Multi-core processor and working method thereof

By introducing status monitoring information tables of main core processing units and microcore array processing units in multi-core processors, tasks matching and allocation are achieved, and the problems of limited microcore capabilities and insufficient resource utilization are solved, and processor performance is improved and avoided. Micronuclear conflict.

CN119025472APending Publication Date: 2024-11-26HUNAN ADVANCECHIP ELECTRONICS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411106281.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

In existing heterogeneous multi-core processors, microcores can only run some main core functions, and their capabilities are limited, which increases the burden on the main core and reduces the overall performance. The lack of effective management mechanisms leads to insufficient resource utilization and the microcore between the Communication conflict.

Method used

Design a multi-core processor, including the main core processing unit and the micro-core array processing unit, match and assign tasks through the status monitoring information table, the main core processing unit loads the target program file, splits it into computer tasks, and separates it in an independent The micro-core is called in the area to perform tasks to achieve reasonable allocation and management of resources and avoid conflicts between the micro-cores.

Benefits of technology

It improves the working efficiency of multi-core processors, improves overall performance, ensures full utilization of resources, and avoids communication conflicts between microcores.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119025472A_ABST
    Figure CN119025472A_ABST
Patent Text Reader

Abstract

The invention provides a multi-core processor and a working method thereof, the multi-core processor comprises a main core processing unit and a micronucleus array processing unit, the main core processing unit is electrically connected with the micronucleus array processing unit, the micronucleus array processing unit comprises a plurality of mutually independent areas, and each area comprises at least one micronucleus. The micronucleus array processing unit is divided into a plurality of areas with fixed sizes, each area comprises at least one micronucleus, each area and the micronucleus in the area are managed through the main core processing unit, and computer tasks are reasonably distributed to the areas, so that the micronucleus can independently and efficiently execute, and the processing efficiency of the micronucleus is improved. Or cooperating with other micronuclei in the same area; the working efficiency of the multi-core processor is effectively improved, and the overall performance is improved; and an effective management mechanism of the main core processing unit also enables resources to be fully utilized, and communication conflicts among a plurality of micro cores are avoided at the same time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of multi-core and many-core processors, and in particular to a multi-core processor and a working method thereof. Background Art

[0002] With the development of information technology, the demand for computer processors' processing power is getting higher and higher. Traditional single-core processors can no longer meet the needs, so multi-core processors and many-core processors have gradually been widely used. A multi-core processor is a processor that integrates two or more processor cores. By dividing tasks between the cores, multiple processor cores can run in parallel, thereby improving the computing power of the entire processor.

[0003] As an extension of multi-core processors and many-core processors, "on-chip heterogeneity" has also emerged. Heterogeneous many-core processors refer to a chip that integrates multiple processors, and different processor cores have different structures and functions, such as a combination of CPU and GPU. The challenge of heterogeneous many-core systems is that they need to solve the coordination problems between different types of processor cores, such as task division, scheduling, mapping, load balancing, synchronization, communication, etc. These problems involve the design methods and key technologies of heterogeneous many-core systems, which are the hot spots and difficulties of current research.

[0004] In the prior art, the micro-core of the heterogeneous multi-core processor is a processor that can only run some of the main core functions and can only recognize simple instructions. The main core needs to split complex instructions and convert them into simple instructions for issuance, which results in very limited capabilities of the micro-core, increases the burden on the main core, reduces overall performance, and the main core can only control a limited number of micro-cores, resulting in insufficient concurrency. At the same time, when the main core issues tasks, it does not explain the issuance rules, nor does it define a reasonable collaboration mechanism. When there are a large number of micro-cores, there is a lack of effective management mechanism, resulting in unreasonable utilization of resources, that is, some micro-cores have heavy tasks, while some micro-cores are always idle; and all micro-cores can communicate with each other, resulting in conflicts between micro-cores when there are a large number of micro-cores. Summary of the invention

[0005] The present application provides a multi-core processor and a working method thereof, which solves the problem of existing heterogeneous many-core processors, in which the micro-core can only run part of the main core functions and can only recognize simple instructions, resulting in very limited capabilities of the micro-core, increasing the burden on the main core, reducing the overall performance, and the main core can only control a limited number of micro-cores, resulting in insufficient concurrency; at the same time, when the main core issues tasks, there are no issuance rules, no reasonable collaboration mechanism is defined, and there is a lack of effective management mechanism, resulting in unreasonable utilization of resources, and all micro-cores can communicate with each other, which may cause conflicts between each other.

[0006] In order to solve the above technical problems, the technical solution adopted in the present application is to provide a multi-core processor, including a main core processing unit and a micro-core array processing unit, the main core processing unit is electrically connected to the micro-core array processing unit, the micro-core array processing unit includes multiple independent slices, each of which includes at least one micro-core; the main core processing unit is used to load the target program file and split the target program file into multiple computer tasks; the main core processing unit is also used to match the corresponding slices for multiple computer tasks according to the status monitoring information table of the micro-core array processing unit, and select the corresponding micro-core in the corresponding slice and call it; the micro-core is used to execute the corresponding computer task and generate the corresponding execution result; the main core processing unit is also used to obtain all the execution results.

[0007] In some embodiments, a shared storage unit is also included, which is electrically connected to the main core processing unit and the micro-core array processing unit respectively, and is used for communication between the main core processing unit and each of the slices in the micro-core array processing unit.

[0008] In some embodiments, the communication mode between the main core processing unit and each of the slices in the micro-core array processing unit is a time-sharing management mode.

[0009] In some embodiments, the main core processing unit includes a first window register, and each of the micro-cores includes a second window register. The first window register and the second window register are used for communication between the main core processing unit and the micro-cores, and the overlapping area of ​​any two of the second window registers is used for communication between the corresponding two micro-cores.

[0010] In some embodiments, the status monitoring information table includes the number of each of the slices, the number of each of the micro-cores, the working status of each of the micro-cores, the name of the computer task currently executed by each of the micro-cores, and the activation status of each of the slices.

[0011] In some embodiments, the target program file contains a task name identifier and a corresponding running micro-kernel identifier, and the main core processing unit is used to split the target program file into multiple computer tasks according to the task name identifier. The main core processing unit is also used to update the status monitoring information table according to the task name identifier and the running micro-kernel identifier.

[0012] In some embodiments, the microcore array processing unit includes 256 slices, and each slice includes 7 microcores.

[0013] The present application also provides a working method of a multi-core processor, the working method is executed by the multi-core processor, the multi-core processor includes a main core processing unit and a micro-core array processing unit, the micro-core array processing unit includes multiple independent slices, each of the slices includes at least one micro-core, the working method includes the steps of: loading a target program file, and splitting the target program file into multiple computer tasks; matching corresponding slices for multiple computer tasks according to a status monitoring information table of the micro-core array processing unit, and selecting corresponding micro-cores in the corresponding slices; calling the corresponding micro-core, executing the corresponding computer task, and generating corresponding execution results; obtaining all the execution results.

[0014] In some embodiments, it includes a pre-step of creating the status monitoring information table of the micro-core array processing unit, and recording the status monitoring information table according to the status information of each micro-core in each of the slices.

[0015] In some embodiments, it includes a post-step of monitoring the current status information of each micro-core, and updating the status monitoring information table according to the status information.

[0016] The beneficial effects of the present application are as follows: the present application discloses a multi-core processor and a working method thereof, the multi-core processor comprises a main core processing unit and a micro-core array processing unit, the main core processing unit is electrically connected to the micro-core array processing unit, the micro-core array processing unit comprises a plurality of independent slices, each slice comprises at least one micro-core; the main core processing unit is used to load a target program file and split the target program file into a plurality of computer tasks; the main core processing unit is also used to match corresponding slices for a plurality of computer tasks according to the status monitoring information table of the micro-core array processing unit, select corresponding micro-cores in the corresponding slices, and call them; the micro-cores are used to execute corresponding computer tasks and generate corresponding execution results; the main core processing unit is also used to obtain all execution results. The working method comprises the steps of: loading a target program file and splitting the target program file into a plurality of computer tasks; matching corresponding slices for a plurality of computer tasks according to the status monitoring information table of the micro-core array processing unit, and selecting corresponding micro-cores in the corresponding slices; calling corresponding micro-cores, executing corresponding computer tasks, and generating corresponding execution results; and obtaining all execution results. The present application divides the micro-core array processing unit into multiple fixed-size slices, each of which includes at least one micro-core. The main core processing unit manages each slice and the micro-cores therein respectively, and reasonably allocates computer tasks to them, so that the micro-cores can execute efficiently alone, or work together with other micro-cores in the same slice to complete the tasks; the working efficiency of the multi-core processor is effectively improved, and the overall performance is improved; and the effective management mechanism of the main core processing unit also enables resources to be fully utilized, while avoiding communication conflicts between multiple micro-cores. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the composition of the multi-core processor provided by this application;

[0018] Figure 2 is a schematic diagram of processing computer tasks in a multi-core processor provided by the present application;

[0019] Figure 3 is a schematic diagram of a first window register and a second window register in a multi-core processor provided by the present application;

[0020] Figure 4 It is a flowchart of the working method of the multi-core processor provided by this application. DETAILED DESCRIPTION

[0021] In order to facilitate the understanding of the present application, the present application is described in more detail below in conjunction with the accompanying drawings and specific embodiments. The preferred embodiments of the present application are provided in the accompanying drawings. However, the present application can be implemented in many different forms and is not limited to the embodiments described in this specification. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive.

[0022] It should be noted that, unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as those commonly understood by those skilled in the art of the present application. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The term "and / or" used in this specification includes any and all combinations of one or more related listed items.

[0023] In order to facilitate the understanding of the present application, some terms that may be involved in the embodiments of the present application are first explained as follows:

[0024] Processor array (matrix processor): a square array composed of multiple processors according to certain rules.

[0025] Multi-core processor and many-core processor: refers to the integration of two or more processor cores in one processor, which enables parallel operation of multiple processor cores by dividing tasks among the cores.

[0026] Homogeneous multi-core processors: processors that contain multiple processor cores with the same structure.

[0027] Heterogeneous multi-core processors: processors that contain multiple processor cores with different structures.

[0028] Vector operations: In digital signal processing (DSP), vector operations usually refer to performing some operations (vector addition, vector subtraction, vector multiplication, vector division, etc.) on a set of data, which is regarded as a vector or an array. In digital signal processing (DSP), vector operations are very common and important because they can effectively improve processing speed and reduce code complexity. Modern DSP processors usually have specialized vector processing instructions that can perform these operations more efficiently.

[0029] Microcore: A processor used to perform computer tasks. Each microcore has independent DRAM (dynamic random access memory), SDRAM (synchronous dynamic random access memory), and PRAM (phase change random access memory); it has its own program space and can perform logical operations and multiplication and addition mathematical operations; the data types it operates on are relatively simple.

[0030] like Figure 1 As shown, the present application provides a multi-core processor, including a main core processing unit 1 and a micro-core array processing unit 2, the main core processing unit 1 is electrically connected to the micro-core array processing unit 2, the micro-core array processing unit 2 includes multiple independent slices 21, and each slice 21 includes at least one micro-core 211.

[0031] The main core processing unit 1 is used to load the target program file and split the target program file into multiple computer tasks; the main core processing unit 1 is also used to match the corresponding slices 21 for multiple computer tasks according to the status monitoring information table of the micro-core array processing unit 2, and select the corresponding micro-core 211 in the corresponding slice 21 and call it; the micro-core 211 is used to execute the corresponding computer task and generate the corresponding execution result; the main core processing unit 1 is also used to obtain all the execution results.

[0032] The present application divides the micro-core array processing unit 2 into a plurality of fixed-size slices 21, each of which includes at least one micro-core 211. The main core processing unit 1 manages each slice 21 and the micro-core 211 therein, and reasonably allocates computer tasks thereto, so that the micro-core 211 can execute efficiently alone, or complete tasks in collaboration with other micro-cores 211 in the same slice 21; the working efficiency of the multi-core processor is effectively improved, and the overall performance is enhanced; and the effective management mechanism of the main core processing unit 1 also enables resources to be fully utilized, while avoiding communication conflicts between multiple micro-cores 211.

[0033] Combine the following Figures 1 to 4 , this application is described in detail with specific embodiments.

[0034] Figure 1 The schematic diagram of the composition of the multi-core processor provided in the embodiment of the present application is shown, including a main core processing unit 1, a micro-core array processing unit 2 and a shared storage unit 3. The main core processing unit 1, the micro-core array processing unit 2 and the shared storage unit 3 are respectively connected to each other through a bus 4. The micro-core array processing unit 2 includes a plurality of independent slices 21, and each slice 21 includes at least one micro-core 211.

[0035] The multi-core processor of the present application is a heterogeneous multi-core processor, that is, the main core processing unit 1 and the multiple micro-cores 211 have different structures; the multiple micro-cores 211 can have the same structure or different structures. In this embodiment, the main core processing unit 1 is a single processor that supports both DSP (digital signal processing: digital signal processor) instructions and RISC (Reduced Instruction Set Computer) instructions.

[0036] Furthermore, the present application divides the micro-core array processing unit 2 into a plurality of different slices 21 according to the different peripheral space addresses that the micro-core 211 can access, and each slice 21 contains a preset number of micro-cores 211. The micro-cores 211 in different slices 21 are independent of each other; multiple micro-cores 211 in the same slice 21 can independently complete the computer tasks sent by the main core processing unit 1 (computer tasks: the main core processing unit 1 loads the target program file from the external memory and splits the target program file into multiple computer tasks), and can also cooperate with other micro-cores 211 in the same slice 21 to complete the computer tasks assigned by the main core processing unit 1.

[0037] In this embodiment, the size of the peripheral space that can be managed by a main core processing unit 1 is 4Gbit (gigabit), and the size of the peripheral space that can be managed by a slice 21 is 16Mbit (megabit). Therefore, the main core processing unit 1 can manage a maximum of 256 slices 21, that is, the micro-core array processing unit 2 is divided into a maximum of 256 slices 21, and each slice 21 can have 1 to 7 micro-cores 211. In this embodiment, each slice 21 includes 7 micro-cores 211, so the micro-core array processing unit 2 has a total of 256×7=1792 micro-cores 211.

[0038] Furthermore, before the multi-core processor of the present application starts working, the main core processing unit 1 is first required to monitor the micro-core array processing unit 2 and create a status monitoring information table of the micro-core array processing unit 2, and record the status monitoring information table according to the status information of each micro-core 211 in each slice 21.

[0039] Specifically, the status information of each micro-core 211 includes whether its working status is busy or idle, and the specific name of the computer task executed when it is busy.

[0040] In this embodiment, the status monitoring information table includes the number of each slice 21, the number of each micro-core 211, the working state of each micro-core 211, the name of the computer task currently executed by each micro-core 211, and the activation state of each slice 21. Specifically, the status monitoring information table is shown in Table 1 below:

[0041] Table 1

[0042]

[0043]

[0044] Among them, "slice 1" represents the first slice 21, "core 1-1" represents the first micro-core 211 in the first slice 21, "B" (BUSY) represents that the micro-core 211 is busy, "I" (IDLE) represents that the micro-core 211 is idle, "taskID" represents the specific name of the currently executed computer task, such as the name of a function, "Y" (YES) represents that the activation state of the slice 21 is activated, and "N" (NO) represents that the activation state of the slice 21 is not activated.

[0045] Furthermore, when the multi-core processor of the present application is working, it first loads the target program file from the external memory through the main core processing unit 1, and splits the target program file into multiple computer tasks, wherein each computer task includes multiple instructions.

[0046] Specifically, in this embodiment, in the coding stage (i.e., the stage of writing the target program file, i.e., before loading the target program file), the developer needs to define which computer tasks should be included in the target program file and which micro-core 211 runs each computer task according to the chip manual (i.e., the reference manual, which can provide detailed product usage information, including the internal structure of each functional module, all possible functional descriptions, the use of various working modes and register configuration and other detailed information; when entering the programming design stage, it provides the specific implementation method of each function and the configuration and use of registers), that is, define the task name identifier, the running micro-core identifier and the corresponding relationship between them; and carry information such as the task name identifier, the running micro-core identifier and the corresponding relationship between them in the target program file.

[0047] Furthermore, after loading the target program file, the main core processing unit 1 obtains the task name identifier, the running micro-core identifier and the corresponding relationship therebetween, and updates the status monitoring information table, namely the above Table 1, based on the above information.

[0048] For example, Figure 2As shown, the main core processing unit 1 loads a target program file (named: tracking and control of intermediate frequency target signal) from the external memory, which contains the following information: the main core processing unit 1 runs task 1 (main function), the core 2-1 of the slice 2 runs task 2 (intermediate frequency signal acquisition), the core 2-2 runs task 3 (first FFT calculation), the core 2-3 runs task 4 (first filter algorithm), the core 2-4 runs task 5 (second filter algorithm), the core 2-5 runs task 6 (tracking algorithm), and the core 2-6 runs task 7 (control algorithm). Among them, FFT represents Fast Fourier Transform.

[0049] According to the above information, task 1 (main function) is assigned to the main core processing unit 1 for execution, task 2 (intermediate frequency signal acquisition) is assigned to the core 2-1 of the slice 2 in Table 1 for execution, task 3 (first FFT calculation) is assigned to the core 2-2 of the slice 2 in Table 1 for execution, task 4 (first filter algorithm) is assigned to the core 2-3 of the slice 2 in Table 1 for execution, task 5 (second filter algorithm) is assigned to the core 2-4 of the slice 2 in Table 1 for execution, task 6 (tracking algorithm) is assigned to the core 2-5 of the slice 2 in Table 1 for execution, task 7 (control algorithm) is assigned to the core 2-6 of the slice 2 in Table 1 for execution, and the status monitoring information table, i.e., Table 1, is updated.

[0050] Furthermore, the micro-kernel 211 assigned with the corresponding computer task needs to load the same target program file, and then send a request carrying the running micro-kernel identifier to the main core processing unit 1; the main core processing unit 1 obtains the name of the corresponding computer task according to the above Table 1 and the running micro-kernel identifier, and queries the operating system which instructions correspond to the name of the computer task, and then feeds back these instruction information to the corresponding micro-kernel 211, indicating which instructions the micro-kernel 211 should execute.

[0051] In some other embodiments, during the coding stage, the developer defines which computer tasks should be included in the target program file based on the chip manual, but does not define which micro-kernel 211 runs each computer task. That is, only the task name identifier is defined and the task name identifier is carried in the target program file.

[0052] Furthermore, after the main core processing unit 1 loads the target program file, it obtains the task name identifier it carries, and consults the status monitoring information table. According to the status monitoring information table, it matches corresponding slices 21 for multiple computer tasks, and selects the corresponding micro-core 211 in the corresponding slice 21 to execute the corresponding computer task.

[0053] For example, if the activation state of slice 2 in Table 1 is N (i.e., not activated), and the working states of cores 2-1 to 2-7 are all I (i.e., idle), multiple computer tasks are matched to slice 2, and a suitable micro-core 211 is selected in slice 2 to execute its corresponding computer task. The status monitoring information table is further updated to modify the content corresponding to slice 2 (e.g., the activation state of slice 2 is set to Y (i.e., activated), and the working state of core 2-1 is set to B (i.e., busy), etc.).

[0054] Furthermore, the micro-core 211 assigned with the corresponding computer task needs to load the same target program file and then send a request to the main core processing unit 1; the main core processing unit 1 obtains the name of the corresponding computer task according to the above Table 1 and the request, and queries which instructions correspond to the name of the computer task through the target program file, and then feeds back these instruction information to the corresponding micro-core 211, indicating which instructions the micro-core 211 should execute.

[0055] Furthermore, the micro-core 211 executes the corresponding computer task and generates the corresponding execution result, and the main core processing unit 1 further obtains all the execution results. Figure 2 When the micro-kernel 211 executes task 2, it transmits the execution result to the main core processing unit 1 executing task 1 and the next micro-kernel 211 executing task 3. The micro-kernel 211 executing task 3 transmits the execution result to the main core processing unit 1 executing task 1 and the next micro-kernel 211 executing task 4. And so on. Each micro-kernel 211 transmits its corresponding execution result to the main core processing unit 1 and its corresponding next micro-kernel 211.

[0056] Furthermore, when all computer tasks are completed, the main core processing unit 1 further monitors the current status information of each micro-core 211, and updates the status monitoring information table according to the status information.

[0057] Furthermore, the main core processing unit 1 in the present application communicates with each slice 21 in the micro-core array processing unit 2 through a shared storage unit 3, and the communication mode between the main core processing unit 1 and each slice 21 in the micro-core array processing unit 2 is a time-sharing management mode. Time-sharing management means that the main core processing unit 1 can only communicate with one of the slices 21 at the same time, thereby achieving high efficiency of the data channel.

[0058] In this embodiment, the shared storage unit 3 is a MSG RAM (Message Random Access Memory) with a fixed direction. The MSG RAM is mainly used for data transmission between processors and exists in the form of Memory.

[0059] Specifically, since the main core processing unit 1 can only manage one slice 21 at the same time, the activation state of the slice 21 that is executing the computer task can be set to Y (i.e., activated). Therefore, when the main core processing unit 1 sends data to the shared storage unit 3, only the slice 21 with the activation state of Y is allowed to read data from the shared storage unit 3. In this way, the main core processing unit 1 can only communicate with the selected slice 21 at the same time.

[0060] Further, the main core processing unit 1 includes a first window register, and each micro core 211 includes a second window register. The first window register and the second window register each include a plurality of different registers. The first window register and the second window register are used for communication between the main core processing unit 1 and the micro core 211, and the overlapping area of ​​any two second window registers is used for communication between the corresponding two micro cores 211.

[0061] The communication between the main core processing unit 1 and the multiple micro cores 211 adopts a parallel management mechanism, that is, the main core processing unit 1 can communicate with multiple micro cores 211 in one slice 21 at the same time.

[0062] Combination Figure 3 In this embodiment, the first window register and the second window register each include three different registers. The three different registers of the first window register and the second window register are: outs return result register, ins input parameter register, and locals local register. Figure 3 The w0-w7 in the registers may correspond to a main core processing unit 1 and seven micro cores 211, respectively. There are 24 registers in total, which are used to store temporary data and pointers and participate in arithmetic and logical operations. Among them, CWP (Current window pointer) is the current window register pointer, CWP-1 is the previous window register pointer, and CWP+1 is the next window register pointer. WIM (window invalid mask) is the window invalid mask, which is used to indicate which window is invalid, that is, it cannot be used anymore. Figure 3 As shown, the two registers w0 ins and w1 outs overlap, which means that the corresponding two micro-cores 211 can communicate with each other through their overlapping areas.

[0063] By means of window registers, the main core processing unit 1 can schedule computer tasks on the seven micro-cores 211 in sequence by means of function calls, thereby improving the efficiency of data access between the main core processing unit 1 and the micro-cores 211. In addition, each micro-core 211 can also run in parallel, and data can be transmitted between micro-cores 211, so that multiple micro-cores 211 can collaborate to complete related computer tasks.

[0064] Corresponding to the multi-core processor in the above embodiment, Figure 4 A flowchart of a working method of a multi-core processor provided in an embodiment of the present application is shown. For ease of explanation, only the part related to the embodiment of the present application is shown.

[0065] refer to Figure 1 , Figure 4 , the working method is executed by a multi-core processor, and the working method comprises the steps of:

[0066] S1: creating a status monitoring information table of the micro-core array processing unit 2 , and recording the status monitoring information table according to the status information of each micro-core 211 in each slice 21 .

[0067] S2: Loading a target program file, and splitting the target program file into a plurality of computer tasks.

[0068] S3: According to the status monitoring information table of the micro-core array processing unit 2 , the corresponding slices 21 are matched for the plurality of computer tasks, and the corresponding micro-cores 211 are selected in the corresponding slices 21 .

[0069] S4: calling the corresponding microkernel 211, executing the corresponding computer task, and generating the corresponding execution result.

[0070] S5: Get all execution results.

[0071] S6: Monitor the current status information of each micro-core 211, and update the status monitoring information table according to the status information.

[0072] It should be noted that the working method of the multi-core processor mentioned above is based on the same concept as the multi-core processor of the present application, and specific details can be referred to the description of the embodiment shown in the aforementioned multi-core processor, which will not be repeated here.

[0073] It can be seen that the present application discloses a multi-core processor and a working method thereof, wherein the multi-core processor includes a main core processing unit and a micro-core array processing unit, wherein the main core processing unit is electrically connected to the micro-core array processing unit, and the micro-core array processing unit includes multiple independent slices, each of which includes at least one micro-core; the main core processing unit is used to load a target program file and split the target program file into multiple computer tasks; the main core processing unit is also used to match corresponding slices for multiple computer tasks according to the status monitoring information table of the micro-core array processing unit, and select corresponding micro-cores in the corresponding slices, and call them; the micro-cores are used to execute corresponding computer tasks and generate corresponding execution results; the main core processing unit is also used to obtain all execution results. The working method includes the steps of: loading a target program file and splitting the target program file into multiple computer tasks; matching corresponding slices for multiple computer tasks according to the status monitoring information table of the micro-core array processing unit, and selecting corresponding micro-cores in the corresponding slices; calling corresponding micro-cores, executing corresponding computer tasks, and generating corresponding execution results; and obtaining all execution results. The present application divides the micro-core array processing unit into multiple fixed-size slices, each of which includes at least one micro-core. The main core processing unit manages each slice and the micro-cores therein respectively, and reasonably allocates computer tasks to them, so that the micro-cores can execute efficiently alone, or work together with other micro-cores in the same slice to complete the tasks; the working efficiency of the multi-core processor is effectively improved, and the overall performance is improved; and the effective management mechanism of the main core processing unit also enables resources to be fully utilized, while avoiding communication conflicts between multiple micro-cores.

[0074] The above are merely embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structural transformations made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A multi-core processor, characterized in that: It includes a main core processing unit and a micro-core array processing unit, wherein the main core processing unit is electrically connected to the micro-core array processing unit, and the micro-core array processing unit includes a plurality of independent slices, each of which includes at least one micro-core; The main core processing unit is used to load the target program file and split the target program file into multiple computer tasks; The main core processing unit is further used to match the corresponding slices for the plurality of computer tasks according to the status monitoring information table of the micro-core array processing unit, and select the corresponding micro-core in the corresponding slice and call it; The microkernel is used to execute the corresponding computer task and generate a corresponding execution result; The main core processing unit is also used to obtain all the execution results.

2. The multi-core processor according to claim 1, characterized in that: It also includes a shared storage unit, which is electrically connected to the main core processing unit and the micro-core array processing unit respectively, and is used for communication between the main core processing unit and each of the slices in the micro-core array processing unit.

3. The multi-core processor according to claim 2, characterized in that: The communication mode between the main core processing unit and each of the slices in the micro-core array processing unit is a time-sharing management mode.

4. The multi-core processor according to claim 1, characterized in that: The main core processing unit includes a first window register, and each of the micro cores includes a second window register. The first window register and the second window register are used for communication between the main core processing unit and the micro cores, and the overlapping area of ​​any two of the second window registers is used for communication between the corresponding two micro cores.

5. The multi-core processor according to claim 1, characterized in that: The status monitoring information table includes the number of each of the slices, the number of each of the micro-cores, the working status of each of the micro-cores, the name of the computer task currently executed by each of the micro-cores, and the activation status of each of the slices.

6. The multi-core processor according to claim 1, characterized in that: The target program file contains a task name identifier and a corresponding running micro-core identifier. The main core processing unit is used to split the target program file into multiple computer tasks according to the task name identifier. The main core processing unit is also used to update the status monitoring information table according to the task name identifier and the running micro-core identifier.

7. The multi-core processor according to claim 1, characterized in that: The microcore array processing unit includes 256 slices, and each slice includes 7 microcores.

8. A method for operating a multi-core processor, characterized in that: The working method is executed by the multi-core processor, the multi-core processor includes a main core processing unit and a micro-core array processing unit, the micro-core array processing unit includes a plurality of mutually independent slices, each of the slices includes at least one micro-core, and the working method includes the steps of: Loading a target program file, and splitting the target program file into a plurality of computer tasks; According to the status monitoring information table of the micro-core array processing unit, the corresponding slices are matched for a plurality of the computer tasks, and the corresponding micro-cores are selected in the corresponding slices; Calling the corresponding microkernel, executing the corresponding computer task, and generating a corresponding execution result; Get all the execution results.

9. The working method according to claim 8, characterized in that: It includes the following pre-steps: The state monitoring information table of the microcore array processing unit is created, and the state monitoring information table is recorded according to the state information of each microcore in each of the slices.

10. The working method according to claim 8, characterized in that: It includes the following post-processing steps: Monitor the current status information of each micro-kernel, and update the status monitoring information table according to the status information.