UnPack

Applicability

Product	Supported
Atlas 350 Accelerator Card	√
Atlas A3 training product/Atlas A3 inference product	x
Atlas A2 training product/Atlas A2 inference product	x
Atlas 200I/500 A2 inference product	x
Atlas inference product AI Core	x
Atlas inference product Vector Core	x
Atlas training product	x

Function Usage

For unsigned integers, writes the elements in the least significant half or most significant half of the source operand srcReg to dstReg with most significant bits filled with 0s to extend the bit width. For signed integers, writes the elements in the least significant half or most significant half of the source operand srcReg to dstReg with the sign bit preserved to extend the bit width.

Prototype

template <typename T = DefaultType, typename U = DefaultType, HighLowPart part = HighLowPart::LOWEST, typename S, typename V>
__simd_callee__ inline void UnPack(S& dstReg, V& srcReg)

Parameters

**Table 1** Template parameters
Parameter	Description
T	Data type of the destination operand. For the Atlas 350 Accelerator Card, the supported data types are int16_t, uint16_t, int32_t, uint32_t, int64_t, and uint64_t. For details about the data type constraints of the source operand and destination operand, see Table 3.
U	Data type of the source operand. For the Atlas 350 Accelerator Card, the supported data types are int8_t, uint8_t, int16_t, uint16_t, int32_t, and uint32_t.
part	Enum type. It is used to specify whether to read the least significant half or most significant half of srcReg. HighLowPart::LOWEST: least significant bit mode, which reads the least significant half of srcReg. HighLowPart::HIGHEST: most significant bit mode, which reads the most significant half of srcReg. Note: RegTraitNumTwo supports only the LOWEST mode.
S	RegTensor type of the destination operand.
V	RegTensor type of the source operand.

**Table 2** Function parameters
Parameter	Description
dstReg	Destination operand. The type is RegTensor.
srcReg	Source operand. The type is RegTensor.

**Table 3** Mapping between the data types of the source and destination operands
T	U
int16_t	int8_t
uint16_t	uint8_t
int32_t	int16_t
uint32_t	uint16_t
uint64_t	uint32_t
int64_t	int32_t

Returns

None

Constraints

None

Example

template<typename T, typename U, int32_t mode>
__simd_vf__ inline void UnPackVF(__ubuf__ T* dstAddr, __ubuf__ U* srcAddr, uint32_t oneDstRepSize, uint16_t repeatTimes, uint32_t oneSrcRepSize)
{
    AscendC::Reg::RegTensor<U> srcReg;
    AscendC::Reg::RegTensor<T> dstReg;
    AscendC::Reg::MaskReg mask = AscendC::Reg::CreateMask<T, AscendC::Reg::MaskPattern::ALL>();
    for (uint16_t i = 0; i < repeatTimes; i++) {
        AscendC::Reg::LoadAlign(srcReg, srcAddr + i * oneSrcRepSize);
        if constexpr (mode == 0) {
            AscendC::Reg::UnPack<T, U, AscendC::Reg::HighLowPart::LOWEST>(dstReg, srcReg);
        } else if constexpr (mode == 1) {
            AscendC::Reg::UnPack<T, U, AscendC::Reg::HighLowPart::HIGHEST>(dstReg, srcReg);
        }
        AscendC::Reg::StoreAlign(dstAddr + i * oneDstRepSize, dstReg, mask);
    }
}

Parent topic: Data Compression