Computerized branch predictions and decisions

A computer-implemented method of processor branch prediction, the method comprising:

• training a machine-learning model of processor branch behavior while a processing unit executes computer instructions that include branch instructions, load instructions and store instructions, wherein the load instructions and store instructions cause a control unit of the processing unit to load data from a memory into processor registers and store data from the processor registers to the memory, respectively,
• wherein,
• training the model comprises, for each of N branch instructions, N>2, encountered whilst the processing unit executes said branch instructions:
• identifying a next branch instruction;
• and
• feeding the machine-learning model with:
• an address, in the memory, of the next branch instruction identified;
• an outcome as to whether a corresponding branch was taken or not by the processing unit while executing it;
• and
• load information as to one or more past load instructions executed by the processing unit, said load information comprises one or each of:
• one or more memory addresses, from which the one or more past load instructions were respectively instructing to load data from the memory into the processor registers;
• data contents accordingly loaded from the memory by the control unit, due to said one or more past load instructions,
• the one or more memory addresses;
• and
• the data contents accordingly loaded from the memory by the control unit due to said one or more past load instructions.