Integrated circuit chip apparatus

A processing system, comprising:

• a neural network computing apparatus;
• a general interconnection interface;
• and
• a general-purpose processing apparatus connected to the neural network computing apparatus via the general interconnection interface,
• wherein the neural network computing apparatus further comprises:
• a main processing circuit;
• and
• a plurality of basic processing circuits arranged in a form of an m*n array comprising m rows and n columns of basic processing circuits, wherein each basic processing circuit is connected to its adjacent basic processing circuits in the array, and the main processing circuit is connected to k basic processing circuits of the plurality of basic processing circuits arranged on one or more sides of the array, wherein the plurality of basic processing circuits comprise data type conversion circuits configured to convert data between a floating data type and a fixed point data type,
• wherein the main processing circuit is configured to transfer data and a computation instruction to the k basic processing circuits,
• wherein the k basic processing circuits are configured to forward data received from the main processing circuit to the other basic processing circuits among the plurality of basic processing circuits, wherein the data forwarded by the k basic processing circuits is in the floating point data type,
• wherein the plurality of basic processing circuits are configured to determine that the data forwarded should be converted, convert the data forwarded into the fixed point data type, perform a first set of neural network computations in parallel in the fixed point data type on the data forwarded to obtain the computation results in the fixed point data type, and transfer computation results to the main processing circuit through the k basic processing circuits,
• wherein the main processing circuit is further configured to perform a second set of neural network computations in series on the computation results.

A method for performing neural network operations using an integrated circuit chip apparatus comprising a main processing circuit, and a plurality of basic processing circuits arranged in a form of an m*n array comprising m rows and n columns of basic processing circuits, wherein each basic processing circuit is connected to its adjacent basic processing circuits in the array, and the main processing circuit is connected to k basic processing circuits of the plurality of basic processing circuits arranged on one or more sides of the array, wherein the plurality of basic processing circuits comprise data type conversion circuits configured to convert data between a floating data type and a fixed point data type, the method comprising:

• transferring data and a computation instruction from the main processing circuit to the k basic processing circuits;
• forwarding, by the k basic processing circuits, data received from the main processing circuit to the other basic processing circuits among the plurality of basic processing circuits, wherein the data forwarded by the k basic processing circuits is in the floating point data type;
• determining, by the plurality of basic processing circuits, that the data forwarded should be converted;
• converting, by the plurality of basic processing circuits, the data forwarded into the fixed point data type;
• performing, by the plurality of basic processing circuits, a first set of neural network computations in parallel in the fixed point data type on the data forwarded to obtain the computation results in the fixed point data type;
• transferring the computation results to the main processing circuit through the k basic processing circuits;
• and
• performing, by the main processing circuit, a second set of neural network computations in series on the computation results.