Branch prediction throughput by skipping over cachelines without branches
According to one general aspect, an apparatus may include a branch prediction circuit configured to predict if a branch instruction will be taken or not. The apparatus may include a branch target buffer circuit configured to store a memory segment empty flag that indicates whether or not the memory segment after a target address includes at least one other branch instruction, wherein the memory segment empty flag was created during a commit stage of a prior occurrence of the branch instruction. The branch prediction circuit may be configured to skip over the memory segment if the memory segment empty flag indicates a lack of other branch instruction(s).
1. An apparatus comprising:
a branch prediction circuit configured to predict that a branch instruction will be taken; and
a return address stack circuit configured to store a memory segment empty flag that indicates that a memory segment after a return address does not include at least one other branch instruction;
determine that a physical address translation for a next memory segment and a following sequential memory segment are available; and
wherein the branch prediction circuit is configured to skip over the memory segment associated with the memory segment empty flag that indicates a lack of at least one other branch instruction.
2. The apparatus of claim 1 , wherein the branch prediction circuit is configured to:
determine that a next memory segment is stored within an instruction cache and the return address stack circuit; and
skip over the memory segment that the memory segment empty flag indicates a lack of branch instructions.
3. The apparatus of claim 1 , wherein the branch prediction circuit is configured to, for a memory segment that includes at least one other branch instruction after a return address, move to a next instruction within the memory segment.
4. The apparatus of claim 1 , wherein the memory segment is a cacheline.
5. The apparatus of claim 1 , wherein the branch prediction circuit is configured to determine if the branch instruction is one of a call instruction or a return instruction.
6. The apparatus of claim 5 , further comprising a branch target buffer circuit configured to store the memory segment empty flag for a target address; and
wherein, in response to the branch instruction being a call instruction, the apparatus is configured to:
determine that the memory segment after an associated return instruction includes at least one other branch instruction, and
store a result of the determination as a memory segment empty flag within the return address stack circuit.
7. The apparatus of claim 5 , wherein branch target buffer circuit is further configured to store a second memory segment empty flag that indicates or not that the memory segment after a target return address does not include at least one other branch instruction, wherein the memory segment empty flag was created during a commit stage of a prior occurrence of the branch instruction.
8. The apparatus of claim 1 , wherein the branch prediction circuit is configured to:
proceed to skip over the memory segment that the memory segment empty flag indicates a lack of branch instructions.
9. The apparatus of claim 1 , wherein the branch detection circuit is configured to, in response to the current instruction not being a branch instruction, determine that a previously stored last committed branch instruction is still valid.
10. The apparatus of claim 9 , further comprising
a branch target buffer circuit configured to store:
a branch instruction address, and
a memory segment empty flag that indicates whether a portion of the memory segment after a target address includes at least one branch instruction; and
wherein the branch detection circuit is configured to, in response to the current instruction not being a branch instruction and the previously stored last committed branch instruction being still valid, set, in the branch target buffer circuit, a memory segment empty flag associated with the previously stored last committed branch instruction.
11. The apparatus of claim 9 , wherein the branch detection circuit is configured to, if the current instruction is not a branch instruction and the previously stored last committed branch instruction is not valid, mark the previously stored last committed branch instruction as not valid.
12. The apparatus of claim 1 , wherein the branch target buffer includes a graph-based branch target buffer circuit.
13. The apparatus of claim 1 , wherein the memory segment empty flag that indicates a number of memory segments, or portions thereof, after the branch instruction address that do not include at least one other branch instruction.
14. An apparatus comprising:
a branch detection circuit configured to detect, during a commit stage of a current instruction, wherein the commit stage includes a pipeline stage in which the apparatus determines that an instruction and the instruction's results will be kept, an existence of at least one branch instruction stored within a portion of a memory segment; and
a return address circuit configured to store:
a return address, and
a memory segment empty flag that indicates whether a portion of the memory segment after a return address includes at least one branch instruction.
15. The apparatus of claim 14 , wherein the memory segment is a cacheline.
16. The apparatus of claim 14 , wherein the apparatus comprises a commit queue circuit;
wherein the commit queue circuit is configured to store currently committing instructions in age order.
17. The apparatus of claim 14 , wherein the apparatus comprises a last committed branch memory configured to store a previously committed branch instruction.
18. The apparatus of claim 17 , wherein branch detection circuit is configured to:
determine that the current instruction is a branch instruction; and
store the current instruction in the last committed branch memory.
19. A system comprising:
a branch detection circuit configured to detect, during a commit stage of a current committing instruction, wherein the commit stage includes a pipeline stage in which the system determines that an instruction and the instruction's results will be kept, an existence or at least one branch instruction stored with a portion of a memory segment;
a branch target buffer circuit configured to store:
a branch instruction address, and
a memory segment empty flag that indicates whether a portion of the memory segment after a target address includes at least one other branch instruction; and
a branch prediction circuit configured to predict that a branch instruction will be taken, and wherein the branch prediction circuit is configured to skip over the memory segment that an associated memory segment empty flag indicates a lack of branch instructions.
20. The system of claim 19 , wherein the memory segment empty flag is only valid for taken branch instructions.