In-core parallelization in a data processing apparatus and method
A data processing apparatus and a method for processing data are disclosed. The data processing apparatus comprises: multithreaded processing circuitry to perform processing operations of a plurality of micro-threads, each micro-thread operating in a corresponding execution context defining an architectural state. Thread control circuitry collects runtime data indicative of a performance metric relating to the processing operations. Decoder circuitry is responsive to a detach instruction in a first micro-thread of instructions executed in a first execution context defining a first architectural state, the detach instruction specifying an address, to provide detach control signals to the thread control circuitry. When the runtime data meet a parallelization criterion, the thread control circuitry is responsive to the detach control signals to spawn a second micro-thread of instructions executed in a second execution context defining a second architectural state based on the first architectural state, the second micro-thread of instructions comprising a subset of instructions of the first micro-thread of instructions starting at the address.
1. A data processing apparatus comprising:
multithreaded processing circuitry to perform processing operations of a plurality of micro-threads, each micro-thread operating in a corresponding execution context defining an architectural state;
thread control circuitry to collect runtime data indicative of a performance metric relating to the processing operations; and
decoder circuitry responsive to a detach instruction in a first micro-thread of instructions executed in a first execution context defining a first architectural state, the detach instruction specifying an address, when the runtime data meet a parallelization criterion at a point where the detach instruction is decoded, to provide detach control signals to the thread control circuitry,
wherein the thread control circuitry is responsive to the detach control signals to spawn a second micro-thread of instructions executed in a second execution context defining a second architectural state based on the first architectural state, the second micro-thread of instructions comprising a subset of instructions of the first micro-thread of instructions starting at the address.
2. The data processing apparatus of claim 1 , wherein the address is a first address, the subset of instructions is a first subset of instructions, and the detach instruction further comprises a second address, and
wherein when the runtime data meet the parallelization criterion, the thread control circuitry is responsive to the detach control signals to cause the first micro-thread of instructions to execute a second subset of instructions starting at the second address.
3. The data processing apparatus of claim 2 , wherein the first subset of instructions and the second subset of instructions are subsets of the first micro-thread of instructions that are non-overlapping during dynamic execution.
4. The data processing apparatus of claim 1 , wherein when the runtime data does not meet the parallelization criterion, the thread control circuitry is adapted to interpret the detach instruction as one of: an unconditional branch instruction, and a no-operation.
5. The data processing apparatus of claim 1 , wherein the decoder circuitry is responsive to a reattach instruction in a particular micro-thread of instructions, to provide reattach control signals to the thread control circuitry, and
wherein the thread control circuitry is responsive to the reattach control signals to terminate the particular micro-thread of instructions.
6. The data processing apparatus of claim 5 , wherein the thread control circuitry is responsive to occurrence of a data hazard between the particular micro-thread of instructions and another micro-thread of instructions to indicate an unsuccessful termination of the particular micro-thread; and
the thread control circuitry is responsive to the indication of the unsuccessful termination to re-spawn the particular micro-thread of instructions.
7. The data processing apparatus of claim 5 , wherein when the runtime data does not meet the parallelization criterion, the thread control circuitry is adapted to interpret the reattach instruction as one of: an unconditional branch instruction, and a no-operation.
8. The data processing apparatus of claim 5 , wherein each micro-thread is identified by an epoch identifier.
9. The data processing apparatus of claim 5 , wherein the particular micro-thread of instructions is the first micro-thread of instructions, and
wherein in response to the detach instruction the thread control circuitry is adapted to assign a new epoch identifier to the second micro-thread of instructions.
10. The data processing apparatus of claim 5 , wherein the particular micro-thread of instructions is the second micro-thread of instructions, and
wherein prior to the detach instruction the first micro-thread was assigned a first epoch identifier and, in response to the detach instruction, the thread control circuitry is adapted to assign a new epoch identifier to the first micro-thread of instructions and to assign the first epoch identifier to the second micro-thread of instructions.
11. The data processing apparatus of claim 5 , wherein the detach instruction specifies a region identifier and at least one subsequent reattach instruction specifies the region identifier.
12. The data processing apparatus of claim 11 , wherein the region identifier is a first region identifier, a subsequent detach instruction specifies a second region identifier and a subsequent reattach instruction specifies the second region identifier,
wherein the epoch identifier comprises an outer epoch identifier and an inner epoch identifier, and
wherein when the subsequent detach instruction and the subsequent reattach instruction are between the detach instruction and the reattach instruction in program order, the outer epoch identifier is assigned in association with the detach instruction, and the inner epoch identifier is assigned in association with the subsequent detach instruction.
13. The data processing apparatus of claim 12 , wherein the runtime data is first runtime data and the performance metric is a first performance metric associated with the first region identifier, and
wherein the thread control circuitry is adapted to collect second runtime data indicative of a second performance metric associated with the second region identifier.
14. The data processing apparatus of claim 11 , wherein the decoder circuitry is responsive to a sync instruction in a micro-thread of instructions with a youngest epoch identifier to pause execution of subsequent instructions in the youngest micro-thread of instructions until all spawned micro-threads with older epoch identifiers and a same region identifier have terminated.
15. The data processing apparatus of claim 14 , wherein the decoder circuitry is further responsive to the sync instruction in the micro-thread of instructions with the youngest epoch identifier to pause execution of subsequent instructions in the youngest micro-thread of instructions until all spawned micro threads with older epoch identifiers and a different region identifier have terminated.
16. The data processing apparatus of claim 5 , wherein, based on a scheduling factor n, the thread control circuitry is adapted to interpret occurrences of the detach instruction and occurrences of the reattach instruction as one of an unconditional branch instruction and a no-operation.
17. The data processing apparatus of claim 16 , wherein the thread control circuitry is adapted to spawn the second micro-thread of instructions in response to every n th occurrence of the detach instruction and to interpret all other occurrences of the detach instruction as one of the unconditional branch instruction and the no-operation, and
wherein the thread control circuitry is adapted to terminate the particular micro-thread of instructions in response to every (n-1) th occurrence of the reattach instruction and to interpret all other occurrences of the reattach instruction as one of the unconditional branch instruction and the no-operation.
18. The data processing apparatus of claim 16 , wherein the scheduling factor n is based on the runtime data.
19. The data processing apparatus of claim 14 , wherein the decoder circuitry is further responsive to the detach instruction specifying an alternative address, and wherein, when n is greater than 1, the thread control circuitry is configured to:
begin execution in the second execution context from the alternative address; or
cause the first micro-thread of instructions to continue execution from the alternative address.
20. A data processing method comprising:
collecting runtime data indicative of a performance metric relating to processing operations of a plurality of micro-threads in a corresponding execution context defining an architectural state on multithreaded processing circuitry;
generating, when the runtime data meet a parallelization criterion at a point where the detach instruction is decoded, detach control signals in response to a detach instruction in a first micro-thread of instructions executed in a first execution context defining a first architectural state, the detach instruction specifying an address; and
spawning, in response to the detach control signals, a second micro-thread of instructions executed in a second execution context defining a second architectural state based on the first architectural state, the second micro-thread of instructions comprising a subset of instructions of the first micro-thread of instructions starting at the address.