[object Object][object Object][object Object]undefined[object Object]

Compares two vectors element by element. If the comparison result is true, the corresponding bit of the output result is 1; otherwise, it is 0. The computation formula is as follows:

$$\overrightarrow{\text{result}} = \text{asc\_gt}(\overrightarrow{src0}, \overrightarrow{src1})$$ $$\begin{bmatrix} \text{result}_0 \\ \text{result}_1 \\ \vdots \\ \text{result}_{n-1} \end{bmatrix} = \begin{bmatrix} \mathbb{I}_{(src0_0 > src1_0)} \\ \mathbb{I}_{(src0_1 > src1_1)} \\ \vdots \\ \mathbb{I}_{(src0_{n-1} > src1_{n-1})} \end{bmatrix}$$ [object Object]

• High-dimensional tiling computation

  [object Object]

• Synchronous computation

  [object Object]

[object Object][object Object]undefined[object Object]

None

[object Object]

PIPE_V

[object Object]

• For operand address overlap constraints, refer to .
• The starting addresses of dst, src0, and src1 must be 32-byte aligned.
• dst is sorted into a binary result in little-endian order, corresponding to the data comparison result at the corresponding position in src.
• When the result is output to the destination address, the mask parameter does not take effect.
• When using the register-based interface, you need to use to obtain the comparison result.
• When the result is output to a register, a fixed 256B of data is processed, and the repeat field does not take effect.
• When the result is output to a register, if a bit in the mask is 0 and the value of the corresponding bit in the output register does not change, the final result may contain residuals from the previous computation. To initialize the register, you can call the interface.

[object Object]