Tracking written addresses of a shared memory of a multi-core processor

A non-transitory machine-readable medium, having encoded thereon program code, wherein, when the program code is executed by a machine, the machine implements a method of controlling processing flow in a network processor having one or more processing modules, the method comprising:

• loading, by a given one of the processing modules, a script into a script memory of a compute engine of the network processor, the script comprising one or more instructions for the compute engine;
• loading, by the given one of the processing modules, a register file into the compute engine, the register file comprising one or more operands for the one or more instructions of the loaded script, wherein the register file comprises N portions, N a positive integer;
• initializing a tracking vector of the compute engine to a default value;
• reading the corresponding instruction of the script from an instruction memory of the compute engine;
• executing, by the compute engine, the one or more instructions of the loaded script based on the one or more operands of the loaded register file;
• updating, by the compute engine, one or more corresponding portions of the register file with updated data corresponding to the executed script;
• tracking, by the tracking vector, the updated ones of the one or more portions of the register file;
• and
• providing, by the compute engine, the tracking vector and the updated register file to the given one of the processing modules.

A network processor comprising:

• a communication link configured to receiving one or more packets;
• a plurality of processing modules, one of the plurality of processing modules configured to generate one or more tasks corresponding to each of the received packets;
• at least one system cache and at least one external memory configured to store data corresponding to each received packet;
• a given one of the processing modules configured to:
• receive a task for operation;
• request an operation by a compute engine of the processing module for a script corresponding to the received task;
• load the script into a script memory of the compute engine, the script comprising one or more instructions for the compute engine;
• load a register file into the compute engine, the register file comprising one or more operands for the one or more instructions of the loaded script, wherein the register file comprises N portions, N a positive integer;
• the compute engine configured to:
• initialize a tracking vector to a default value;
• read the corresponding instruction of the script from an instruction memory;
• execute the one or more instructions of the loaded script based on the one or more operands of the loaded register file;
• update, one or more corresponding portions of the register file with updated data corresponding to the executed script;
• update an indicator of the tracking vector corresponding to each updated portion of the register file, wherein the tracking vector comprises N indicators, each indicator corresponding to one of the N portions of the register file;
• and
• provide the tracking vector and the updated register file to the given one of the processing modules, wherein the given one of the processing modules is further configured to determine further processing of the received task based on the provided tracking vector.