Get

// Allocate buffer for TBuf initialization. The length of the allocated buffer is 1024 bytes.
AscendC::TPipe pipe;
AscendC::TBuf<AscendC::TPosition::VECCALC> calcBuf; // The template parameter is the VECCALC type in TPosition.
uint32_t byteLen = 1024;
pipe.InitBuffer(calcBuf, byteLen);
// Obtain a tensor from calcBuf. The tensor size is the size of all buffers allocated by the pipe (1024 bytes).
AscendC::LocalTensor<int32_t> tempTensor1 = calcBuf.Get<int32_t>();
// Obtain a tensor from calcBuf. The tensor size is the buffer size (512 bytes) of 128 int32_t elements.
AscendC::LocalTensor<int32_t> tempTensor1 = calcBuf.Get<int32_t>(128);
/* In complex computing scenarios, TBuf can be used as a temporary variable to store intermediate calculation results, avoiding complex dequeue and enqueue processes.
* The following code comes from the API for calculating distances. Matrix C needs to be divided by the point multiplication result of matrix A and matrix B. In this algorithm, all matrices are converted into vectors in advance.
 */
auto normADotB = calcBuf.Get <int32_t>(); // Store the result of point multiplication of matrix A and matrix B.
auto normB = qidVecIn.AllocTensor<DTypeOut>();
...
normB= qidVecIn.DeQue<DTypeOut>(); // Obtain matrix B.
for(int i = 0; i < tiling.baseM; i++) {
   AscendC::Muls(normADotB[i * tiling.baseN], normB, normA.GetValue(i), tiling.baseN); // Matrices A and B are converted into vectors for number multiplication. normADotB is used as a temporary variable to store the result.
}
qidVecIn.FreeTensor(normB);
...
for(int i = 0; i < tiling.baseM; i++) {
   AscendC::Mul(baseCVecFloat[i * tiling.baseN], baseCVecFloat[i * tiling.baseN], normADotB[i * tiling.baseN], tiling.baseN); // Obtain matrix C through calculation and divide matrix C by normADotB.
}