COMPMID-2805: Add QASYMM8_SIGNED support in NEGEMMLowpOutputStage
Add support from requantizing down from S32 to Int8 with fixed point
requantization. This involves the following:
- Compute fixed point multiplication between each entry of input by
result_fixedpoint_multiplier
- Add bias to final result if bias tensor is not a nullptr
- Round to nearest division by a power-of-two using result_shift
- Add offset to each result
- Clamp the value between the specified min and max bounds
- Cast to int8 data type
Change-Id: I641b3fac0833c568d8565ccb859bbc561a24c17d
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-on: https://review.mlplatform.org/c/2340
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
diff --git a/tests/validate_examples/cl_gemm.cpp b/tests/validate_examples/cl_gemm.cpp
index 128c5f6..39fe111 100644
--- a/tests/validate_examples/cl_gemm.cpp
+++ b/tests/validate_examples/cl_gemm.cpp
@@ -321,11 +321,11 @@
SimpleTensor<int32_t> biases{ TensorShape(N), DataType::S32, 1 };
// Fill bias
fill(biases, 3);
- ref_dst = reference::gemmlowp_quantize_down_int32_to_uint8_scale_by_fixedpoint<int32_t>(ref_tmp_dst, biases, dst_multiplier_vec, dst_shift_vec, offset_dst);
+ ref_dst = reference::gemmlowp_quantize_down_scale_by_fixedpoint<int32_t, uint8_t>(ref_tmp_dst, biases, dst_multiplier_vec, dst_shift_vec, offset_dst);
}
else
{
- ref_dst = reference::gemmlowp_quantize_down_int32_to_uint8_scale_by_fixedpoint<int32_t>(ref_tmp_dst, dst_multiplier_vec, dst_shift_vec, offset_dst);
+ ref_dst = reference::gemmlowp_quantize_down_scale_by_fixedpoint<int32_t, uint8_t>(ref_tmp_dst, dst_multiplier_vec, dst_shift_vec, offset_dst);
}
validate(CLAccessor(dst), ref_dst);
break;