From 8d4cdd43a74574e0f99f83f1adb1d391c0c85abe Mon Sep 17 00:00:00 2001 From: Pablo Marquez Tello Date: Tue, 21 Nov 2023 10:10:01 +0000 Subject: BatchNorm changes to enable fp16 in armv8a multi_isa builds * Moved NCHW kernels fp16 and fp32 to their corresponding files src/cpu/kernels/fuse_batch_normalization/nchw/neon/fp16.cpp and src/cpu/kernels/fuse_batch_normalization/nchw/neon/fp32.cpp * Changes in filelist.json to include the new fp16 and fp32 files * Moved the template batch_normalization_nchw to impl.h as we need to instantiate it from fp16.cpp and fp32.cpp * Pooling layer: removed the guard __ARM_FEATURE_FP16_VECTOR_ARITHMETIC that prevented the FP16 kernel execution. * Partially resolves MLCE-1102 Change-Id: Ia8c85e9ffb76c9e387f9ae2685e5df5e52c8dc27 Signed-off-by: Pablo Marquez Tello Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10777 Reviewed-by: Viet-Hoa Do Comments-Addressed: Arm Jenkins Tested-by: Arm Jenkins Benchmark: Arm Jenkins --- Android.bp | 2 ++ 1 file changed, 2 insertions(+) (limited to 'Android.bp') diff --git a/Android.bp b/Android.bp index 31ec9b2716..23b264a3a9 100644 --- a/Android.bp +++ b/Android.bp @@ -515,6 +515,8 @@ cc_library_static { "src/cpu/kernels/fuse_batch_normalization/generic/fp16.cpp", "src/cpu/kernels/fuse_batch_normalization/generic/fp32.cpp", "src/cpu/kernels/fuse_batch_normalization/nchw/all.cpp", + "src/cpu/kernels/fuse_batch_normalization/nchw/neon/fp16.cpp", + "src/cpu/kernels/fuse_batch_normalization/nchw/neon/fp32.cpp", "src/cpu/kernels/fuse_batch_normalization/nhwc/neon/fp16.cpp", "src/cpu/kernels/fuse_batch_normalization/nhwc/neon/fp32.cpp", "src/cpu/kernels/gemm_matrix_add/generic/neon/fp16.cpp", -- cgit v1.2.1