On-chip integrated circuit, data processing device, and data processing method

Patent metadata

Field	Value
Publication number	US-11914540-B2
Application number	US-202217737415-A
Country	US
Kind code	B2
Filing date	May 5, 2022
Priority date	May 7, 2021
Publication date	Feb 27, 2024
Grant date	Feb 27, 2024

Abstract

An on-chip integrated circuit, a data processing device and a method are provided. The on-chip integrated circuit includes: a processor circuit and an accelerator circuit. The processor circuit includes a processor and a data storage area, the processor is connected to the data storage area through a first bus in the processor circuit. The accelerator circuit includes an accelerator and a second bus, the accelerator is connected to the second bus, and the second bus is bridged with the first bus corresponding to the data storage area, to perform data interaction between the accelerator and the data storage area, which can reduce the congestion on a bus of the processor and improve the quality of service of the application.

First claim

The invention claimed is:

1. An on-chip integrated circuit, comprising: a processor circuit and an accelerator circuit, wherein
the processor circuit comprises a processor and a data storage area, and the processor is connected to the data storage area through a first bus arranged in the processor circuit, and
the accelerator circuit is configured to perform data interaction between an accelerator and the data storage area, and comprises the accelerator and a second bus, the accelerator is connected to the second bus, and the second bus is bridged with the first bus,
wherein the data storage area is one of a plurality of data storage areas included in the processor circuit,
the processor circuit further comprises a plurality of first bridge points, and there is a one-to-one correspondence between the data storage areas and the first bridge points,
the accelerator circuit further comprises a plurality of second bridge points, and there is a one-to-one correspondence between the first bridge points and the second bridge points, and
each of the second bridge points is connected to a corresponding first bridge point through the second bus.

2. The on-chip integrated circuit according to claim 1 , wherein the processor circuit further comprises a plurality of processors, the plurality of data storage areas are connected through data channels provided by a first network-on-chip, and the first network-on-chip comprises the first bus.

3. The on-chip integrated circuit according to claim 1 , wherein the data storage area has a storage area identifier, and
the processor is configured to determine a target storage area identifier by using a first preset hash function, and perform data interaction with a target data storage area corresponding to the target storage area identifier through a data channel provided by a network-on-chip.

4. The on-chip integrated circuit according to claim 3 , wherein the accelerator circuit further comprises a positioning unit,
the positioning unit is configured to determine the target storage area identifier according to an address of data to be processed, and determine a target first bridge point and a target second bridge point according to the target storage area identifier; and
the accelerator is configured to perform data interaction with the target storage area corresponding to the target storage area identifier through the target second bridge point, the target first bridge point and a second bus between the target second bridge point and the target first bridge point.

5. The on-chip integrated circuit according to claim 4 , wherein a length of a target path for the accelerator to access the target data storage area through the target first bridge point and the target second bridge point is smaller than a length of a path for the accelerator to access the target data storage area through another first bridge point and another second bridge point.

6. The on-chip integrated circuit according to claim 4 , wherein a time duration for the accelerator to access the target data storage area through the target first bridge point and the target second bridge point is shorter than a time duration for the accelerator to access the target data storage area through another first bridge point and another second bridge point.

7. The on-chip integrated circuit according to claim 4 , wherein the accelerator circuit further comprises a second network-on-chip,
the second network-on-chip comprises a plurality of second data channels, each of second bridge points is connected to one of the second data channels, and the second data channels are connected to the accelerator through a gate.

8. The on-chip integrated circuit according to claim 7 , wherein the accelerator circuit comprises a plurality of accelerators, and each of the accelerators is connected to one of the second bridge points through a corresponding one of the second data channels of the second network-on-chip.

9. The on-chip integrated circuit according to claim 1 , wherein the second bus comprises one of the following topologies: a ring topology, a mesh topology, and a star topology.

10. The on-chip integrated circuit according to claim 1 , wherein the second bus comprises: a cache coherent bus or a non-cache coherent bus.

11. A data processing device, comprising an on-chip integrated circuit, wherein the on-chip integrated circuit comprises: a processor circuit and an accelerator circuit, and wherein
the processor circuit comprises a processor and a data storage area, and the processor is connected to the data storage area through a first bus arranged in the processor circuit, and
the accelerator circuit is configured to perform data interaction between an accelerator and the data storage area, and comprises the accelerator and a second bus, the accelerator is connected to the second bus, the second bus is bridged with the first bus,
wherein the data storage area is one of a plurality of data storage areas included in the processor circuit,
the processor circuit further comprises a plurality of first bridge points, and there is a one-to-one correspondence between the data storage areas and the first bridge points,
the accelerator circuit further comprises a plurality of second bridge points, and there is a one-to-one correspondence between the first bridge points and the second bridge points, and
each of the second bridge points is connected to a corresponding first bridge point through the second bus.

12. The data processing device according to claim 11 , wherein the processor circuit further comprises a plurality of processors, the plurality of data storage areas are connected through data channels provided by a first network-on-chip, and the first network-on-chip comprises the first bus,
channels are connected to the accelerator through a gate.

13. The data processing device according to claim 11 , wherein the data storage area has a storage area identifier, and
the processor is configured to determine a target storage area identifier by using a first preset hash function, and perform data interaction with a target data storage area corresponding to the target storage area identifier through a data channel provided by a network-on-chip.

14. The data processing device according to claim 13 , wherein the accelerator circuit further comprises a positioning unit,
the positioning unit is configured to determine the target storage area identifier according to an address of data to be processed, and determine a target first bridge point and a target second bridge point according to the target storage area identifier; and
the accelerator is configured to perform data interaction with the target storage area corresponding to the target storage area identifier through the target second bridge point, the target first bridge point and a second bus between the target second bridge point and the target first bridge point.

15. The data processing device according to claim 14 , wherein a length of a target path for the accelerator to access the target data storage area through the target first bridge point and the target second bridge point is smaller than a length of a path for the accelerator to access the target data storage area through another first bridge point and another second bridge point.

16. The data processing device according to claim 14 , wherein a time duration for the accelerator to access the target data storage area through the target first bridge point and the target second bridge point is shorter than a time duration for the accelerator to access the target data storage area through another first bridg

Assignees

• Lemon Inc

Inventors

• Chen Yimin
• Lu Shan
• Zhang Chuang
• Zhang Junmou
• Cheng Yuanlin
• Wang Jian

Classifications

• G06F13/4027 Primary

  using bus bridges (G06F13/4022 takes precedence) · CPC title

• G06F2213/40

  Bus coupling · CPC title

• G06F15/7807 Primary

  System on chip, i.e. computer system on a single chip; System in package, i.e. computer system on one or more chips in a single package · CPC title

• G06F2015/763

  ASIC · CPC title