COMPMID-1226 Extend CLMeanStdDev to support FP32 / FP16

- Extend support for FP16 in CLReduction.
- For F16/F32 MeanStdDev we perform one reduction operation for mean
and one for stddev and we calculate the final result in the host CPU.

Change-Id: Iad2099f26c0ba7969737d22f00c6c275634d875c
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/135870
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>

1 files changed, 3 insertions, 2 deletions

diff --git a/tests/validation/reference/ReductionOperation.cpp b/tests/validation/reference/ReductionOperation.cpp
index acfcc09cea..871a761b1a 100644
--- a/tests/validation/reference/ReductionOperation.cpp
+++ b/tests/validation/reference/ReductionOperation.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017 ARM Limited.
+ * Copyright (c) 2017-2018 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -53,7 +53,7 @@ T reduce_operation(T *ptr, int reduce_elements, ReductionOperation op)
     switch(op)
     {
         case ReductionOperation::SUM_SQUARE:
-            return std::accumulate(ptr, ptr + reduce_elements, 0.f, square<T>());
+            return std::accumulate(ptr, ptr + reduce_elements, static_cast<T>(0), square<T>());
         default:
             ARM_COMPUTE_ERROR("Unsupported reduction operation");
     }
@@ -87,6 +87,7 @@ SimpleTensor<T> reduction_operation(const SimpleTensor<T> &src, const TensorShap
 }
 
 template SimpleTensor<float> reduction_operation(const SimpleTensor<float> &src, const TensorShape &dst_shape, unsigned int axis, ReductionOperation op);
+template SimpleTensor<half> reduction_operation(const SimpleTensor<half> &src, const TensorShape &dst_shape, unsigned int axis, ReductionOperation op);
 } // namespace reference
 } // namespace validation
 } // namespace test