Scheduling method for a multi-layer convolutional neural network, electronic device and storage medium

An electronic device, comprising:

• a processor;
• and
• a memory, comprising at least one computer program module;
• wherein the at least one computer program module is stored in the memory and is configured to be executed by the processor, and the at least one computer program module is configured to implement a scheduling method,
• wherein the scheduling method comprises:
• respectively performing, by a plurality of compute units, a first convolution computation on a plurality of data groups that correspond to the plurality of compute units respectively, to obtain a plurality of first computation result groups, wherein a first convolutional layer obtained through the first convolution computation comprises the plurality of first computation result groups, and the plurality of compute units comprise a first compute unit and a second compute unit;
• determining a data replication-transmission mode corresponding to the plurality of first computation result groups in the plurality of compute units, according to a configuration rule of a second convolutional layer in the plurality of compute units, wherein the second convolutional layer is obtained after performing, by the plurality of compute units, a second convolution computation on the first convolutional layer;
• copying a first intermediate data row required by the first compute unit for padding in the second convolution computation process from a first computation result group in the second compute unit and transmitting the first intermediate data row that is copied to the first compute unit based on the data replication-transmission mode by:
• determining a size of the first intermediate data row based on a relationship between a second computation result group and a first computation result group in the first compute unit, wherein the second computation result group is a computation result obtained after the second convolution computation by the first compute unit, the first computation result group in the first compute unit is a computation result obtained after the first convolution computation by the first compute unit;
• and
• copying the first intermediate data row from the first computation result group in the second compute unit and transmitting the first intermediate data row that is copied to the first compute unit based on the data replication-transmission mode and the size of the first intermediate data row.