aboutsummaryrefslogtreecommitdiff
path: root/src/gpu/cl/kernels/ClSoftmaxKernel.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'src/gpu/cl/kernels/ClSoftmaxKernel.cpp')
-rw-r--r--src/gpu/cl/kernels/ClSoftmaxKernel.cpp114
1 files changed, 73 insertions, 41 deletions
diff --git a/src/gpu/cl/kernels/ClSoftmaxKernel.cpp b/src/gpu/cl/kernels/ClSoftmaxKernel.cpp
index 59299fa441..1b5a2666bc 100644
--- a/src/gpu/cl/kernels/ClSoftmaxKernel.cpp
+++ b/src/gpu/cl/kernels/ClSoftmaxKernel.cpp
@@ -22,12 +22,14 @@
* SOFTWARE.
*/
#include "src/gpu/cl/kernels/ClSoftmaxKernel.h"
+
#include "arm_compute/core/CL/ICLTensor.h"
-#include "arm_compute/core/Utils.h"
#include "arm_compute/core/experimental/Types.h"
+#include "arm_compute/core/Utils.h"
#include "arm_compute/core/utils/helpers/AdjustVecSize.h"
#include "arm_compute/core/utils/quantization/AsymmHelpers.h"
#include "arm_compute/core/utils/StringUtils.h"
+
#include "src/core/CL/CLValidate.h"
#include "src/core/helpers/AutoConfiguration.h"
#include "src/core/helpers/WindowHelpers.h"
@@ -60,15 +62,16 @@ CLBuildOptions prepare_quantized_softmax_build_options(float input_scale, float
// Number of integer bits used in temporary fixed-point representation of exponent accumulator
static const int exp_accumulation_in_bits = 12;
- const double beta_multiplier = std::min(
- 1.0 * beta * input_scale * (1 << (31 - scaled_diff_int_bits)),
- (1LL << 31) - 1.0);
+ const double beta_multiplier =
+ std::min(1.0 * beta * input_scale * (1 << (31 - scaled_diff_int_bits)), (1LL << 31) - 1.0);
int input_beta_multiplier;
int input_beta_left_shift;
- quantization::calculate_quantized_multiplier_greater_than_one(beta_multiplier, &input_beta_multiplier, &input_beta_left_shift);
+ quantization::calculate_quantized_multiplier_greater_than_one(beta_multiplier, &input_beta_multiplier,
+ &input_beta_left_shift);
- const double max_input_rescaled = 1.0 * ((1 << scaled_diff_int_bits) - 1) * (1LL << (31 - scaled_diff_int_bits)) / (1LL << input_beta_left_shift);
- const int diff_min = -1.f * std::floor(max_input_rescaled);
+ const double max_input_rescaled =
+ 1.0 * ((1 << scaled_diff_int_bits) - 1) * (1LL << (31 - scaled_diff_int_bits)) / (1LL << input_beta_left_shift);
+ const int diff_min = -1.f * std::floor(max_input_rescaled);
CLBuildOptions build_opts;
build_opts.add_option("-DSCALED_DIFF_INT_BITS=" + support::cpp11::to_string(scaled_diff_int_bits));
@@ -80,18 +83,22 @@ CLBuildOptions prepare_quantized_softmax_build_options(float input_scale, float
return build_opts;
}
-Status validate_arguments_1DMaxShiftExpSum(const ITensorInfo &src, const ITensorInfo &max, const ITensorInfo &dst, const ITensorInfo &sum)
+Status validate_arguments_1DMaxShiftExpSum(const ITensorInfo &src,
+ const ITensorInfo &max,
+ const ITensorInfo &dst,
+ const ITensorInfo &sum)
{
ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(&src);
- ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&src, 1, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::F16, DataType::F32);
+ ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&src, 1, DataType::QASYMM8, DataType::QASYMM8_SIGNED,
+ DataType::F16, DataType::F32);
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(&src, &max);
const bool is_quantized_asymmetric = is_data_type_quantized_asymmetric(src.data_type());
// Checks performed when output is configured
- if(dst.total_size() != 0)
+ if (dst.total_size() != 0)
{
- if(is_quantized_asymmetric)
+ if (is_quantized_asymmetric)
{
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&dst, 1, DataType::S32);
}
@@ -103,9 +110,9 @@ Status validate_arguments_1DMaxShiftExpSum(const ITensorInfo &src, const ITensor
}
// Checks performed when sum is configured
- if(sum.total_size() != 0)
+ if (sum.total_size() != 0)
{
- if(is_quantized_asymmetric)
+ if (is_quantized_asymmetric)
{
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&sum, 1, DataType::S32);
}
@@ -119,7 +126,10 @@ Status validate_arguments_1DMaxShiftExpSum(const ITensorInfo &src, const ITensor
return Status{};
}
-Status validate_arguments_1DNorm(const ITensorInfo &src, const ITensorInfo &sum, const ITensorInfo &dst, const SoftmaxKernelInfo &info)
+Status validate_arguments_1DNorm(const ITensorInfo &src,
+ const ITensorInfo &sum,
+ const ITensorInfo &dst,
+ const SoftmaxKernelInfo &info)
{
ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(&src);
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&src, 1, DataType::S32, DataType::F16, DataType::F32);
@@ -127,14 +137,15 @@ Status validate_arguments_1DNorm(const ITensorInfo &src, const ITensorInfo &sum,
ARM_COMPUTE_RETURN_ERROR_ON(info.is_log && !is_data_type_float(info.input_data_type));
// Note: output should always have a scale of 1/256 and offset 0
- const QuantizationInfo allowed_quantization_info = get_softmax_output_quantization_info(info.input_data_type, info.is_log);
- const bool is_quantized_asymmetric = is_data_type_quantized_asymmetric(info.input_data_type);
+ const QuantizationInfo allowed_quantization_info =
+ get_softmax_output_quantization_info(info.input_data_type, info.is_log);
+ const bool is_quantized_asymmetric = is_data_type_quantized_asymmetric(info.input_data_type);
// Checks performed when output is configured
- if(dst.total_size() != 0)
+ if (dst.total_size() != 0)
{
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(&src, &dst);
- if(!is_quantized_asymmetric)
+ if (!is_quantized_asymmetric)
{
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(&src, &dst);
}
@@ -161,9 +172,14 @@ ClLogits1DMaxShiftExpSumKernel::ClLogits1DMaxShiftExpSumKernel()
_type = CLKernelType::ELEMENTWISE;
}
-void ClLogits1DMaxShiftExpSumKernel::configure(const CLCompileContext &compile_context, const ITensorInfo &src, ITensorInfo &max, ITensorInfo &dst, ITensorInfo &sum, const SoftmaxKernelInfo &info)
+void ClLogits1DMaxShiftExpSumKernel::configure(const CLCompileContext &compile_context,
+ const ITensorInfo &src,
+ ITensorInfo &max,
+ ITensorInfo &dst,
+ ITensorInfo &sum,
+ const SoftmaxKernelInfo &info)
{
- auto padding_info = get_padding_info({ &src, &max, &dst, &sum });
+ auto padding_info = get_padding_info({&src, &max, &dst, &sum});
// Output auto initialization if not yet initialized
auto_init_if_empty(sum, src.clone()->set_tensor_shape(max.tensor_shape()));
@@ -191,15 +207,21 @@ void ClLogits1DMaxShiftExpSumKernel::configure(const CLCompileContext &compile_c
build_opts.add_option("-DLOG_VECTOR_SIZE=" + support::cpp11::to_string(lround(log2(vector_size))));
build_opts.add_option_if((reduction_dim_size % vector_size) != 0, "-DNON_MULTIPLE_OF_VECTOR_SIZE");
build_opts.add_option_if(is_signed_qasymm8, "-DQASYMM8_SIGNED");
- build_opts.add_option_if(is_data_type_float(dt) && (beta != 1.0f), "-DBETA=" + float_to_string_with_full_precision(beta));
+ build_opts.add_option_if(is_data_type_float(dt) && (beta != 1.0f),
+ "-DBETA=" + float_to_string_with_full_precision(beta));
build_opts.add_option_if(is_data_type_float(dt) && info.is_log, "-DLOG_SOFTMAX");
- build_opts.add_option_if(is_data_type_float(dt), "-DMINVAL=" + ((dt == DataType::F16) ? std::string("-HALF_MAX") : std::string("-FLT_MAX")));
- build_opts.add_option_if(is_data_type_quantized_asymmetric(dt), "-DSCALE=" + float_to_string_with_full_precision(qinfo.scale));
- build_opts.add_option_if(is_data_type_quantized_asymmetric(dt), "-DBETA=" + float_to_string_with_full_precision(beta));
- build_opts.add_options_if(is_data_type_quantized_asymmetric(dt), prepare_quantized_softmax_build_options(qinfo.scale, beta).options());
+ build_opts.add_option_if(is_data_type_float(dt), "-DMINVAL=" + ((dt == DataType::F16) ? std::string("-HALF_MAX")
+ : std::string("-FLT_MAX")));
+ build_opts.add_option_if(is_data_type_quantized_asymmetric(dt),
+ "-DSCALE=" + float_to_string_with_full_precision(qinfo.scale));
+ build_opts.add_option_if(is_data_type_quantized_asymmetric(dt),
+ "-DBETA=" + float_to_string_with_full_precision(beta));
+ build_opts.add_options_if(is_data_type_quantized_asymmetric(dt),
+ prepare_quantized_softmax_build_options(qinfo.scale, beta).options());
cl::NDRange lws_hint(cl::NullRange);
- std::string kernel_name = std::string("softmax_layer_max_shift_exp_sum_") + (is_data_type_quantized_asymmetric(dt) ? "quantized_" : "") + "serial";
+ std::string kernel_name = std::string("softmax_layer_max_shift_exp_sum_") +
+ (is_data_type_quantized_asymmetric(dt) ? "quantized_" : "") + "serial";
// Create kernel.
_kernel = create_kernel(compile_context, kernel_name, build_opts.options());
@@ -211,7 +233,10 @@ void ClLogits1DMaxShiftExpSumKernel::configure(const CLCompileContext &compile_c
ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
}
-Status ClLogits1DMaxShiftExpSumKernel::validate(const ITensorInfo &src, const ITensorInfo &max, const ITensorInfo &dst, const ITensorInfo &sum)
+Status ClLogits1DMaxShiftExpSumKernel::validate(const ITensorInfo &src,
+ const ITensorInfo &max,
+ const ITensorInfo &dst,
+ const ITensorInfo &sum)
{
ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments_1DMaxShiftExpSum(src, max, dst, sum));
return Status{};
@@ -241,7 +266,7 @@ void ClLogits1DMaxShiftExpSumKernel::run_op(ITensorPack &tensors, const Window &
// Reconfigure window in case of parallel reduction
ParallelReductionInfo parallel_reduction_info = is_parallel_reduction(src->info()->dimension(0));
- if(std::get<0>(parallel_reduction_info))
+ if (std::get<0>(parallel_reduction_info))
{
// Launch grid_size parallel work items
window_collapsed.set(Window::DimX, Window::Dimension(0, _grid_size, 1));
@@ -258,8 +283,7 @@ void ClLogits1DMaxShiftExpSumKernel::run_op(ITensorPack &tensors, const Window &
add_3D_tensor_argument(idx, dst, slice);
add_3D_tensor_argument(idx, sum, slice);
enqueue(queue, *this, slice, lws_hint());
- }
- while(window_collapsed.slide_window_slice_3D(slice));
+ } while (window_collapsed.slide_window_slice_3D(slice));
}
ClLogits1DNormKernel::ClLogits1DNormKernel()
@@ -267,18 +291,24 @@ ClLogits1DNormKernel::ClLogits1DNormKernel()
_type = CLKernelType::ELEMENTWISE;
}
-void ClLogits1DNormKernel::configure(const CLCompileContext &compile_context, const ITensorInfo &src, const ITensorInfo &sum, ITensorInfo &dst, const SoftmaxKernelInfo &info)
+void ClLogits1DNormKernel::configure(const CLCompileContext &compile_context,
+ const ITensorInfo &src,
+ const ITensorInfo &sum,
+ ITensorInfo &dst,
+ const SoftmaxKernelInfo &info)
{
- auto padding_info = get_padding_info({ &src, &dst, &sum });
+ auto padding_info = get_padding_info({&src, &dst, &sum});
// Note: output should always have a scale of 1/256 and offset 0
- const bool is_quantized_asymmetric = is_data_type_quantized_asymmetric(info.input_data_type);
- const DataType output_data_type = info.input_data_type;
- const QuantizationInfo allowed_quantization_info = get_softmax_output_quantization_info(info.input_data_type, info.is_log);
- const UniformQuantizationInfo qinfo = src.quantization_info().uniform();
+ const bool is_quantized_asymmetric = is_data_type_quantized_asymmetric(info.input_data_type);
+ const DataType output_data_type = info.input_data_type;
+ const QuantizationInfo allowed_quantization_info =
+ get_softmax_output_quantization_info(info.input_data_type, info.is_log);
+ const UniformQuantizationInfo qinfo = src.quantization_info().uniform();
// Output auto initialization if not yet initialized
- auto_init_if_empty(dst, src.clone()->set_data_type(output_data_type).set_quantization_info(allowed_quantization_info));
+ auto_init_if_empty(dst,
+ src.clone()->set_data_type(output_data_type).set_quantization_info(allowed_quantization_info));
// Perform validation step
ARM_COMPUTE_ERROR_THROW_ON(validate_arguments_1DNorm(src, sum, dst, info));
@@ -311,7 +341,10 @@ void ClLogits1DNormKernel::configure(const CLCompileContext &compile_context, co
ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
}
-Status ClLogits1DNormKernel::validate(const ITensorInfo &src, const ITensorInfo &sum, const ITensorInfo &dst, const SoftmaxKernelInfo &info)
+Status ClLogits1DNormKernel::validate(const ITensorInfo &src,
+ const ITensorInfo &sum,
+ const ITensorInfo &dst,
+ const SoftmaxKernelInfo &info)
{
ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments_1DNorm(src, sum, dst, info));
@@ -343,9 +376,8 @@ void ClLogits1DNormKernel::run_op(ITensorPack &tensors, const Window &window, ::
add_3D_tensor_argument(idx, sum, sum_slice);
add_3D_tensor_argument(idx, dst, slice);
enqueue(queue, *this, slice, lws_hint());
- }
- while(window_collapsed.slide_window_slice_3D(slice));
+ } while (window_collapsed.slide_window_slice_3D(slice));
}
} // namespace kernels
} // namespace opencl
-} // namespace arm_compute \ No newline at end of file
+} // namespace arm_compute
generated by cgit v1.2.1 (git 2.23.0) at 2024-05-20 00:19:49 +0000