0_hello_cann

Description

This sample is a quick start example using Runtime interfaces. Using the aclnnAdd vector addition operator from the CANN neural network operator library as the entry point, it demonstrates a minimum computation closed loop: complete initialization configuration, specify Device and create Stream, prepare input/output Tensor and output DataBuffer, query and apply workspace, execute out = self + alpha * other, synchronously get results and release resources.

Product Support

This sample supports the following products:

Product	Supported
Ascend 950PR/Ascend 950DT	Yes
Atlas A3 training series products/Atlas A3 inference series products	Yes
Atlas A2 training series products/Atlas A2 inference series products	Yes

Compile and Run

For environment installation details and general running steps, refer to README in the example directory.

CANN RUNTIME API

The key functionality points and their key interfaces involved in this sample are as follows:

• Initialization
  • Call aclInit interface for initialization configuration.
  • Call aclFinalize interface for deinitialization.
• Device and Stream Management
  • Call aclrtSetDevice interface to specify the Device for computation.
  • Call aclrtCreateStream interface to create a Stream.
  • Call aclrtSynchronizeStream interface to block and wait for Stream tasks to complete.
  • Call aclrtDestroyStream interface to destroy a Stream.
  • Call aclrtResetDeviceForce interface to forcefully reset the current computation Device and reclaim Device resources.
• Tensor and Data Description
  • Call aclCreateTensor interface to create input and output Tensors.
  • Call aclCreateScalar interface to create scaling factor alpha.
  • Call aclCreateDataBuffer and aclGetDataBufferAddr interfaces to manage output Buffer.
  • Call aclDestroyTensor, aclDestroyScalar, and aclDestroyDataBuffer interfaces to release description objects.
• Memory Management and Data Transfer
  • Call aclrtMalloc interface to allocate memory on Device.
  • Call aclrtMemcpy interface to complete Host/Device data transfer.
  • Call aclrtFree interface to release memory on Device.
• Operator Execution
  • Call aclnnAddGetWorkspaceSize interface to query workspace required for operator execution.
  • Call aclnnAdd interface to execute vector addition.

Core API

aclnnAdd

aclError aclnnAdd(
    void* workspace,         // workspace address
    uint64_t workspaceSize,  // workspace size
    aclOpExecutor* executor, // operator executor
    aclrtStream stream       // Stream
);

aclnnAddGetWorkspaceSize

aclError aclnnAddGetWorkspaceSize(
    const aclTensor* self,   // first input tensor
    const aclTensor* other,  // second input tensor
    const aclScalar* alpha,  // scaling factor
    aclTensor* out,          // output tensor
    uint64_t* workspaceSize, // [output] workspace size
    aclOpExecutor** executor // [output] operator executor
);

Computation Formula

out = self + alpha * other

Sample Output

ACL init successfully
Set device 0 successfully
Create stream successfully
Input vectors:
  self:   [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0]
  other:  [0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0]
  alpha:  1.0
Create output aclDataBuffer successfully, buffer addr = 0x...
Get workspace size successfully, workspace size = ...
Launch aclnnAdd successfully
Synchronize stream successfully

Vector addition result:
  result[0] = 1.5 (expected: 1.5)
  result[1] = 3.0 (expected: 3.0)
  result[2] = 4.5 (expected: 4.5)
  result[3] = 6.0 (expected: 6.0)
  result[4] = 7.5 (expected: 7.5)
  result[5] = 9.0 (expected: 9.0)
  result[6] = 10.5 (expected: 10.5)
  result[7] = 12.0 (expected: 12.0)

Sample run successfully!

Related Samples

• 4_custom_kernel_launch: If you need to use <<<>>> to call custom AscendC Kernel, refer to this sample.