MLModel prediction gets blocked

Is there any limitation for using MLModel in multiple NSOperations?

I'm trying to get predictions from a MLModel, but sometimes the prediction function gets blocked.

I'm running 15 NSOperations concurrently in a queue with qualityOfService = .userInitiated.

Each operation is calling 45 times the prediction method:

  ....
  for groupedData in subArrayOfRawData {

  let mlm = try MLMultiArray(shape:[NSNumber(value:SIZE_MULTIARRAY)], dataType:.double)

  for (index, element) in groupedData.enumerated() {
  mlm[index] = NSNumber(value: element)
  }

  let x = try myModel().prediction(input1: mlm)

  results.append(x.output1)
  }
  ....

I'm running a lot of tasks in the CPU just before calling to the prediction. Is there any limitation if there is not enough resources, so it gets block?

Thanks