branch prediction vs branch target prediction

问题:

Have I understood this right, if statements are more dependent on branch prediction and v-table look-up is more dependent on branch target prediction? Regarding v-tables, there is no "branch prediction", just the target prediction?

Trying to understand how a v-table is processed by the CPU.

回答1:

Branch prediction is predicting whether or not the branch will be taken. Branch target prediction is prediction where the branch is going to. These two things are independent and can occur in all combinations.

Examples of these might be:

Unconditional branch, fixed target

• Infinite loop
• goto statement
• break or continue statement
• End of the 'then' clause of an if/else statement (to jump past the else clause)
• Non-virtual function call

Unconditional branch, variable target

• Returning from a function
• Virtual function call
• Function pointer call
• switch statement (if compiled into a jump table)

Conditional branch, fixed target

• if statement
• switch statement (if compiled into a series of if/else statements)
• Loop condition tests
• The && and || operators
• The ternary ?: operator

Conditional branch, variable target

• Less likely to show up under normal conditions, but the compiler may synthesize one as an optimization, combining two of the above cases. For example, on x86, the compiler may optimize code like if (condition) { obj->VirtualFunctionCall(); } into a conditional indirect jump like jne *%eax if it appears at the end of a function due to tail call optimization.