SetFixpipeNz2ndFlag

Product	Supported
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	√
Atlas inference product's AI Core	x
Atlas inference product's Vector Core	x
Atlas training products	x

Sets the configuration related to in-line format conversion (from NZ to ND) during the DataCopy (CO1 -> GM and CO1 -> A1) process.

__aicore__ inline void SetFixpipeNz2ndFlag(uint16_t ndNum, uint16_t srcNdStride, uint16_t dstNdStride)

**Table 1** Parameters
Parameter	Input/Output	Meaning
ndNum	Input	Number of NDs. The type is uint16_t. Value range: ndNum ∈ [1, 65535].
srcNdStride	Input	Source stride in fractal size, indicating head-to-head offset between adjacent source NZ matrices. Atlas A2 training products/Atlas A2 inference products: srcNdStride ∈ [1, 512] (unit: fractal_size 1024 bytes) Atlas A3 training products/Atlas A3 inference products: srcNdStride ∈ [1, 512] (unit: fractal_size 1024 bytes) Atlas 200I/500 A2 inference products: srcNdStride ∈ [1, 512] (unit: fractal_size 1024 bytes)
dstNdStride	Input	Head-to-head offset between adjacent destination ND matrices. The unit is element. Atlas A2 training products/Atlas A2 inference products: dstNdStride ∈ [1, 65535] Atlas A3 training products/Atlas A3 inference products: dstNdStride ∈ [1, 65535] Atlas 200I/500 A2 inference products: dstNdStride ∈ [1, 65535]

None

None

For details, see the complete example.

uint16_t ndNum = 2;
uint16_t srcNdStride = 2;
uint16_t dstNdStride = 1;
AscendC::SetFixpipeNz2ndFlag(ndNum, srcNdStride, dstNdStride);

Parent topic: Data Movement