Tensor

Function

Onboard tensor abstraction, in which the memory location, data type, size, and format of tensors can be specified as the data dependency identifiers of instructions.

Prototype

class Tensor(mem_type, dtype=None, size=None, format=None, is_inited=False)

Parameters

Parameter	Input Type	Description
mem_type	String	Location of the memory space where the abstracted tensor is located, such as GM, UB, L1, L0A, L0B, L0C, FB, and BT.
dtype	String	Data type, such as BOOL, UINT1, UINT2, UINT8, UINT16, UINT32, BF16, UINT64, INT4, INT8, INT16, INT32, INT64, FP16, and FP32.
size	list	Shape of a tensor.
format	String	Data layout format. For details, see the Format in Operator Development through the path: .
is_inited	bool	Switch that indicates whether the tensor class is ready. Once enabled, instructions that utilize the tensors as the input can be initiated.

Members

Member	Description
tensor.set_valid()	Enables the current tensor to be ready. Once enabled, instructions that utilize the tensor as the input can be initiated immediately.
tensor.set_invalid()	Disables the current tensor to be ready. Once disabled, instructions that utilize the tensor as the input cannot be initiated immediately.
tensor.is_valid()	Obtains the current tensor ready status.

Constraints

You need to create a tensor whose shape is [1] and is_inited is True for scalar creation.

Example

from mskpp import Tensor, Core
gm_tmp= Tensor("GM", "FP16", [48, 16], format="ND")
with Core("AIV0") as aiv:  # Computing logic on AIV0.
    ...
    gm_tmp.load(result, set_value=0)
with Core("AIC0") as aic:
    in_x = Tensor("GM", "FP16", [48, 16], format="ND")
    in_x.load(gm_tmp, expect_value=0) # Computing logic on AIC0.
    ...

Parent topic: Basic APIs