* Copyright (c) 2025 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
#include <sstream>
#include "ascir_ops.h"
#include "common_utils.h"
#include "common/checker.h"
#include "api_call/utils/api_call_factory.h"
#include "reg_api_call_utils.h"
#include "reg_nddma_api_call.h"
using namespace af::ops;
using namespace af::ascir_op;
namespace {
constexpr uint64_t kNddmaMaxLen = 5UL;
}
namespace codegen {
Status NddmaApiCall::ParseAttr(const ascir::NodeView &node) {
(void)node->attr.ir_attr->GetAttrValue("offset", offset_);
return ge::SUCCESS;
}
Status NddmaApiCall::Generate(const TPipe &tpipe, const std::vector<ascir::AxisId> ¤t_axis,
const std::vector<std::reference_wrapper<const Tensor>> &inputs,
const std::vector<std::reference_wrapper<const Tensor>> &outputs,
std::string &result) const {
std::stringstream ss;
const auto &gm = inputs[0].get();
const auto &ub = outputs[0].get();
if (tpipe.cv_fusion_type == ascir::CubeTemplateType::kUBFuse) {
GE_ASSERT_TRUE(gm.axis_size.size() >= 2U, "Nddma src axis-size less than 2 is invalid in CV-Fusion case");
auto last_index = gm.axis_size.size() - 1U;
auto second_to_last_index = gm.axis_size.size() - 2U;
ss << "const int64_t output_dims_" << ub.id << "[2] = {curAivM, curAlignN};" << std::endl;
std::string gm_offset;
if ((gm.axis_size[second_to_last_index] == One) && (gm.axis_size[last_index] == One)) {
ss << "const int64_t input_stride_" << ub.id << "[2] = {0, 0};" << std::endl;
ss << "const int64_t output_stride_" << ub.id << "[2] = {curAlignN, 1};" << std::endl;
gm_offset = "batch_num";
} else if (gm.axis_size[second_to_last_index] == One) {
ss << "const int64_t input_stride_" << ub.id << "[2] = {0, 1};" << std::endl;
ss << "const int64_t output_stride_" << ub.id << "[2] = {curAlignN, 1};" << std::endl;
gm_offset = "offset % shapeN + batch_num * shapeN";
} else if (gm.axis_size[last_index] == One) {
ss << "const int64_t input_stride_" << ub.id << "[2] = {1, 0};" << std::endl;
ss << "const int64_t output_stride_" << ub.id << "[2] = {curAlignN, 1};" << std::endl;
gm_offset = "offset / shapeN";
}
ss << api_name_ << "(" << ub << ", " << gm << "[" << gm_offset << "], " << "output_dims_" << ub.id << ", "
<< "output_stride_" << ub.id << ", " << "input_stride_" << ub.id << ");" << std::endl;
} else {
DataCopyParams param;
GE_ASSERT_TRUE(CalculateDmaParams(tpipe, ub, ub, param, true), "CalculateDmaParams failed");
const std::string gm_offset = ub.is_ub_scalar ? "0" : tpipe.tiler.Offset(current_axis, ub.axis, ub.axis_strides);
if (param.repeats.size() <= kNddmaMaxLen) {
NddmaParams nddma_param;
SetNddmaParams(tpipe, param, nddma_param, ub.id, ss);
ss << api_name_ << "(" << ub << ", " << gm << "[" << gm_offset << " + " << tpipe.tiler.Size(offset_) << "], "
<< "output_dims_" << ub.id << ", " << "output_stride_" << ub.id << ", " << "input_stride_" << ub.id << ");"
<< std::endl;
} else {
CreateNddmaCall(tpipe, gm, ub, gm_offset, param, offset_, ss);
}
}
result = ss.str();
return ge::SUCCESS;
}
static ApiCallRegister<NddmaApiCall> register_nddma_api_call("NddmaApiCall");
}