#include "op_plugin/AclOpsInterface.h"
#include "op_plugin/OpApiInterface.h"
#include "op_plugin/utils/op_api_common.h"
namespace op_api {
at::Tensor &le_out(const at::Tensor &self, const at::Tensor &other, at::Tensor &result)
{
DO_COMPATIBILITY(aclnnLeTensor, acl_op::le_out(self, other, result));
auto outputSize = op_infer::broadcast_ops_npu_output_size(self, other);
at_npu::native::OpPreparation::check_tensor({self}, result, result.scalar_type(), outputSize);
if (at_npu::native::OpPreparation::IsCPUScalar(self)) {
const at::Scalar self_scalar = self.item();
EXEC_NPU_CMD(aclnnGeScalar, other, self_scalar, result);
} else if (at_npu::native::OpPreparation::IsCPUScalar(other)) {
const at::Scalar other_scalar = other.item();
EXEC_NPU_CMD(aclnnLeScalar, self, other_scalar, result);
} else {
EXEC_NPU_CMD(aclnnLeTensor, self, other, result);
}
return result;
}
at::Tensor &le_out(const at::Tensor &self, const at::Scalar &other, at::Tensor &result)
{
DO_COMPATIBILITY(aclnnLeScalar, acl_op::le_out(self, other, result));
auto outputSize = self.sizes();
at_npu::native::OpPreparation::check_tensor({self}, result, result.scalar_type(), outputSize);
EXEC_NPU_CMD(aclnnLeScalar, self, other, result);
return result;
}
at::Tensor le(const at::Tensor &self, const at::Tensor &other)
{
DO_COMPATIBILITY(aclnnLeTensor, acl_op::le(self, other));
auto outputSize = op_infer::broadcast_ops_npu_output_size(self, other);
at::Tensor result =
at_npu::native::OpPreparation::apply_tensor_without_format(outputSize, self.options().dtype(at::kBool));
if (at_npu::native::OpPreparation::IsCPUScalar(self)) {
const at::Scalar self_scalar = self.item();
EXEC_NPU_CMD(aclnnGeScalar, other, self_scalar, result);
} else if (at_npu::native::OpPreparation::IsCPUScalar(other)) {
const at::Scalar other_scalar = other.item();
EXEC_NPU_CMD(aclnnLeScalar, self, other_scalar, result);
} else {
EXEC_NPU_CMD(aclnnLeTensor, self, other, result);
}
return result;
}
at::Tensor le(const at::Tensor &self, const at::Scalar &other)
{
DO_COMPATIBILITY(aclnnLeScalar, acl_op::le(self, other));
auto outputSize = op_infer::input_same_output_size(self);
at::Tensor result =
at_npu::native::OpPreparation::apply_tensor_without_format(outputSize, self.options().dtype(at::kBool));
EXEC_NPU_CMD(aclnnLeScalar, self, other, result);
return result;
}
at::Tensor &le_(at::Tensor &self, const at::Scalar &other)
{
DO_COMPATIBILITY(aclnnInplaceLeScalar, acl_op::le_(self, other));
EXEC_NPU_CMD(aclnnInplaceLeScalar, self, other);
return self;
}
at::Tensor &le_(at::Tensor &self, const at::Tensor &other)
{
DO_COMPATIBILITY(aclnnInplaceLeTensor, acl_op::le_(self, other));
if (other.dim() == 0 && !torch_npu::utils::is_npu(other)) {
return op_api::le_(self, other.item());
} else {
TORCH_CHECK(self.device() == other.device(),
"Expected all tensors to be on the same device, but found at least two devices", OPS_ERROR(ErrCode::INTERNAL));
at_npu::native::OpPreparation::CheckMemory({self, other}, {self});
EXEC_NPU_CMD(aclnnInplaceLeTensor, self, other);
return self;
}
}
}