Sep 17 00:00:00 2001 From: Usama Arif Date: Thu, 23 May 2019 14:20:33 +0100 Subject: COMPMID-2160: Implement Round for NEON Change-Id: Ie80e2ad294eaf95bc823d979842c320e8fb41f67 Signed-off-by: Usama Arif Reviewed-on: https://review.mlplatform.org/c/1215 Comments-Addressed: Arm Jenkins Reviewed-by: Georgios Pinitas Tested-by: Arm Jenkins --- src/core/NEON/kernels/NEReductionOperationKernel.cpp | 2 ++ 1 file changed, 2 insertions(+) (limited to 'src/core/NEON/kernels/NEReductionOperationKernel.cpp') diff --git a/src/core/NEON/kernels/NEReductionOperationKernel.cpp b/src/core/NEON/kernels/NEReductionOperationKernel.cpp index e6edf22083..c6e853659c 100644 --- a/src/core/NEON/kernels/NEReductionOperationKernel.cpp +++ b/src/core/NEON/kernels/NEReductionOperationKernel.cpp @@ -180,6 +180,7 @@ inline uint8x8_t calculate_max(uint8x16_t in) return wrapper::vpmax(pmax, pmax); } +template <> uint32_t calculate_vector_index(uint32x4x4_t vec_res_idx, uint8x16_t vec_res_value, ReductionOperation op) { uint32x4x4_t res_idx_mask{ { 0 } }; @@ -272,6 +273,7 @@ inline float16x4_t calculate_max(float16x8_t in) return wrapper::vpmax(pmax, pmax); } +template <> uint32_t calculate_vector_index(uint32x4x4_t vec_res_idx, float16x8_t vec_res_value, ReductionOperation op) { uint32x4x2_t res_idx_mask{ 0 }; -- cgit v1.2.1