Time-based synchronization descriptors
In one embodiment, a system includes a peripheral device including a hardware clock, and processing circuitry to read a given work request entry stored with a plurality of work request entries in at least one work queue in a memory, the given work request entry including timing data and an operator, the timing data being indicative of a time at which a work request should be executed, retrieve a clock value from the hardware clock, and execute the work request with a workload while execution of the work request is timed responsively to the timing data and the operator and the retrieved clock value.
1 . A peripheral device, comprising:
a hardware clock; and
processing circuitry to:
read a given work request entry stored with a plurality of work request entries in at least one work queue in a memory, respective ones of the work request entries being fence entries without work requests, the fence entries including timing data that defines execution timing for other work requests, the given work request entry including timing data and an operator, the timing data being indicative of a time at which a work request should be executed;
retrieve a clock value from the hardware clock; and
execute the work request with a workload while execution of the work request is timed responsively to the timing data and the operator and the retrieved clock value, wherein:
the operator defines a condition for executing the work request based on the timing data and retrieved clock value; and
the processing circuitry includes a packet processing pipeline to perform a steering decision process based on packets associated with the work request entries.
2 . A system comprising the peripheral device according to claim 1 , and further comprising multiple host devices, each host device including:
an interface to connect with the peripheral device and transfer data with the peripheral device; and
a processor to run at least one software application, which is configured to write host-specific work request entries of the work request entries into at least one respective work queue of the at least one work queue.
3 . A system comprising the peripheral device according to claim 1 , and further comprising a host device including:
an interface to connect with the peripheral device and transfer data with the peripheral device; and
a processor to run at least one software application, which is configured to write the work request entries into the at least one work queue in the memory, wherein the peripheral device includes an interface configured to connect with the host device and transfer the data with the host device.
4 . The system according to claim 3 , wherein the host device includes the memory.
5 . The system according to claim 3 , wherein the at least one software application is configured to write the timing data according to Coordinated Universal Time (UTC) format.
6 . The system according to claim 3 , wherein the at least one software application is configured to generate respective timing data of respective ones of the work request entries for writing into the at least one work queue in the memory responsively to a time division multiplexing schedule.
7 . The system according to claim 6 , wherein the processing circuitry is configured to process packets to be sent over the network timed according to the time division multiplexing schedule responsively to the respective timing data of the respective ones of the work request entries and respective clock values retrieved from the hardware clock.
8 . The device according to claim 1 , wherein the operator is selected from any one or more of the following: a conditional operator; a comparison operator; is less than; is greater than; includes a pattern; includes a periodic pattern; or is in a range.
9 . The device according to claim 1 , wherein the operator includes a mask, which masks out most significant bits of the retrieved clock value.
10 . The device according to claim 1 , wherein respective ones of the work request entries define execution timing for respective work requests included in the respective ones of the work request entries.
11 . The device according to claim 1 , wherein at least some work request entries of the work request entries include references to locations in the memory where respective workloads are stored or should be scattered.
12 . The device according to claim 1 , further comprising a network interface controller and a network interface to share packets with remote devices over a packet data network.
13 . The device according to claim 12 , further comprising a data processing unit (DPU) comprising the network interface controller, the network interface, and at least one processing core.
14 . The device according to claim 12 , wherein:
the network interface is configured to receive packets over the network; and
the processing circuitry is configured to process the received packets timed responsively to respective timing data and respective operators of respective ones of the work request entries.
15 . The system according to claim 12 , wherein:
the processing circuitry is configured to process packets to be sent over the network timed responsively to respective timing data and respective operators of respective ones of the work request entries; and
the network interface is configured to send the packets over the network.
16 . The system according to claim 1 , wherein the processing circuitry comprises a clock controller to discipline the hardware clock responsively to a Precision Time Protocol (PTP).
17 . The system according to claim 1 , wherein the processing circuitry comprises a hardware accelerator to execute at least one of the work requests.
18 . The system according to claim 17 , wherein the hardware accelerator is configured to perform any one or more of the following:
encrypt the workload;
decrypt the workload;
compute a cyclic redundancy check (CRC) of the workload; or
compute a cryptographic hash of the workload.
19 . A system, comprising:
a plurality of peripheral devices to connect with each other over a network, and comprising respective hardware clocks to be synchronized with each other over the network; and
multiple host devices connected to the network via the peripheral devices, the host devices comprising respective processors to run respective software applications to share data with each other over the network, each of the respective software applications being configured to write into a respective work queue a plurality of work request entries including a respective work request entry including a respective operator and respective timing data indicative of a time at which a respective work request should be executed respective ones of the work request entries being fence entries without work requests, the fence entries including timing data that defines execution timing for other work requests, wherein each of the peripheral devices comprises processing circuitry to:
read the respective work request entry from the respective work queue;
retrieve a respective clock value from a respective one of the hardware clocks; and
execute the respective work request with a workload timed responsively to the respective timing data, the respective operator, and the retrieved respective clock value wherein the respective operator defines a condition for executing the respective work request based on the respective timing data and respective retrieved clock value, wherein the processing circuitry of each of the peripheral devices includes a packet processing pipeline to perform a steering decision process based on a packet associated with the respective work request entry.
20 . The system according to claim 19 , wherein the respective software applications are configured to share the data with each other over the network and generate respective work requests entries responsively to a time division multiplexing schedule.
21 . A method comprising:
reading a given work request entry stored with a plurality of work request entries in at least one work queue in a memory, respective ones of the work request entries being fence entries without work requests, the fence entries including timing data that defines execution timing for other work requests, the given work request entry including timing data and an operator, the timing data being indicative of a time at which a work request should be executed;
retrieving a clock value from a hardware clock;
executing the work request with a workload while execution of the work request is timed responsively to the timing data and the operator and the retrieved clock value wherein the operator defines a condition for executing the work request based on the timing data and retrieved clock value; and
performing a steering decision process with a packet processing pipeline based on packets associated with the work request entries.
22 . The peripheral device according to claim 1 , wherein the packet processing pipeline is to perform the steering decision process based on the packets yielding results indicating which of the work request entries the packets belong to.
23 . The peripheral device according to claim 22 , further comprising a network interface controller including a physical layer and a medium access control (MAC) chip and to process the packets which are received, or for sending, over a network.
24 . The peripheral device according to claim 23 , further comprising a clock controller to discipline the hardware clock responsively to a clock synchronization protocol.