Chip auxiliary development method and device, storage medium and equipment
By obtaining chip computing functions and parameter templates, simulate operation and filtering performance results, the chip design parameters are automatically obtained, which solves the problem of long chip design cycle and improves design efficiency and versatility.
Patent Information
- Application Number
- CN202410956388.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-26
AI Technical Summary
The chip design cycle is long, and due to the component power consumption and area factors, it is difficult to balance computing power, bandwidth and functional modules under fixed conditions, resulting in high design difficulty and low efficiency.
By obtaining the calculation functions set by the chip and the corresponding calculation function parameter templates, simulate the operation logic to obtain performance results, and filter the configuration parameter combination using preset performance evaluation standards to automatically obtain the final design parameter template, shorten the design cycle and improve efficiency.
It realizes the rapid acquisition of final design parameters templates, reduces manual workload, shortens chip design cycle, improves design efficiency, and adapts to different application scenarios to enhance versatility.
Smart Images

Figure CN119026540A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of chip development, and in particular to a chip-assisted development method, device, storage medium and equipment. Background Art
[0002] In recent years, artificial intelligence technology has been widely used in various fields, such as face recognition, target detection, autonomous driving, etc., and is closely related to people's lives.
[0003] With the further development of technology, the requirements of artificial intelligence technology for computing power are also further improved. Therefore, chips of different types, different computing power, and different application scenarios are also emerging, such as: Graphics Processing Unit (GPU), Neural Network Processing Unit (NPU), Tensor Processing Unit (TPU), etc.
[0004] However, the application scenarios of chips are varied, the functions are complex, and the design is difficult. When designing chips, the performance of chips can only be designed under the conditions of fixed area and limited power consumption, such as computing power, bandwidth, and functional modules, due to the influence of factors such as power consumption and area of components. In order to balance these factors and design the most reasonable chip parameters, chip designers need to spend a lot of manpower and material resources on simulation experiments, which leads to a long chip design cycle. Summary of the invention
[0005] The problem to be solved by the present invention is to shorten the chip design cycle.
[0006] To solve the above problems, an embodiment of the present invention provides a chip-assisted development method, the method comprising:
[0007] Get the calculation function set by the chip;
[0008] Acquire a computing function parameter template of the chip, wherein the computing function parameter template includes: configuration parameters related to executing the computing function and value information corresponding to each configuration parameter;
[0009] Simulating the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result;
[0010] The chip simulation performance results are screened using the preset chip performance evaluation standards to obtain a configuration parameter combination that meets the chip performance evaluation standards as a final design parameter template for the chip.
[0011] In a possible embodiment, the method further includes:
[0012] A chip performance evaluation standard is obtained, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
[0013] In a possible embodiment, the obtaining the computing function parameter template corresponding to the computing function includes:
[0014] More than two first computing function parameter templates corresponding to the computing function are obtained, each of the first computing function templates including: configuration parameter identification information related to executing the computing function and unique value information corresponding to each configuration parameter.
[0015] In a possible embodiment, simulating the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result includes:
[0016] Each of the first computing function parameter templates is traversed, and the operation logic of the computing function on each of the first computing function parameter templates is simulated to obtain simulation operation performance results respectively corresponding to each of the first computing function parameter templates.
[0017] In a possible embodiment, the obtaining a computing function parameter template corresponding to the computing function includes:
[0018] A second computing function parameter template corresponding to the computing function is obtained, where the second computing function template includes: configuration parameter identification information related to executing the computing function and value range information corresponding to each configuration parameter.
[0019] In a possible embodiment, simulating the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result includes:
[0020] The value ranges of each configuration parameter on the second computing function parameter template are traversed to obtain the simulation running performance results corresponding to each configuration parameter combination on the second computing function parameter template.
[0021] In a possible embodiment, simulating the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result includes:
[0022] A binary search algorithm is used to search for the values of each configuration parameter on the second computing function parameter template to obtain the chip simulation running performance result corresponding to each search result.
[0023] In a possible embodiment, the chip performance evaluation criteria include at least one of the following:
[0024] The value range of the chip MAC computing power utilization rate;
[0025] The value range of the total power consumption of the chip.
[0026] In a possible embodiment, the configuration parameters related to executing the computing function include:
[0027] The chip's DRAM size, SRAM size, DRAM bandwidth, SRAM bandwidth, MAC computing power, DRAM power consumption, SRAM power consumption, and MAC computing unit power consumption.
[0028] The embodiment of the present invention further provides a chip-assisted development device, the device comprising:
[0029] A first acquisition unit, adapted to acquire a computing function set by the chip;
[0030] A second acquisition unit is adapted to acquire a computing function parameter template corresponding to the computing function, wherein the computing function parameter template includes: configuration parameters related to executing the computing function and value information corresponding to each configuration parameter;
[0031] A simulation unit, adapted to simulate the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result;
[0032] The screening unit is adapted to screen the chip simulation operation performance results using a preset chip performance evaluation standard to obtain a configuration parameter combination that meets the chip performance evaluation standard as a final design parameter template for the computing function.
[0033] In a possible embodiment, the device further includes:
[0034] The third acquisition unit is adapted to acquire a chip performance evaluation standard, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
[0035] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. The computer program is executed by a processor to implement the steps of any of the above methods.
[0036] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the steps of any of the above methods when running the computer program.
[0037] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:
[0038] By applying the scheme of the present invention, the chip simulation operation performance results are obtained by obtaining the calculation function set by the chip and the corresponding calculation function parameter template, and simulating the operation logic of the calculation function on the calculation function parameter template, so as to screen the chip simulation operation performance results by using the preset chip performance evaluation standard, obtain the configuration parameter combination that meets the chip performance evaluation standard, and finally obtain the final design parameter template of the calculation function set by the chip, without the need to manually obtain the final design parameter template through a large number of experiments, so that the final design parameter template can be obtained quickly, which not only reduces the manual workload, shortens the chip design cycle, and improves the chip design efficiency, but also reduces the difficulty of obtaining the final design parameter template. In addition, by adopting the scheme of the present invention, even if the calculation function and the calculation function parameter template set by the chip change, the final design parameter template can be obtained, thereby dynamically adapting to different application scenarios, and having better versatility. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 is a flow chart of a chip-assisted development method in an embodiment of the present invention;
[0040] Figure 2 is a flow chart of another chip-assisted development method in an embodiment of the present invention;
[0041] Figure 3 is a flow chart of another chip-assisted development method in an embodiment of the present invention;
[0042] Figure 4 is a flow chart of another chip-assisted development method in an embodiment of the present invention;
[0043] Figure 5 It is a structural schematic diagram of a chip-assisted development device in an embodiment of the present invention. DETAILED DESCRIPTION
[0044] In actual applications, due to the changing application scenarios, complex functions and high design difficulty of chips, when designing chips, the performance of chips can only be designed under the conditions of fixed area and limited power consumption, such as computing power, bandwidth and functional modules, due to the influence of factors such as power consumption and area of components. In order to balance these factors and design the most reasonable chip parameters, chip designers need to spend a lot of manpower and material resources on simulation experiments, which leads to a long chip design cycle.
[0045] In view of this problem, the present invention provides a chip-assisted development method, which can automatically obtain the final design parameter template based on the calculation function set by the chip and the corresponding calculation function parameter template, thereby shortening the chip design cycle, improving the chip design efficiency, and reducing the chip design difficulty. In addition, the method can be applied to different application scenarios, meet the chip design requirements of different calculation functions, and has better versatility.
[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0047] Reference Figure 1 The embodiment of the present invention provides a chip-assisted development method, which may include the following steps:
[0048] Step 11, obtaining the computing function set by the chip.
[0049] In a specific implementation, the computing functions obtained refer to all computing functions set for the chip, that is, all computing functions that the chip can support.
[0050] In an embodiment of the present invention, the computing function refers to a logic unit in the chip that implements any calculation, operation or algorithm and the maximum scenario parameters required to run the logic unit. The logic unit can be implemented by a single operator or by multiple operators, and multiple operators can be combined through predetermined or dynamically specified logical relationships. For example, a computing function of a chip can be a residual network (ResNet) or an LLaMA (Large Language Model MetaAI) network, which can specifically include the model structure itself and the parameter configuration file of the model.
[0051] In the embodiment of the present invention, the operator refers to a unit in the chip that implements a basic operation algorithm or function, which can be implemented by any existing or future developed software or hardware technology. Various basic operations in the chip can be implemented by the operator, for example, convolution, various numerical operations, various vector operations, various matrix operations, various character operations, etc.
[0052] Step 12: Obtain a computing function parameter template corresponding to the chip, wherein the computing function parameter template includes: configuration parameter identification information related to executing the computing function and value information corresponding to each configuration parameter.
[0053] In a specific implementation, the computing function parameter template refers to the configuration parameters required for designing various computing functions on the chip, which may specifically include the configuration parameters and specific values of each component used to perform calculations on the chip. Among them, the components used to perform calculations on the chip may include dynamic random access memory (DRAM), static random access memory (SRAM), multiply accumulate unit (MAC), etc. Correspondingly, the configuration parameters of DRAM and SRAM may include the size, bandwidth, power consumption, etc. of DRAM and SRAM. The configuration parameters of the MAC computing unit may include: the computing power and power consumption of the MAC computing unit.
[0054] For example, in one embodiment, the configuration parameter values corresponding to each computing function on the chip can be: DRAM size is 4GB to 16GB, granularity is 1GB, SRAM size is 64KB-1024KB, MAC computing unit size is 4096 or 9128, and DRAM bandwidth is 10GB / s to 40GB / s.
[0055] In a specific implementation, only one computing function parameter template may be obtained, but each configuration parameter on the computing function parameter template may have multiple values. Alternatively, multiple computing function parameter templates may be obtained, but the configuration parameter values on each computing function parameter template are fixed.
[0056] Step 13, simulating the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result.
[0057] In a specific implementation, the so-called simulation of the operating logic of the computing function on the computing function parameter template is to configure the computing function with the information on the computing function parameter template, run the logic corresponding to each computing function, simulate the actual operation of the chip under the computing function parameter template, and obtain the chip simulation operation performance result.
[0058] In a specific implementation, the same computing function parameter template may obtain only one chip simulation operation performance result, or may obtain multiple chip simulation operation performance results, which is specifically related to the values of each configuration parameter on the computing function parameter template.
[0059] Step 14, using a preset chip performance evaluation standard, screening the chip simulation operation performance results, and obtaining a configuration parameter combination that meets the chip performance evaluation standard as a final design parameter template for the chip.
[0060] In a specific implementation, the chip performance evaluation standard can be obtained in advance. After obtaining all chip simulation operation performance results, the chip simulation operation performance results can be screened using the chip performance evaluation standard to obtain chip simulation operation performance results that meet the chip performance evaluation standard, thereby combining the configuration parameters corresponding to the chip simulation operation performance results that meet the chip performance evaluation standard as the final design parameter template for the chip.
[0061] In one embodiment of the present invention, the chip performance evaluation standard may include: the value interval of the best chip performance parameter, that is, the value interval of the expected chip performance. For example, the value interval of the best chip performance parameter may be the value interval of the chip MAC computing power utilization, or the value interval of the chip total power consumption, or the value interval of the chip bandwidth utilization, or the value interval of the DRAM utilization on the chip, or may simultaneously include two or more of the value intervals of the chip MAC computing power utilization, the value interval of the chip total power consumption, the value interval of the chip bandwidth utilization, and the value interval of the DRAM utilization. Among them, when the value interval of the best chip performance parameter is the value interval of the chip MAC computing power utilization, it indicates the value interval of the MAC computing power utilization that the chip is expected to achieve after the design is completed. When the value interval of the best chip performance parameter is the value interval of the chip total power consumption, it indicates the value interval of the total power consumption that the chip is expected to achieve after the design is completed. When the value range of the optimal chip performance parameter is the value range of the chip bandwidth utilization, it is expected that the chip can achieve the value range of the chip bandwidth utilization after the design is completed.
[0062] For example, in one embodiment, the chip performance evaluation criteria may include: MAC computing power utilization>80%, total chip power consumption less than <10W, bandwidth utilization>70%, and DRAM utilization>80%.
[0063] In the specific implementation, the chip performance evaluation standard is used to screen the chip simulation operation performance results, that is, the chip simulation operation performance results that fall within the value interval of the best chip performance parameters are screened from the chip simulation operation performance results, and the configuration parameter combination corresponding to the chip simulation operation performance results that fall within the value interval of the best chip performance parameters is used as the final design parameter template of the chip.
[0064] Figure 2 FIG. 1 is a flow chart of another chip-assisted development method in an embodiment of the present invention. Figure 2 , the method may include:
[0065] Step 21, obtaining the computing function set by the chip.
[0066] Regarding step 21, please refer to the above description of step 11 for details, which will not be repeated here.
[0067] Step 22: Obtain two or more first computing function parameter templates corresponding to the computing function.
[0068] Each of the first computing function templates includes: configuration parameters related to the execution of the computing function and unique value information corresponding to each configuration parameter. That is, on the first computing function parameter template, each configuration parameter has a value. On each first computing function parameter template, the value of the same configuration parameter may be the same or different. However, on any two first computing function parameter templates, at least some of the configuration parameters have different values.
[0069] Step 23, obtaining a chip performance evaluation standard, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
[0070] In a specific implementation, step 23 may be performed before step 25. For example, step 23 may be performed after step 22, step 23 may be performed after step 24, or step 23 may be performed after step 21. Figure 2 This merely provides an embodiment, and does not limit the execution order of step 23.
[0071] Step 24, traverse each of the first computing function parameter templates, simulate the operation logic of the computing function on each of the first computing function parameter templates, and obtain the simulation operation performance results corresponding to each of the first computing function parameter templates.
[0072] After obtaining multiple first computing function parameter templates, simulate the running logic of the computing function on each first computing function parameter template, that is, configure the computing function with each first computing function parameter template respectively, run the logic corresponding to each computing function, so as to obtain the chip simulation running performance results corresponding to each first computing function parameter template.
[0073] Step 25, using the preset chip performance evaluation standard, the chip simulation operation performance results are screened to obtain a configuration parameter combination that meets the chip performance evaluation standard as the final design parameter template of the chip.
[0074] Specifically, the chip performance evaluation standard can be used to screen the simulation performance results of each first computing function parameter template chip, and the first computing function parameter template corresponding to the chip simulation performance result falling within the optimal performance parameter value range is used as the final design parameter template of the chip.
[0075] Figure 3FIG. 1 is a flowchart of another chip-assisted development method in an embodiment of the present invention. Figure 3 , the method may include:
[0076] Step 31, obtaining the computing function set by the chip.
[0077] Regarding step 31, please refer to the above description about step 11 for details, which will not be repeated here.
[0078] Step 32: Obtain a second computing function parameter template corresponding to the computing function.
[0079] The second computing function template includes: configuration parameters related to the execution of the computing function and value range information corresponding to each configuration parameter. That is to say, in this embodiment, only one computing function parameter template is obtained, namely, the second function parameter template, and each configuration parameter has a corresponding value range in the second function parameter template, such as the value range of DRAM size, the value range of SRAM size, the value range of DRAM bandwidth, the value range of SRAM bandwidth, the value range of MAC computing unit power consumption, etc.
[0080] Step 33, obtaining a chip performance evaluation standard, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
[0081] In a specific implementation, step 33 may be performed before step 35. For example, step 33 may be performed after step 32, step 33 may be performed after step 24, or step 33 may be performed after step 31. Figure 2 This merely provides an embodiment, and does not limit the execution order of step 33.
[0082] Step 34, traverse the value ranges of each configuration parameter on the second computing function parameter template to obtain the simulation running performance results corresponding to each configuration parameter combination on the second computing function parameter template.
[0083] After obtaining the second computing function parameter template, since each configuration parameter has a corresponding value range, a value can be selected from the value range of each configuration parameter for combination to obtain all combinations of values of each configuration parameter, and the computing function is set with each configuration parameter value combination, and the logic corresponding to each computing function is run, thereby obtaining the chip simulation running performance results corresponding to each configuration parameter value combination.
[0084] Step 35, using the preset chip performance evaluation standard, the chip simulation operation performance results are screened to obtain a configuration parameter combination that meets the chip performance evaluation standard as the final design parameter template of the chip.
[0085] Specifically, the chip performance evaluation standard can be used to screen the chip simulation performance results corresponding to each configuration parameter value combination, and the configuration parameter value combination corresponding to the chip simulation performance result falling within the best performance parameter value range is used as the final design parameter template of the chip.
[0086] Figure 4 FIG. 1 is a flow chart of a chip-assisted development method in another embodiment of the present invention. Figure 4 , the method may include the following steps:
[0087] Step 41, obtaining the computing function set by the chip.
[0088] In a specific implementation, a user may input a computing function set by a chip through a graphical user interface (GUI).
[0089] Step 42: Obtain a second computing function parameter template corresponding to the computing function.
[0090] In a specific implementation, the user may input the second calculation function parameter template through a graphical user interface (GUI).
[0091] Step 43, obtaining a chip performance evaluation standard, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
[0092] In a specific implementation, a user may input chip performance evaluation criteria through a graphical user interface (GUI).
[0093] Step 44, using a binary search algorithm, searches for the values of each configuration parameter on the second computing function parameter template to obtain the chip simulation running performance result corresponding to each search result.
[0094] In a specific implementation, the values of each configuration parameter on the second computing function parameter template can be initialized according to the input task requirements, and a binary search algorithm can be used to search for the values of each configuration parameter on the second computing function parameter template to improve the search rate of the final design parameter template, thereby improving chip development efficiency.
[0095] Step 45, using the preset chip performance evaluation standard, the chip simulation operation performance results are screened to obtain a configuration parameter combination that meets the chip performance evaluation standard as the final design parameter template of the chip.
[0096] Based on each search result, the operating logic of the chip computing function is simulated respectively, so that the chip simulation operation performance results corresponding to each search result can be obtained. Subsequently, the configuration parameter combination corresponding to the chip simulation operation performance results that meet the chip performance evaluation standards is selected from the chip simulation operation performance results as the final design parameter template of the chip.
[0097] For example, in the second computing function parameter template, the setting range of DRAM size is [4GB, 16GB]. When searching with the binary search algorithm, the initially selected DRAM size is 10GB. Then, according to the logic corresponding to the computing function running with DRAM size of 10GB, the chip simulation running performance results corresponding to DRAM size of 10GB are obtained. Among them, for the chip simulation running performance results corresponding to DRAM size of 10GB, if the DRAM utilization rate is low, it can be judged that the parameter value of the initial DRAM setting is large, and it needs to be adjusted downward, and the parameter value with the best DRAM utilization rate is continued to be found in the DRAM size range of [4GB, 10GB].
[0098] By using a binary search algorithm to search for parameter values, it is possible to find chip parameter configurations that meet the requirements of deep learning inference tasks in a shorter time, thereby improving chip development efficiency.
[0099] In some embodiments, the chip-assisted development method in the embodiments of the present invention can be used to output the chip simulation operation performance results corresponding to the optimal configuration parameter combination on the second computing function parameter template, including DRAM size, bandwidth, power consumption, etc., and provide a detailed evaluation report to explain how each parameter meets the task requirements and chip performance evaluation standards.
[0100] It should be noted that in the embodiment of the present invention, the computing function supported by the chip can be implemented using any deep learning model, including but not limited to computer vision models ResNet50, YOLO-3, and natural language large model LLaMA-3-8B, etc.
[0101] From the above content, it can be seen that the chip-assisted development method in the embodiment of the present invention can automatically search for the optimal parameter template according to the chip function and configuration parameter settings, thereby greatly simplifying the difficulty of manually obtaining the final design parameter template, and can shorten the chip design cycle, improve chip design efficiency, and have better versatility.
[0102] In order to enable those skilled in the art to better understand and implement the present invention, the device, test system, electronic device and computer-readable storage medium corresponding to the above method are described in detail below.
[0103] Reference Figure 5The embodiment of the present invention further provides a chip-assisted development device 50, which may include: a first acquisition unit 51, a second acquisition unit 52, a simulation unit 53, and a screening unit 54. Among them:
[0104] The first acquisition unit 51 is adapted to acquire a computing function set by the chip;
[0105] The second acquisition unit 52 is adapted to acquire a computing function parameter template corresponding to the computing function, wherein the computing function parameter template includes: configuration parameter identification information related to executing the computing function and value information corresponding to each configuration parameter;
[0106] The simulation unit 53 is adapted to simulate the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result;
[0107] The screening unit 54 is adapted to screen the chip simulation operation performance results using a preset chip performance evaluation standard to obtain a configuration parameter combination that meets the chip performance evaluation standard as a final design parameter template for the computing function.
[0108] In one embodiment, the apparatus 50 may further include: a third acquisition unit 55. The third acquisition unit 55 is adapted to acquire a chip performance evaluation standard, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
[0109] For details about each functional unit in the device 50, please refer to the above Figures 1 to 4 The corresponding steps are described in the implementation, which will not be repeated here.
[0110] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. The computer program is executed by a processor to implement the steps of any of the above methods.
[0111] In a specific implementation, the computer-readable storage medium may include: ROM, RAM, magnetic disk or optical disk, etc.
[0112] An embodiment of the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the steps of any of the above methods when running the computer program.
[0113] Regarding the various modules / units included in the various devices and products described in the above embodiments, they can be software modules / units, or hardware modules / units, or they can be partially software modules / units and partially hardware modules / units. For example, for various devices and products applied to or integrated in a chip, the various modules / units included therein can all be implemented in the form of hardware such as circuits, or at least some of the modules / units can be implemented in the form of software programs, which run on a processor integrated inside the chip, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits; for various devices and products applied to or integrated in a chip module, the various modules / units included therein can all be implemented in the form of hardware such as circuits, and different modules / units can be located in the same component (such as a chip, circuit module, etc.) or different components of the chip module, or at least some of the modules / units can be implemented in the form of hardware such as circuits. The element can be implemented in the form of a software program, which runs on a processor integrated inside the chip module, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits; for various devices and products applied to or integrated in the terminal, the various modules / units contained therein can be implemented in the form of hardware such as circuits, and different modules / units can be located in the same component (for example, chip, circuit module, etc.) or in different components in the terminal, or, at least some modules / units can be implemented in the form of a software program, which runs on a processor integrated inside the terminal, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits.
[0114] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.
Claims
1. A chip-assisted development method, characterized in that: include: Get the calculation function set by the chip; Acquire a computing function parameter template of the chip, wherein the computing function parameter template includes: configuration parameter identification information related to executing the computing function and value information corresponding to each configuration parameter; simulate the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result; The chip simulation performance results are screened using the preset chip performance evaluation standards to obtain a configuration parameter combination that meets the chip performance evaluation standards as a final design parameter template for the chip.
2. The chip-assisted development method according to claim 1, characterized in that: Also includes: A chip performance evaluation standard is obtained, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
3. The chip-assisted development method according to claim 1, characterized in that: The obtaining of the calculation function parameter template corresponding to the calculation function includes: More than two first computing function parameter templates corresponding to the computing function are obtained, each of the first computing function templates including: configuration parameter identification information related to executing the computing function and unique value information corresponding to each configuration parameter.
4. The chip-assisted development method according to claim 3, characterized in that: The simulating the operation logic of the computing function on the computing function parameter template to obtain the chip simulation operation performance result includes: Each of the first computing function parameter templates is traversed, and the operation logic of the computing function on each of the first computing function parameter templates is simulated to obtain simulation operation performance results respectively corresponding to each of the first computing function parameter templates.
5. The chip-assisted development method according to claim 1, characterized in that: The obtaining of the calculation function parameter template corresponding to the calculation function includes: A second computing function parameter template corresponding to the computing function is obtained, where the second computing function template includes: configuration parameter identification information related to executing the computing function and value range information corresponding to each configuration parameter.
6. The chip-assisted development method according to claim 5, characterized in that: The simulating the operation logic of the computing function on the computing function parameter template to obtain the chip simulation operation performance result includes: The value ranges of each configuration parameter on the second computing function parameter template are traversed to obtain the simulation running performance results corresponding to each configuration parameter combination on the second computing function parameter template.
7. The chip-assisted development method according to claim 5, characterized in that: The simulating the operation logic of the computing function on the computing function parameter template to obtain the chip simulation operation performance result includes: A binary search algorithm is used to search for the values of each configuration parameter on the second computing function parameter template to obtain the chip simulation running performance result corresponding to each search result.
8. The chip-assisted development method according to claim 1, characterized in that: The chip performance evaluation criteria include at least one of the following: The value range of the chip MAC computing power utilization rate; a value range of the total power consumption of the chip; The chip bandwidth utilization rate range.
9. The chip-assisted development method according to claim 1, characterized in that: The configuration parameters related to executing the computing function include: The chip's DRAM size, SRAM size, DRAM bandwidth, SRAM bandwidth, MAC computing power, DRAM power consumption, SRAM power consumption, and MAC computing unit power consumption.
10. A chip-assisted development device, characterized in that: include: A first acquisition unit, adapted to acquire a computing function set by the chip; A second acquisition unit is adapted to acquire a computing function parameter template corresponding to the computing function, wherein the computing function parameter template includes: configuration parameter identification information related to executing the computing function and value information corresponding to each configuration parameter; A simulation unit, adapted to simulate the operation logic of the computing function on the computing function parameter template to obtain a chip simulation operation performance result; The screening unit is adapted to screen the chip simulation operation performance results using a preset chip performance evaluation standard to obtain a configuration parameter combination that meets the chip performance evaluation standard as a final design parameter template for the computing function.
11. The chip-assisted development device according to claim 10, characterized in that: Also includes: The third acquisition unit is adapted to acquire a chip performance evaluation standard, wherein the chip performance evaluation standard includes: a value range of an optimal chip performance parameter.
12. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the steps of the method according to any one of claims 1 to 9.
13. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor runs the computer program, the steps of the method according to any one of claims 1 to 9 are performed.
