Using a vector processor to configure a direct memory access system for feature tracking operations in a system on a chip
In various examples, a VPU and associated components may be optimized to improve VPU performance and throughput. For example, the VPU may include a min/max collector, automatic store predication functionality, a SIMD data path organization that allows for inter-lane sharing, a transposed load/store with stride parameter functionality, a load with permute and zero insertion functionality, hardware, logic, and memory layout functionality to allow for two point and two by two point lookups, and per memory bank load caching capabilities. In addition, decoupled accelerators may be used to offload VPU processing tasks to increase throughput and performance, and a hardware sequencer may be included in a DMA system to reduce programming complexity of the VPU and the DMA system. The DMA and VPU may execute a VPU configuration mode that allows the VPU and DMA to operate without a processing controller for performing dynamic region based data movement operations.
1. A system comprising:
a memory;
a direct memory access (DMA) system coupled to the memory; and
a processor coupled to the memory,
wherein the system executes operations comprising:
computing, using the processor and based at least in part on first data written to memory using the DMA system, a first output corresponding to one or more first updated locations of a tracked feature;
updating, using the processor, the memory to include second data representative of one or more address/data pairs corresponding to the one or more first updated locations;
updating, using the DMA system and based at least in part on the one or more address/data pairs, one or more descriptors corresponding to the tracked feature;
writing, using the DMA system and based at least in part on the one or more descriptors, third data to the memory; and
computing, using the processor and based at least in part on the third data, a second output corresponding to one or more second updated locations of the tracked feature.
2. The system of claim 1 , further comprising a processing controller, wherein the system further executes operations comprising configuring, using the processing controller, the DMA system and the processor.
3. The system of claim 1 , further comprising a processing controller, wherein the system further executes operations comprising triggering, using the processing controller, the DMA system and the processor.
4. The system of claim 1 , further comprising a descriptor memory, wherein the one or more descriptors are updated in the descriptor memory.
5. The system of claim 1 , wherein the operations are performed in a loop at each iteration of a plurality of iterations.
6. The system of claim 1 , wherein the second data corresponds to a data format including one or more bytes that represent an address of the one or more address/data pairs and one or bytes that represent data of the one or more address/data pairs.
7. The system of claim 1 , wherein the system is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing deep learning operations;
a system on chip (SoC);
a system including a programmable vision accelerator (PVA);
a system including a vison processing unit;
a system implemented using an edge device;
a system implemented using a robot;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.
8. A system comprising:
a vector memory (VMEM);
a direct memory access (DMA) system coupled to the VMEM; and
a vector processing unit (VPU) coupled to the VMEM,
wherein the system executes operations comprising:
computing, using the VPU and based at least in part on first data written to VMEM using the DMA system, a first output corresponding to one or more first updated locations of a tracked feature;
updating, using the VPU, the VMEM to include second data representative of one or more address/data pairs corresponding to the one or more first updated locations;
updating, using the DMA system and based at least in part on the one or more address/data pairs, one or more descriptors corresponding to the tracked feature;
writing, using the DMA system and based at least in part on the one or more descriptors, third data to the VMEM; and
computing, using the VPU and based at least in part on the third data, a second output corresponding to one or more second updated locations of the tracked feature.
9. The system of claim 8 , further comprising a processing controller, wherein the system further executes operations comprising configuring, using the processing controller, the DMA system and the VPU.
10. The system of claim 8 , further comprising a processing controller, wherein the system further executes operations comprising triggering, using the processing controller, the DMA system and the VPU.
11. The system of claim 8 , further comprising a descriptor memory, wherein the one or more descriptors are updated in the descriptor memory.
12. The system of claim 8 , wherein the operations are performed in a loop at each iteration of a plurality of iterations.
13. The system of claim 8 , wherein the second data corresponds to a data format including one or more bytes that represent an address of the one or more address/data pairs and one or bytes that represent data of the one or more address/data pairs.
14. The system of claim 8 , wherein the system is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing deep learning operations;
a system on chip (SoC);
a system including a programmable vision accelerator (PVA);
a system including a vison processing unit;
a system implemented using an edge device;
a system implemented using a robot;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.
15. A method comprising:
computing, using a vector processing unit (VPU) and based at least in part on first data written to vector memory (VMEM) using a direct memory access (DMA) system, a first output corresponding to one or more first updated locations of a tracked feature;
updating, using the VPU, the VMEM to include second data representative of one or more address/data pairs corresponding to the one or more first updated locations;
updating, using the DMA system and based at least in part on the one or more address/data pairs, one or more descriptors corresponding to the tracked feature;
writing, using the DMA system and based at least in part on the one or more descriptors, third data to the VMEM; and
computing, using the VPU and based at least in part on the third data, a second output corresponding to one or more second updated locations of the tracked feature.
16. The method of claim 15 , further comprising configuring, using a processing controller, the DMA system and the VPU.
17. The method of claim 15 , further comprising triggering, using a processing controller, the DMA system and the VPU.
18. The method of claim 15 , further comprising updating the one or more descriptors in a descriptor memory.
19. The method of claim 15 , wherein the method is performed in a loop at each iteration of a plurality of iterations.
20. The method of claim 15 , wherein the second data corresponds to a data format including one or more bytes that represent an address of the one or more address/data pairs and one or bytes that represent data of the one or more address/data pairs.