Parallel access to volatile memory by a processing device for machine learning

1. A system, comprising:

a processing device;

a plurality of memory regions configured to store data used by the processing device; and

at least one controller coupled to the plurality of memory regions and configured to:

read data from a first memory region of the plurality of memory regions, including reading first data from the first memory region, the first data for use by the processing device in processing associated with machine learning; and

write data, received as output data from a computer model, to a second memory region of the plurality of memory regions, including writing second data to the second memory region so that the second data can be accessed sequentially when subsequently read from the second memory region by the processing device for use by the computer model;

wherein reading the first data and writing the second data are performed in parallel.

2. The system of claim 1 , wherein the at least one controller comprises a respective controller used for read or write access to each of the memory regions.

3. The system of claim 1 , wherein the first memory region is used in a continuous burst mode when the first data is read.

4. The system of claim 3 , wherein the second memory region is used in a continuous burst mode when the second data is written.

5. The system of claim 1 , wherein:

the first data is used as an input to a neural network;

the second data is an output from the neural network; and

during the processing associated with machine learning, the first memory region operates in a read-only mode, and the second memory region operates in a write-only mode.

6. The system of claim 1 , wherein each of the plurality of memory regions is a bank in a volatile memory.

7. The system of claim 6 , wherein the volatile memory is a dynamic random access memory.

8. The system of claim 1 , wherein the processing device, the plurality of memory regions, and the at least one controller are disposed on the same chip or die, and the processing device uses the first data as an input to the computer model for machine learning.

9. The system of claim 8 , wherein the computer model is a neural network.

10. The system of claim 8 , wherein the processing device provides the second data as an output from the computer model.

11. A method, comprising:

reading, by a first controller and in response to a request from a processing device, first data from a first memory region of a plurality of memory regions;

performing, by the processing device, processing associated with a neural network, wherein the first data is an input to the neural network, and second data is provided as an output from the neural network;

structuring, by the processing device, the second data for storage in a second memory region of the plurality of memory regions so that the second data can be accessed sequentially for use by the neural network when subsequently read from the second memory region; and

writing, by a second controller, the second data to the second memory region, wherein writing the second data is performed in parallel with reading the first data.

12. The method of claim 11 , further comprising reading, by the second controller, the second data from the second memory region, wherein the second data is accessed sequentially from the second memory region for use as an input to the neural network.

13. The method of claim 11 , further comprising reading, by a third controller, third data from a third memory region of the plurality of memory regions, wherein reading the second data and reading the third data are performed in parallel, and wherein the third data is used as an input to the neural network.

14. The method of claim 11 , wherein the processing associated with the neural network comprises training the neural network using data stored in the first memory region.

15. The method of claim 11 , further comprising:

receiving, by the processing device, data from a sensor;

wherein the processing associated with the neural network comprises using the data from the sensor as an input to the neural network.

16. The method of claim 11 , further comprising determining, by the processing device, a type of the neural network, wherein structuring the second data for storage in the second memory region is based on the determined type.

17. The method of claim 11 , wherein the first memory region is used in a continuous burst mode when reading the first data, and the second memory region is used in a continuous burst mode when writing the second data.

18. The method of claim 11 , wherein, during the processing associated with the neural network, the first memory region operates in a read-only mode, and the second memory region operates in a write-only mode.

19. A non-transitory computer-readable storage medium storing instructions that, when executed by a processing device, cause the processing device to perform a method, the method comprising:

reading first data from a first bank of a dynamic random access memory;

performing processing associated with a neural network, wherein the first data is an input to the neural network, and second data is provided as an output by the neural network; and

writing the second data to a second bank of the dynamic random access memory so that the second data can be accessed sequentially when subsequently read from the second memory region by the processing device for use by the neural network, wherein writing the second data is performed in parallel with reading the first data.