Add in-place calculation support for CL elementwise arithmetic kernels
- Add in-place calculation support in ClArithmeticKernel, ClSaturatedArithmeticKernel and ClMulKernel
- Add in-place test cases
Resolves: COMPMID-4431
Signed-off-by: Sheri Zhang <sheri.zhang@arm.com>
Change-Id: Id484bdb76b74478a33fedb471ae0c7f799c599f6
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5885
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
diff --git a/src/core/CL/cl_kernels/pixelwise_mul_float.cl b/src/core/CL/cl_kernels/pixelwise_mul_float.cl
index 0016775..1087529 100644
--- a/src/core/CL/cl_kernels/pixelwise_mul_float.cl
+++ b/src/core/CL/cl_kernels/pixelwise_mul_float.cl
@@ -77,7 +77,9 @@
__kernel void pixelwise_mul_float(
TENSOR3D_DECLARATION(in1),
TENSOR3D_DECLARATION(in2),
+#if !defined(IN_PLACE)
TENSOR3D_DECLARATION(out),
+#endif // !defined(IN_PLACE)
const float scale)
{
// Get pixels pointer
@@ -87,7 +89,16 @@
__global uchar *in1_addr = in1_ptr + in1_offset_first_element_in_bytes + x * in1_stride_x + y * in1_stride_y + z * in1_stride_z;
__global uchar *in2_addr = in2_ptr + in2_offset_first_element_in_bytes + x * in2_stride_x + y * in2_stride_y + z * in2_stride_z;
- __global uchar *out_addr = out_ptr + out_offset_first_element_in_bytes + x * out_stride_x + y * out_stride_y + z * out_stride_z;
+ __global uchar *
+#if !defined(IN_PLACE)
+ out_addr = out_ptr + out_offset_first_element_in_bytes + x * out_stride_x + y * out_stride_y + z * out_stride_z;
+#else // !defined(IN_PLACE)
+#if defined(SRC1_IN_PLACE)
+ out_addr = in1_addr;
+#else //defined(SRC1_IN_PLACE)
+ out_addr = in2_addr;
+#endif //defined(SRC1_IN_PLACE)
+#endif // !defined(IN_PLACE)
// Load data
VEC_ACC_TYPE in1_data = CONVERT((VEC_DATA_TYPE(DATA_TYPE_IN1, VEC_SIZE_OUT))(VLOAD(VEC_SIZE_IN1)(0, (__global DATA_TYPE_IN1 *)in1_addr)), VEC_ACC_TYPE);