COMPMID-3687: Remove deprecated functions in 20.05 release

Change-Id: I90e09e460b5d5d4f9ead8e3905833c5da3b9fbd6
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3762
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Sheri Zhang <sheri.zhang@arm.com>

6 files changed, 7 insertions, 288 deletions

diff --git a/arm_compute/runtime/CL/functions/CLGEMMLowpOutputStage.h b/arm_compute/runtime/CL/functions/CLGEMMLowpOutputStage.h
index 8f7a7c397e..c6e95888e5 100644
--- a/arm_compute/runtime/CL/functions/CLGEMMLowpOutputStage.h
+++ b/arm_compute/runtime/CL/functions/CLGEMMLowpOutputStage.h
@@ -38,76 +38,6 @@ namespace arm_compute
 {
 class ITensor;
 
-/** Basic function to execute CLGEMMLowpQuantizeDownInt32ToUint8Scale on OpenCL.
- *
- *  CLGEMMLowpQuantizeDownInt32ToUint8Scale depends on 3 parameters: result_offset, result_mult_int, result_shift
- *  The final result is:
- *
- *  ((input[i][k] + result_offset) * result_mult_int) >> result_shift
- *
- * In case the bias tensor is provided, the final result is:
- *
- *  ((input[i][k] + bias[k] + result_offset) * result_mult_int) >> result_shift
- *
- *  This function calls the following OpenCL kernels:
- *
- * -# @ref CLGEMMLowpQuantizeDownInt32ScaleKernel
- *
- * @note The function accepts also 2 optional input arguments (min and max) which can be used to implement "rectified linear unit" activation functions
- *       after the result is shifted right by result_shift
-*/
-class CLGEMMLowpQuantizeDownInt32ToUint8Scale : public ICLSimpleFunction
-{
-public:
-    /** Initialise the kernel's inputs, output
-     *
-     * @param[in]  input           Input tensor. It is the output of @ref CLGEMMLowpMatrixMultiplyCore function. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: QASYMM8
-     * @param[in]  result_offset   Offset to be added to each element of the input matrix
-     * @param[in]  result_mult_int Value to be multiplied to each element of the input matrix when once the result_offset has been add
-     * @param[in]  result_shift    Number of bits to shift right the result before converting back to QASYMM8
-     * @param[in]  min             (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in]  max             (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                             Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, int result_offset, int result_mult_int, int result_shift, int min = std::numeric_limits<int32_t>::lowest(),
-                   int max = std::numeric_limits<int32_t>::max());
-    /** Initialise the kernel's inputs, output
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. It is the output of @ref CLGEMMLowpMatrixMultiplyCore function. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: QASYMM8
-     * @param[in]  result_offset   Offset to be added to each element of the input matrix
-     * @param[in]  result_mult_int Value to be multiplied to each element of the input matrix when once the result_offset has been add
-     * @param[in]  result_shift    Number of bits to shift right the result before converting back to QASYMM8
-     * @param[in]  min             (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in]  max             (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                             Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, int result_offset, int result_mult_int, int result_shift,
-                   int min = std::numeric_limits<int32_t>::lowest(), int max = std::numeric_limits<int32_t>::max());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpQuantizeDownInt32ToUint8Scale
-     *
-     * @param[in] input  Input tensor. It is the output of @ref CLGEMMLowpMatrixMultiplyCore function. Data type supported: S32
-     * @param[in] bias   Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                   Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output Output tensor. Data type supported: QASYMM8
-     * @param[in] min    (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in] max    (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                   Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     *
-     * @return a status
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits<int32_t>::lowest(), int max = std::numeric_limits<int32_t>::max());
-};
-
 /** Basic function to execute CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint on OpenCL.
  *
  *  CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint depends on 3 parameters:
@@ -264,65 +194,6 @@ public:
     static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits<int32_t>::lowest(), int max = std::numeric_limits<int32_t>::max());
 };
 
-/** Basic function to execute CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat on OpenCL.
- *
- *  This function calls the following OpenCL kernels:
- *
- * -# @ref CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel
- *
- * @note The function accepts also 2 optional input arguments (min and max) which can be used to implement "rectified linear unit" activation functions
- *       after the result is shifted right by result_shift
-*/
-class CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat : public ICLSimpleFunction
-{
-public:
-    /** Initialise the kernel's inputs, output
-     *
-     * @param[in]  input      Input tensor. Data type supported: S32
-     * @param[in]  bias       Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                        Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output     Output tensor. Data type supported: QASYMM8
-     * @param[in]  multiplier Float multiplier to be multiplied to each element of the input matrix
-     * @param[in]  offset     Offset to be applied to result before converting it back to QASYMM8
-     * @param[in]  min        (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in]  max        (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                        Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, float multiplier, int offset, int min = std::numeric_limits<int32_t>::lowest(),
-                   int max = std::numeric_limits<int32_t>::max());
-    /** Initialise the kernel's inputs, output
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: QASYMM8
-     * @param[in]  multiplier      Float multiplier to be multiplied to each element of the input matrix
-     * @param[in]  offset          Offset to be applied to result before converting it back to QASYMM8
-     * @param[in]  min             (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in]  max             (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                        Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, float multiplier, int offset,
-                   int min = std::numeric_limits<int32_t>::lowest(),
-                   int max = std::numeric_limits<int32_t>::max());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
-     *
-     * @param[in] input  Input tensor. It is the output of @ref CLGEMMLowpMatrixMultiplyCore function. Data type supported: S32
-     * @param[in] bias   Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                   Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output Output tensor. Data type supported: QASYMM8
-     * @param[in] min    (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in] max    (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                   Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     *
-     * @return a status
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits<int32_t>::lowest(), int max = std::numeric_limits<int32_t>::max());
-};
 /** Basic function to execute CLGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint on OpenCL.
  *
  *  CLGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint depends on 2 parameters:
@@ -442,4 +313,4 @@ public:
     static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo &info);
 };
 } // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGEMMLOWPOUTPUTSTAGE_H */
\ No newline at end of file
+#endif /*ARM_COMPUTE_CLGEMMLOWPOUTPUTSTAGE_H */
diff --git a/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h b/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h
index 486b9c6791..f29d5d464b 100644
--- a/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h
+++ b/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h
@@ -38,59 +38,6 @@ namespace arm_compute
 {
 class ITensor;
 
-/** Basic function to execute NEGEMMLowpQuantizeDownInt32ToUint8Scale on NEON.
- *
- *  NEGEMMLowpQuantizeDownInt32ToUint8Scale depends on 3 parameters: result_offset, result_mult_int, result_shift
- *  The final result is:
- *
- *  ((input[i][k] + result_offset) * result_mult_int) >> result_shift
- *
- * In case the bias tensor is provided, the final result is:
- *
- *  ((input[i][k] + bias[k] + result_offset) * result_mult_int) >> result_shift
- *
- *  This function calls the following NEON kernels:
- *
- * -# @ref NEGEMMLowpQuantizeDownInt32ScaleKernel
- *
- * @note The function accepts also 2 optional input arguments (min and max) which can be used to implement "rectified linear unit" activation functions
- *       after the result is shifted right by result_shift
-*/
-class NEGEMMLowpQuantizeDownInt32ToUint8Scale : public INESimpleFunctionNoBorder
-{
-public:
-    /** Initialise the kernel's inputs, output
-     *
-     * @param[in]  input           Input tensor. It is the output of @ref NEGEMMLowpMatrixMultiplyCore function. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: Data type supported: QASYMM8
-     * @param[in]  result_offset   Offset to be added to each element of the input matrix
-     * @param[in]  result_mult_int Value to be multiplied to each element of the input matrix when once the result_offset has been add
-     * @param[in]  result_shift    Number of bits to shift right the result before converting back to QASYMM8
-     * @param[in]  min             (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in]  max             (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                             Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_offset, int result_mult_int, int result_shift, int min = std::numeric_limits<int32_t>::lowest(),
-                   int max = std::numeric_limits<int32_t>::max());
-    /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToUint8Scale
-     *
-     * @param[in] input  Input tensor. It is the output of @ref NEGEMMLowpMatrixMultiplyCore function. Data type supported: S32
-     * @param[in] bias   Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                   Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8
-     * @param[in] min    (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer.
-     * @param[in] max    (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
-     *                   Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer.
-     *
-     * @return a status
-     */
-    ARM_COMPUTE_DEPRECATED_REL(20.05)
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits<int32_t>::lowest(), int max = std::numeric_limits<int32_t>::max());
-};
-
 /** Basic function to execute NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint on NEON.
  *
  *  NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint depends on 3 parameters:
diff --git a/docs/00_introduction.dox b/docs/00_introduction.dox
index c281cb21b7..35fff366c9 100644
--- a/docs/00_introduction.dox
+++ b/docs/00_introduction.dox
@@ -257,6 +257,11 @@ v20.08 Public major release
    - graph_yolov3_output_detector
  - Removed padding from:
    - @ref NEPixelWiseMultiplicationKernel
+ - Removedd OpenCL kernels / functions:
+     - CLGEMMLowpQuantizeDownInt32ToUint8Scale
+     - CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat
+ - Removed NEON kernels / functions:
+     - NEGEMMLowpQuantizeDownInt32ToUint8Scale
  - GEMMTuner improvements:
      - Added fp16 support
      - Output json files for easier integration
diff --git a/docs/06_functions_list.dox b/docs/06_functions_list.dox
index 4a7c9fac6a..99eab12f8f 100644
--- a/docs/06_functions_list.dox
+++ b/docs/06_functions_list.dox
@@ -91,7 +91,6 @@ namespace arm_compute
         - @ref NEFullyConnectedLayerReshapeWeights
         - @ref NEGather
         - @ref NEGEMMInterleave4x4
-        - @ref NEGEMMLowpQuantizeDownInt32ToUint8Scale
         - @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
         - @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
         - @ref NEGEMMTranspose1xW
@@ -297,10 +296,8 @@ namespace arm_compute
         - @ref CLFullyConnectedLayerReshapeWeights
         - @ref CLGather
         - @ref CLGaussian3x3
-        - @ref CLGEMMLowpQuantizeDownInt32ToUint8Scale
         - @ref CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
         - @ref CLGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint
-        - @ref CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat
         - @ref CLMagnitude
         - @ref CLMeanStdDevNormalizationLayer
         - @ref CLMedian3x3
diff --git a/src/runtime/CL/functions/CLGEMMLowpOutputStage.cpp b/src/runtime/CL/functions/CLGEMMLowpOutputStage.cpp
index d27b764a3d..a499e1858d 100644
--- a/src/runtime/CL/functions/CLGEMMLowpOutputStage.cpp
+++ b/src/runtime/CL/functions/CLGEMMLowpOutputStage.cpp
@@ -33,45 +33,6 @@
 
 namespace arm_compute
 {
-void CLGEMMLowpQuantizeDownInt32ToUint8Scale::configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, int result_offset, int result_mult_int, int result_shift, int min, int max)
-{
-    GEMMLowpOutputStageInfo info = GEMMLowpOutputStageInfo();
-    info.gemmlowp_offset         = result_offset;
-    info.gemmlowp_multiplier     = result_mult_int;
-    info.gemmlowp_shift          = result_shift;
-    info.gemmlowp_min_bound      = min;
-    info.gemmlowp_max_bound      = max;
-
-    auto k = arm_compute::support::cpp14::make_unique<CLGEMMLowpQuantizeDownInt32ScaleKernel>();
-    k->configure(CLKernelLibrary::get().get_compile_context(), input, bias, output, &info);
-    _kernel = std::move(k);
-}
-
-void CLGEMMLowpQuantizeDownInt32ToUint8Scale::configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, int result_offset,
-                                                        int result_mult_int,
-                                                        int result_shift, int min, int max)
-{
-    GEMMLowpOutputStageInfo info = GEMMLowpOutputStageInfo();
-    info.gemmlowp_offset         = result_offset;
-    info.gemmlowp_multiplier     = result_mult_int;
-    info.gemmlowp_shift          = result_shift;
-    info.gemmlowp_min_bound      = min;
-    info.gemmlowp_max_bound      = max;
-
-    auto k = arm_compute::support::cpp14::make_unique<CLGEMMLowpQuantizeDownInt32ScaleKernel>();
-    k->configure(compile_context, input, bias, output, &info);
-    _kernel = std::move(k);
-}
-
-Status CLGEMMLowpQuantizeDownInt32ToUint8Scale::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max)
-{
-    GEMMLowpOutputStageInfo info = GEMMLowpOutputStageInfo();
-    info.gemmlowp_min_bound      = min;
-    info.gemmlowp_max_bound      = max;
-
-    return CLGEMMLowpQuantizeDownInt32ScaleKernel::validate(input, bias, output, &info);
-}
-
 void CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output,
                                                                     int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift,
                                                                     int min, int max)
@@ -118,45 +79,6 @@ Status CLGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint::validate(const ITenso
     return CLGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::validate(input, bias, output, min, max);
 }
 
-void CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat::configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output,
-                                                               float multiplier, int offset,
-                                                               int min, int max)
-{
-    GEMMLowpOutputStageInfo info  = GEMMLowpOutputStageInfo();
-    info.gemmlowp_offset          = offset;
-    info.gemmlowp_real_multiplier = multiplier;
-    info.gemmlowp_min_bound       = min;
-    info.gemmlowp_max_bound       = max;
-
-    auto k = arm_compute::support::cpp14::make_unique<CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel>();
-    k->configure(CLKernelLibrary::get().get_compile_context(), input, bias, output, &info);
-    _kernel = std::move(k);
-}
-
-void CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat::configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output,
-                                                               float multiplier, int offset,
-                                                               int min, int max)
-{
-    GEMMLowpOutputStageInfo info  = GEMMLowpOutputStageInfo();
-    info.gemmlowp_offset          = offset;
-    info.gemmlowp_real_multiplier = multiplier;
-    info.gemmlowp_min_bound       = min;
-    info.gemmlowp_max_bound       = max;
-
-    auto k = arm_compute::support::cpp14::make_unique<CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel>();
-    k->configure(compile_context, input, bias, output, &info);
-    _kernel = std::move(k);
-}
-
-Status CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloat::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output,
-                                                                int min, int max)
-{
-    GEMMLowpOutputStageInfo info = GEMMLowpOutputStageInfo();
-    info.gemmlowp_min_bound      = min;
-    info.gemmlowp_max_bound      = max;
-    return CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel::validate(input, bias, output, &info);
-}
-
 void CLGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint::configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output,
                                                                     int result_fixedpoint_multiplier, int result_shift,
                                                                     int min, int max)
@@ -268,4 +190,4 @@ Status CLGEMMLowpOutputStage::validate(const ITensorInfo *input, const ITensorIn
             return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported GEMMLowpOutputStage type.");
     }
 }
-} // namespace arm_compute
\ No newline at end of file
+} // namespace arm_compute
diff --git a/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp b/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp
index ed9b449629..239a8e668a 100644
--- a/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp
+++ b/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp
@@ -33,29 +33,6 @@
 
 namespace arm_compute
 {
-void NEGEMMLowpQuantizeDownInt32ToUint8Scale::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_offset, int result_mult_int, int result_shift, int min, int max)
-{
-    GEMMLowpOutputStageInfo info = GEMMLowpOutputStageInfo();
-    info.gemmlowp_offset         = result_offset;
-    info.gemmlowp_multiplier     = result_mult_int;
-    info.gemmlowp_shift          = result_shift;
-    info.gemmlowp_min_bound      = min;
-    info.gemmlowp_max_bound      = max;
-
-    auto k = arm_compute::support::cpp14::make_unique<NEGEMMLowpQuantizeDownInt32ScaleKernel>();
-    k->configure(input, bias, output, &info);
-    _kernel = std::move(k);
-}
-
-Status NEGEMMLowpQuantizeDownInt32ToUint8Scale::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max)
-{
-    GEMMLowpOutputStageInfo info = GEMMLowpOutputStageInfo();
-    info.gemmlowp_min_bound      = min;
-    info.gemmlowp_max_bound      = max;
-
-    return NEGEMMLowpQuantizeDownInt32ScaleKernel::validate(input, bias, output, &info);
-}
-
 void NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift,
                                                                     int result_offset_after_shift, int min, int max)
 {