Initialisation of worker threads and associated operand registers

A method comprising:

• interleaving execution of a plurality of worker threads, wherein each of the worker threads is configured to execute a same set of code to perform operations on a different set of data held in an input buffer of a memory of a processing device;
• in response to a first instruction, which is executed prior to execution of the same set of code by the worker threads, for a first worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determining one or more variables enabling the first worker thread to locate in the input buffer a first set of data on which the operations are to be performed by the first worker thread when it executes the same set of code;
• causing at least one operand register that is associated with the first worker thread to be preloaded with the one or more variables determined for the first worker thread prior to execution of the same set of code, wherein for the first worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the first worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the first set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the first worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop;
• wherein further in response to the first instruction, which is executed prior to execution of the same set of code by the worker threads, for a second worker thread of the plurality of worker threads:
• in dependence upon an identifier of the second worker thread, determining one or more variables enabling the second worker thread to locate in the input buffer a second set of data on which the operations are to be performed by the second worker thread when it executes the same set of code;
• causing at least one operand register that is associated with the second worker thread to be preloaded with the one or more variables determined for the second worker thread prior to execution of the same set of code, wherein for the second worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the second worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the second set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the second worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop.

A non-transitory computer readable medium storing a computer program comprising a set of executable instructions, which when executed by at least one processor causes a method to be performed, the method comprising:

• interleaving execution of a plurality of worker threads, wherein each of the worker threads is configured to execute a same set of code to perform operations on a different set of data held in an input buffer of a memory of a processing device;
• and
• in response to a first instruction, which is executed prior to execution of the same set of code by the worker threads, for a first worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determining one or more variables enabling the first worker thread to locate in the input buffer a first set of data on which the operations are to be performed by the first worker thread when it executes the same set of code;
• causing at least one operand register for the first worker thread to be preloaded with the one or more variables determined for the first worker thread prior to execution of the same set of code, wherein for the first worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the first worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the first set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the first worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop,
• wherein further in response to the first instruction, which is executed prior to execution of the same set of code by the worker threads, for a second worker thread of the plurality of worker threads:
• in dependence upon an identifier of the second worker thread, determining one or more variables enabling the second worker thread to locate in the input buffer a second set of data on which the operations are to be performed by the second worker thread when it executes the same set of code;
• causing at least one operand register that is associated with the second worker thread to be preloaded with the one or more variables determined for the second worker thread prior to execution of the same set of code, wherein for the second worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the second worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the second set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the second worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop.

A processing device comprising:

• at least one execution unit configured to interleave execution of a plurality of worker threads, wherein each of the worker threads is configured to:
• execute a same set of code to perform operations on a different set of input data to produce a different set of results;
• store its set of results from the operations to an output buffer of a memory of the processing device;
• and
• a plurality of operand registers, each of which is associated with one of the worker threads and is configured to store at least one operand for at least one instruction executed by its associated one of the worker threads;
• wherein the at least one execution unit comprises a hardware module comprising processing circuitry configured to, in response to a first instruction, which is executed by the at least one execution unit prior to execution of the same set of code, for a first worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determine from control information accessible by the hardware module one or more variables enabling the first worker thread to determine locations in the output buffer to which a first set of results will be written when the same set of code is executed by the first worker thread;
• cause at least one of the operand registers that is associated with the first worker thread to be preloaded with the one or more variables determined for the first worker thread prior to execution of the same set of code, wherein for the first worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the first worker thread;
• and
• execute each iteration of the loop on a different one of a plurality of subsets of the first set of data wherein the processing circuitry is configured to receive a set of input information enabling it to, for the first worker thread, determine the one or more variables for the first worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop,
• wherein the processing circuitry is further configured to, in response to the first instruction, which is executed by the at least one execution unit prior to execution of the same set of code, for a second worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determine from control information accessible by the hardware module one or more variables enabling the second worker thread to determine locations in the output buffer to which a second set of results will be written when the same set of code is executed by the second worker thread;
• cause at least one of the operand registers that is associated with the second worker thread to be preloaded with the one or more variables determined for the second worker thread prior to execution of the same set of code, wherein for the second worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the second worker thread;
• and
• execute each iteration of the loop on a different one of a plurality of subsets of the second set of data wherein the processing circuitry is configured to receive a set of input information enabling it to, for the second worker thread, determine the one or more variables for the second worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop.

A method comprising:

• interleaving execution of a plurality of worker threads, wherein each of the worker threads is configured to:
• execute a same set of code to perform operations on a different set of input data to produce a different set of results;
• and
• store its set of results to an output buffer of a memory of a processing device;
• and
• in response to a first instruction, which is executed prior to execution of the same set of code, for a first worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determining one or more variables enabling the first worker thread to determine locations in the output buffer to which a first set of results will be written when the first set of code is executed;
• causing at least one operand register that is associated with the first worker thread to be preloaded with the one or more variables determined for the first worker thread prior to execution of the same set of code, wherein for the first worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the first worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the first set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the first worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop;
• wherein further in response to the first instruction, which is executed prior to execution of the same set of code, for a second worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determining one or more variables enabling the first worker thread to determine locations in the output buffer to which a first set of results will be written when the first set of code is executed;
• causing at least one operand register that is associated with the first worker thread to be preloaded with the one or more variables determined for the first worker thread prior to execution of the same set of code, wherein for the first worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the first worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the first set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the first worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop.

A non-transitory computer readable medium storing a computer program comprising a set of executable instruction which, when executed by at least one processor causes a method to be carried out, the method comprising:

• interleaving execution of a plurality of worker threads, wherein each of the worker threads is configured to:
• execute a same set of code to perform operations on a different set of input data to produce a different set of results;
• and
• store its set of results to an output buffer of a memory of a processing device;
• and
• in response to a first instruction, which is executed prior to execution of the same set of code, for a first worker thread of the plurality of worker threads:
• in dependence upon an identifier of the first worker thread, determining one or more variables enabling the first worker thread to determine locations in the output buffer to which a first set of results will be written when the first set of code is executed;
• causing at least one operand register that is associated with the first worker thread to be preloaded with the one or more variables determined for the first worker thread prior to execution of the same set of code,
• wherein for the first worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the first worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the first set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the first worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop;
• wherein further in response to the first instruction, which is executed prior to execution of the same set of code, for a second worker thread of the plurality of worker threads:
• in dependence upon an identifier of the second worker thread, determining one or more variables enabling the second worker thread to determine locations in the output buffer to which a second set of results will be written when the second set of code is executed;
• causing at least one operand register that is associated with the second worker thread to be preloaded with the one or more variables determined for the second worker thread prior to execution of the same set of code,
• wherein for the second worker thread, the operations comprise a same set of operations executed repeatedly in a loop and the one or more variables comprises an iteration count indicating a number of iterations of the loop to be performed by the second worker thread;
• executing each iteration of the loop on a different one of a plurality of subsets of the first set of data;
• and
• receiving a set of input information enabling a determination of the one or more variables for the second worker thread, wherein the set of input information comprises:
• a size of the subsets of each of the sets of data with respect to which the worker threads are configured to perform a single iteration of a same subset of the operations performed repeatedly in each loop.