Performance evaluation method and device for vehicle-mounted chip, medium and program product

Through multi-angle and comprehensive vehicle chip performance evaluation methods, combined with static and dynamic performance evaluation, the one-sided problems of existing evaluation methods are solved, and a more accurate and comprehensive evaluation of vehicle chips is achieved, ensuring the performance of intelligent driving systems and Security.

CN118937949APending Publication Date: 2024-11-12VOLKSWAGEN AG
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202310531427.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-11
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing vehicle-mounted chip performance evaluation methods are one-sided, and they cannot fully understand the static and dynamic performance of the chip, resulting in difficulty in selecting and risk of functional degradation.

Method used

A multi-angle and comprehensive performance evaluation method is adopted to determine the static and dynamic performance scores of the on-board chip through static and dynamic metric weights and scores, and comprehensively evaluate its comprehensive performance scores to provide a standardized performance testing system.

Benefits of technology

It improves the accuracy and comprehensiveness of the performance evaluation of on-board chips, helps development engineers to select suitable chips and reduces the risk of functional degradation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118937949A_ABST
    Figure CN118937949A_ABST
Patent Text Reader

Abstract

The embodiment of the invention relates to a performance evaluation method and device for a vehicle-mounted chip, a medium and a program product. The method includes determining a static performance score for the on-board chip based on a plurality of static metric weights and a plurality of static metric scores for a plurality of static metrics for the on-board chip, where the plurality of static metrics indicate hardware performance of the on-board chip in a non-operating state. The method further includes determining a dynamic performance score for the on-board chip based on the plurality of dynamic metric weights and the plurality of dynamic metric scores for the plurality of dynamic metrics for the on-board chip, where the plurality of dynamic metrics indicate operational performance of the on-board chip in the operational state. And determining a comprehensive performance score of the vehicle-mounted chip based on the static performance score and the dynamic performance score. In this way, standardized performance testing of the vehicle-mounted chip can be achieved, and the accuracy of performance evaluation of the vehicle-mounted chip can be improved by comprehensively considering the static performance and the dynamic performance of the vehicle-mounted chip.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] Embodiments of the present disclosure generally relate to the field of chips, and more specifically, to a performance evaluation method, device, medium, and program product for an in-vehicle chip. Background Art

[0002] With the rapid development of chip technology, its application areas are constantly expanding, such as communications, medical care, aerospace, smart home, etc. Combining chip technology with automotive products is a promising research direction, which can improve the intelligence, safety, performance and comfort of automotive products, and thus promote the development of the automotive industry.

[0003] Intelligent driving is a representative driving application achieved with the help of high-performance chips. Intelligent driving is a technology that uses artificial intelligence technology and sensors to upgrade cars to intelligent levels, enabling them to automatically perceive the surrounding environment, process information, and make corresponding decisions and actions. By using computer vision, machine learning, deep learning, speech recognition and other technologies, intelligent driving can achieve functions such as vehicle autonomous navigation, obstacle detection, traffic signal recognition, automatic parking, collision avoidance, etc., thereby improving driving safety, reducing traffic accident rates, and reducing traffic congestion.

[0004] Intelligent driving requires a lot of data processing and computing power, and vehicle chips are one of the core components necessary to achieve intelligent driving tasks. With the continuous advancement of technology, vehicle chips are also developing rapidly, and more and more manufacturers have launched dedicated chips for intelligent driving. By continuously improving the performance and functions of vehicle chips, intelligent driving will have a wider application prospect. Summary of the invention

[0005] The embodiments of the present disclosure provide a performance evaluation solution for an on-board chip, which can perform performance evaluation on the on-board chip from multiple angles and in all directions, and obtain an accurate understanding of the on-board chip.

[0006] In the first aspect of the present disclosure, a performance evaluation method for an on-board chip is provided. The performance evaluation method includes determining the static performance score of the on-board chip based on multiple static metric weights and multiple static metric scores of multiple static metrics for the on-board chip, wherein the multiple static metrics indicate the hardware performance of the on-board chip in a non-operating state. The performance evaluation method also includes determining the dynamic performance score of the on-board chip based on multiple dynamic metric weights and multiple dynamic metric scores of multiple dynamic metrics for the on-board chip, wherein the multiple dynamic metrics indicate the operating performance of the on-board chip in an operating state. The performance evaluation method also includes determining the comprehensive performance score of the on-board chip based on the static performance score and the dynamic performance score. According to the performance evaluation method for on-board chips of an embodiment of the present disclosure, standardized performance testing for on-board chips is realized, so that different on-board chips can be tested and evaluated through a unified evaluation system. In addition, by comprehensively considering the static and dynamic performance performance of the on-board chip, a more comprehensive understanding of the on-board chip can be obtained, and the accuracy of the performance evaluation of the on-board chip is improved.

[0007] In a second aspect of the present disclosure, a performance evaluation device for an on-board chip is provided, the performance evaluation device comprising a processor, and a memory, the memory being coupled to the processor and storing instructions, which, when executed by the processor, cause the device to perform a performance evaluation method according to the first aspect of the present disclosure.

[0008] In a third aspect of the present disclosure, a computer-readable storage medium is provided, on which computer-executable instructions are stored. When the computer-executable instructions are executed, the computer executes the performance evaluation method according to the first aspect of the present disclosure.

[0009] According to a fourth aspect of the present disclosure, a computer program product is provided, which is tangibly stored on a non-volatile computer-readable medium and includes machine executable instructions, which, when executed, cause a machine to perform the performance evaluation method according to the first aspect of the present disclosure.

[0010] Please note that the present summary is provided to introduce some concepts in a simplified form, which will be further described in the following detailed description. The present summary is not intended to identify key features or essential features of the present disclosure, nor is it intended to limit the scope of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The above and other objects, features and advantages of the present disclosure will become more apparent through a more detailed description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, in which:

[0012] Figure 1A schematic diagram schematically illustrates an example environment in which methods and / or processes according to embodiments of the present disclosure may be implemented;

[0013] Figure 2 A flow chart of a performance evaluation method for a vehicle-mounted chip according to an embodiment of the present disclosure is schematically illustrated;

[0014] Figure 3 A schematic diagram of a comprehensive performance evaluation model for an on-vehicle chip according to an embodiment of the present disclosure is schematically illustrated;

[0015] Figure 4 An example process of determining static metrics for an on-board chip according to an embodiment of the present disclosure is schematically illustrated;

[0016] Figure 5 Schematically illustrates examples of multiple scenarios adapted by the vehicle-mounted chip;

[0017] Figure 6 An example process of determining a static performance score of a static metric for an on-board chip according to an embodiment of the present disclosure is schematically illustrated;

[0018] Figure 7 An example process of determining a dynamic performance score of a dynamic metric for an on-board chip according to an embodiment of the present disclosure is schematically illustrated;

[0019] Figure 8 The example process of load testing for a vehicle-mounted chip according to an embodiment of the present disclosure is schematically illustrated;

[0020] Fig. 9 A schematic block diagram of an example device that may be used to implement an embodiment according to the present disclosure is schematically illustrated.

[0021] Throughout the drawings, the same or similar reference numerals refer to the same or similar elements. Specific embodiments

[0022] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0023] In the description of the embodiments of the present disclosure, the term "including" and its variations should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. can refer to different or the same objects, unless explicitly indicated to be different.

[0024] As mentioned above, intelligent driving refers to the use of advanced technologies, sensors, etc. to automatically control and navigate vehicles during driving. With the rapid development of artificial intelligence and automation technology, intelligent driving is regarded as the trend of future driving. Intelligent driving systems can reduce the risk of traffic accidents, make more efficient use of road resources, and reduce the burden on drivers, thereby providing a safer, more efficient and convenient driving experience. Intelligent driving is a representative driving application that relies on the support of high-performance chips. With the rapid development of chip technology, high-performance chips are widely used in various fields and bring convenience to our lives.

[0025] In order to achieve intelligent driving, a large amount of data processing and computing power is required, because the massive data collected from various sensors such as cameras needs to be processed and analyzed in order to make accurate decisions and perform complex operations in real time. On-board chips are one of the core components necessary to achieve intelligent driving tasks. They can provide high-performance computing, support artificial intelligence algorithms such as machine learning and deep learning, and ensure multi-task collaborative processing. With the continuous development of smart cars and autonomous driving technology, the demand for on-board chips is also increasing. More and more automakers are beginning to adopt such dedicated chips to provide high-performance intelligent driving experience.

[0026] In the design stage of automotive products, it is particularly important to choose the right automotive chip. However, the selection of automotive chips is a complex process that requires consideration of many factors and parameters, and combined with the actual application scenarios, in order to select the most suitable automotive chip. Generally, it is necessary to consider the functional requirements and internal and external environment of automotive products, the performance and safety of automotive chips, and even the cost and supply chain of automotive chips. For example, for automotive products that focus on in-vehicle entertainment, it is particularly necessary to consider the image processing and audio processing capabilities of the automotive chip, while for automotive products that focus on safe driving assistance, it is particularly necessary to consider the data processing capabilities of the automotive chip.

[0027] Before selecting the vehicle-mounted chips, it is necessary to test the different vehicle-mounted chips produced by various manufacturers to obtain an accurate judgment of these vehicle-mounted chips. By testing different chips, you can choose vehicle-mounted chips with better performance and more powerful functions, thereby improving the performance of the intelligent driving system.

[0028] However, it is very difficult to systematically evaluate the automotive chips launched by various manufacturers. This is because, for example, most automotive chips are integrated chips rather than single-function chips, and their performance is closely related to the design. In addition, for example, the part of the automotive chip that undertakes the artificial intelligence (AI) acceleration function is subject to the software algorithm. The industry does not have a unified evaluation standard for automotive chips, and the various performance evaluation methods make it difficult to compare the chip products of various suppliers horizontally. Furthermore, the existing performance evaluation methods are often very one-sided and can only evaluate the performance of automotive chips from one aspect, while ignoring the influence of other aspects, such as the operating performance or application scenarios of the automotive chips.

[0029] At least in order to solve at least some of the above and other potential problems, an embodiment of the present disclosure proposes a solution for an on-board chip. The solution includes determining the static performance score of the on-board chip based on multiple static metric weights and multiple static metric scores of multiple static metrics for the on-board chip, wherein the multiple static metrics indicate the hardware performance of the on-board chip in a non-operating state. The solution also includes determining the dynamic performance score of the on-board chip based on multiple dynamic metric weights and multiple dynamic metric scores of multiple dynamic metrics for the on-board chip, wherein the multiple dynamic metrics indicate the operating performance of the on-board chip in an operating state. The solution also includes determining the comprehensive performance score of the on-board chip based on the static performance score and the dynamic performance score. According to the performance evaluation method for on-board chips of the embodiment of the present disclosure, standardized performance testing for on-board chips is realized, so that different on-board chips can be tested and evaluated through a unified evaluation system. In addition, by comprehensively considering the static and dynamic performance performance of the on-board chip, a more comprehensive understanding of the on-board chip can be obtained, and the accuracy of the performance evaluation of the on-board chip is improved.

[0030] Reference below Figures 1 to 8 It should be understood that these exemplary embodiments are provided only to enable those skilled in the art to better understand and implement the embodiments of the present disclosure, and are not intended to limit the scope of the present disclosure in any way.

[0031] Figure 1 Schematically illustrates a schematic diagram of an example environment 100 in which methods and / or processes according to embodiments of the present disclosure may be implemented. Figure 1 As shown in FIG. 1 , in some embodiments, an example environment 100 includes a vehicle chip 110 and a test device 120 coupled to each other (eg, via a wire or a network).

[0032] In order to accurately judge the performance of the automotive chips produced by various suppliers, it is necessary to test these automotive chips. Figure 1As shown in , the on-board chip 110 is deployed on an automotive product such as a car 111. According to an embodiment of the present disclosure, the on-board chip 110 may include an on-board system on chip (System on Chip, SoC) chip. A SoC chip is a single chip that integrates multiple functional modules. In other words, a complex electronic system is implemented on a single integrated circuit chip. SoC chips are widely used in various frontier fields including intelligent driving due to their high integration, low power consumption, and simple and compact design. It should be understood that the description of on-board chips and SoC chips herein is only exemplary and non-restrictive. The present disclosure is not limited to the examples described, and the on-board chip 110 may also include other different on-board chips.

[0033] In the example environment 100, the test device 120 is coupled to the vehicle chip 110 to perform a comprehensive and all-round test on the vehicle chip 110. According to an embodiment of the present disclosure, in addition to evaluating the hardware parameters and design layout of the vehicle chip 110, the performance of the vehicle chip 110 during operation can also be tested. In the example, the test device 120 can monitor various parameters of the car 111 including the vehicle chip 110 during operation to accurately evaluate the vehicle chip 110. In another example, through the test device 120, it is possible to simulate the operating environment of the vehicle chip 110 when executing various functions of the automotive product, and accurately monitor and characterize the performance of the vehicle chip 110 in the simulated operating environment.

[0034] According to an embodiment of the present disclosure, the test device 120 may be a device with computing power and storage capacity. According to an embodiment of the present disclosure, the comprehensive performance evaluation model for the on-board chip 110 may be deployed on the test device 120 for performing a comprehensive performance evaluation of the on-board chip 110. Examples of the test device 120 may include a personal computer, a server computer, a multi-processor system, etc. Examples of the test device 120 may also include a distributed computing environment including any one of the above-mentioned devices or systems, etc. It should be understood that the description of the test device 120 herein is only exemplary and not restrictive. The present disclosure is not limited to the described examples, and the test device 120 may also include other different devices. In the following, the performance evaluation method for the on-board chip 110 according to an embodiment of the present disclosure and the comprehensive performance evaluation model will be further described in detail.

[0035] According to an embodiment of the present disclosure, the chip description 130 for the vehicle-mounted chip 110 can be input into the test device 120 for further improving the test (for example, for determining multiple static metrics for the vehicle-mounted chip 110 as described below). The chip description 130 can be a document associated with the vehicle-mounted chip 110, including parameters and uses of the vehicle-mounted chip 110, etc., such as in Figure 1 As shown in the exemplary embodiment of FIG. 1 . It should be understood that the description of the chip description 130 herein is only exemplary and not restrictive. The present disclosure is not limited to the described examples, and the chip description 130 may also be an image, audio, video, etc. associated with the vehicle-mounted chip 110. The performance evaluation process for the vehicle-mounted chip 110 using the chip description 130 will be described in more detail below.

[0036] After completing the performance test for the on-board chip 110 according to an embodiment of the present disclosure, the test results can be transmitted to a display device for display to, for example, a development engineer of an automotive product. The display device can be any device with a display function. In an example, the display device can be a display dedicated to display, such as a television or a projector, or can be an electronic device with a display unit, such as a laptop computer or a smart phone. In another example, the display device can be in the test device 120, and after the test device 120 completes the performance test for the on-board chip 110, the test results are transmitted to the display device located therein for display. In this way, a separate display device is no longer required. It should also be noted that each of the above examples is not intended to be restrictive, but is only exemplarily shown for the purpose of helping understanding, and the embodiments of the present disclosure are not limited to the above examples.

[0037] Combined with the above Figure 1 A schematic diagram of an example environment 100 is depicted in which methods and / or processes according to embodiments of the present disclosure may be implemented. Figure 2 The flowchart of the performance evaluation method 200 for the vehicle-mounted chip 110 according to the embodiment of the present disclosure is described. As described above for the test device 120, the performance evaluation method 200 for the vehicle-mounted chip 110 according to the embodiment of the present disclosure can be executed at a local device with computing power and storage power, or at a cloud server, and the present disclosure does not limit this.

[0038] Figure 2The flowchart of the performance evaluation method 200 for the vehicle chip 110 according to an embodiment of the present disclosure is schematically illustrated. At box 210, the static performance score of the vehicle chip 110 is determined based on multiple static metric weights and multiple static metric scores for multiple static metrics of the vehicle chip 110, wherein the multiple static metrics indicate the hardware performance of the vehicle chip 110 in a non-operating state. In order to obtain an accurate evaluation of the vehicle chip 110, it is first necessary to test the hardware performance of the vehicle chip 110 in a non-operating state.

[0039] A typical AI chip is a heterogeneous SoC that integrates multiple functional cores into one physical semiconductor chip. Typically, SoC OEMs provide descriptions such as technical data sheets to show the hardware parameters, design layout, etc. of their products. Such descriptions help automotive product development engineers select the SoC that is most suitable for a particular development work. According to the performance evaluation method 200 for the on-board chip 110 of an embodiment of the present disclosure, a quantitative evaluation method can be used to characterize the hardware parameters, design layout, etc. of the SoC as multiple static metrics, and an accurate evaluation of the hardware performance of the on-board chip 110 can be obtained based on the weights and scores of these static metrics. In the following, an example process for determining static metrics and static performance scores for the on-board chip 110 will be further described in detail.

[0040] At block 220, a dynamic performance score of the onboard chip 110 is determined based on multiple dynamic metric weights and multiple dynamic metric scores for the multiple dynamic metrics of the onboard chip 110, wherein the multiple dynamic metrics indicate the operating performance of the onboard chip 110 under the operating state. In order to further obtain an accurate evaluation of the onboard chip 110, it is also necessary to test the operating performance of the onboard chip 110 under the operating condition.

[0041] If only the hardware performance of the vehicle-mounted chip 110 is evaluated, it will not be enough to fully understand its performance in actual application scenarios. If only the hardware performance of the vehicle-mounted chip 110 is considered, the evaluation results will be one-sided and inaccurate. This may cause the development engineers of automotive products to mistakenly select a certain SoC, which in turn leads to undesirable adverse consequences such as functional degradation of automotive products. For example, descriptions such as technical data sheets as mentioned above often have specific biases. In some extreme cases, the selected SoC obviously has excellent hardware properties, but is not suitable for the application scenarios of automotive products and exhibits degraded performance.

[0042] According to the performance evaluation method 200 for the vehicle-mounted chip 110 of the embodiment of the present disclosure, the operating performance of the vehicle-mounted chip 110 under operating conditions can be characterized by multiple dynamic metrics, and an accurate evaluation of the operating performance of the vehicle-mounted chip 110 can be obtained based on the weights and scores of these dynamic metrics. Hereinafter, an example process of determining the dynamic metric score of the vehicle-mounted chip 110 will be described in further detail.

[0043] At box 230, the comprehensive performance score of the vehicle chip 110 is determined based on the static performance score and the dynamic performance score. According to an embodiment of the present disclosure, based on the static performance score indicating the hardware performance of the vehicle chip 110 determined at box 210, and the dynamic performance score indicating the operating performance of the vehicle chip 110 determined at box 220, the comprehensive performance of the vehicle 110 can be evaluated in all directions, thereby helping automotive product development engineers to perform accurate chip selection. In this way, a standardized performance test for the vehicle chip 110 is provided, so that different vehicle chips can be tested and evaluated through a unified evaluation system. In addition, by comprehensively considering the static and dynamic performance of the vehicle chip, a more comprehensive understanding of the vehicle chip can be obtained, and the accuracy of the performance evaluation of the vehicle chip is improved.

[0044] Combined with the above Figure 2 The flowchart of the performance evaluation method 200 for the vehicle-mounted chip 110 according to an embodiment of the present disclosure is described. Figure 3 1 is a schematic diagram of a comprehensive performance evaluation model 300 for a vehicle-mounted chip 110 according to an embodiment of the present disclosure. It should be understood that: Figure 3 The various modules shown in the figure are merely exemplary and non-limiting, and such a comprehensive performance evaluation model 300 may also include more or fewer modules.

[0045] Figure 3 The schematic diagram of the comprehensive performance evaluation model for the vehicle-mounted chip 110 according to the embodiment of the present disclosure is schematically illustrated. Figure 3As shown in , the comprehensive performance evaluation model 300 for the vehicle-mounted chip 110 according to an embodiment of the present disclosure may include a hardware performance evaluation module 310, an operation performance evaluation module 320, and a comprehensive performance evaluation module 330. The hardware performance evaluation module 310 may be configured to obtain a chip description (such as a chip description 130) for the vehicle-mounted chip 110, and determine multiple static metrics for the vehicle-mounted chip 110 based on the chip description. The hardware performance evaluation module 310 may determine the static performance score 311 of the vehicle-mounted chip 110 based on the static metric weights and static metric scores of these static metrics, thereby obtaining an accurate evaluation of the hardware performance of the vehicle-mounted chip 110. The hardware performance evaluation module 310 may use different static metric weights depending on the application scenario adapted by the target automotive product or the functional system emphasized, so that the performance evaluation of the vehicle-mounted chip 110 is more accurate.

[0046] According to an embodiment of the present disclosure, the operation performance evaluation module 320 may be configured to monitor the performance of the onboard chip 110 in the operation state. The operation performance evaluation module 320 may simulate the operation environment of the onboard chip 110 in the application scenario of the automobile product, and monitor the performance of the onboard chip 110 in the operation state at this time. The operation performance evaluation module 320 may determine the dynamic performance score 312 of the onboard chip 110 based on the dynamic metric weight and dynamic metric score of the dynamic metric of the onboard chip 100, thereby obtaining an accurate evaluation of the operation performance of the onboard chip 110.

[0047] According to an embodiment of the present disclosure, the comprehensive performance evaluation module 330 determines the comprehensive performance score of the vehicle chip 110 based on the static performance score 311 of the vehicle chip 110 from the hardware performance evaluation module 310 and the dynamic performance score 312 of the vehicle chip 110 from the operation performance evaluation module 320. In this way, the comprehensive performance evaluation model 300 for the vehicle chip 110 according to the embodiment of the present disclosure can comprehensively consider the static and dynamic performance of the vehicle chip 110, so as to accurately obtain a more comprehensive understanding of the vehicle chip. More details of each module of the comprehensive performance evaluation model 300 for the vehicle chip 110 according to the embodiment of the present disclosure will be further described in detail below.

[0048] Figure 4 The exemplary process of determining multiple static metrics 430 for the vehicle-mounted chip 110 according to an embodiment of the present disclosure is schematically illustrated. Figure 4 As shown in , the hardware performance evaluation module 310 according to an embodiment of the present disclosure may include an extraction module 410 and a parsing module 420 . Figure 4 The various modules shown in the figure are merely exemplary and non-limiting, and such a hardware performance evaluation module 310 may also include more or fewer modules.

[0049] According to an embodiment of the present disclosure, the extraction module 410 may be configured to extract at least a portion associated with the hardware performance of the vehicle chip 110 from the chip description 130 for the vehicle chip 100. The chip description 130 may be a document associated with the vehicle chip 110 (such as a description such as a technical data sheet provided by an original equipment manufacturer of the SoC), including parameters and uses of the vehicle chip 110 (such as hardware parameters, design layout, etc. of the SoC). In some embodiments, the extraction module 410 may be configured to extract at least a portion associated with the hardware performance of the vehicle chip 110 from the chip description 130 based on keywords associated with the performance of the vehicle chip 110. Keywords associated with the performance of the vehicle chip 110 may include speed, size, capacity, processing power, etc.

[0050] It should be understood that the extraction is performed here based on keywords associated with the performance of the vehicle chip 110. However, this is only exemplary and not restrictive, and the extraction can also be performed based on other identifiers (such as numbers, types, etc.) associated with the performance of the vehicle chip 110, and the embodiments of the present disclosure are not limited to this. In addition, the description of the chip description 130 here is only exemplary and not restrictive. The present disclosure is not limited to the examples described, and the chip description 130 can also be an image, audio, video, etc. associated with the vehicle chip 110.

[0051] According to an embodiment of the present disclosure, the parsing module 420 may be configured to parse at least a portion of the information associated with the hardware performance of the vehicle-mounted chip 110 extracted by the extraction module 410 to determine a plurality of static metrics 430 (such as Figure 4 In the example of FIG. 4 , the analysis module 420 may use machine learning, statistical analysis, image processing, natural language processing and other techniques to identify and analyze the extracted data, and the embodiments of the present disclosure are not limited thereto.

[0052] Figure 5 An example 500 of multiple application scenarios adapted by the vehicle-mounted chip 110 is schematically illustrated. During the design and development process, the development engineers of the automotive product may focus on the implementation of specific functions adapted to specific scenarios as the unique selling point of the product. Here, the functional system corresponding to the specific function is further described in detail.

[0053] Figure 5Shown in (1) is an example of an advanced driver assistance system (ADAS). An ADAS system is a system that improves the safety performance of a car through on-board electronic devices and sensor technology. These ADAS systems can identify the environment around the vehicle, including other vehicles, pedestrians, road signs, and traffic lights, and take appropriate measures to help the driver avoid accidents. For example, many vehicles are now equipped with features such as adaptive cruise control, lane departure warning, blind spot monitoring, automatic emergency braking, and parking assistance. These ADAS systems help improve driving safety, reduce the risk of accidents, and improve the driving experience.

[0054] In the ADAS scenario, it is necessary to process massive amounts of data, as well as programs with high requirements for input and output IO, computing, and real-time performance, especially AI programs. For IO ports, speed is extremely important, not quantity. Accelerator units for heavy-duty autonomous AI applications should process several types of massive data generated by different sensors (such as cameras, lidars, radars, etc.) in parallel. Therefore, when designing and developing automotive products that focus on ADAS systems, it should be considered whether the on-board chip can adapt to such application scenarios.

[0055] Figure 5 (2) shows an example of an in-vehicle infotainment system (IVI). The IVI system is a multimedia system built into a car, including functions such as audio, video, navigation, and communication. The IVI system can be operated through a touch screen, voice commands, or buttons on the vehicle, so that the driver and passengers can enjoy driving and entertainment at the same time. The IVI system can provide functions such as listening to the radio, playing music, viewing maps, searching for destinations, receiving calls, sending text messages, etc., and with the development of technology, modern IVI systems can also support advanced functions such as smart home control and remote vehicle monitoring.

[0056] In the IVI scenario, many human-computer interaction and entertainment applications are run, some of which run on the Android system. With more IO interfaces, the IVI system requires higher graphics computing performance and IO throughput performance. In addition, the IVI system can include multiple high-resolution screens, that is, basic multi-screen support. Nowadays, full LCD instruments and large central control screens have become basic configurations of automotive products. For multiple screens, the graphics processing unit (GPU) requires at least 100GFLOPS (billion floating-point operations per second). The neural network (NN) accelerator unit has a speed of up to 300GOPS (billion operations per second) and can run lightweight AI applications for human-computer interaction with the natural language processing (NLP) function enabled. Therefore, when designing and developing automotive products focusing on the IVI system, it should be considered whether the on-board chip can adapt to such application scenarios.

[0057] It should be noted that the above-mentioned ADAS system and IVI system (or, ADAS scenario and IVI scenario) are only examples of functional systems exemplified for the purpose of ease of understanding and convenience of explanation, and more functional systems and application scenarios may be considered during the comprehensive performance evaluation of the vehicle-mounted chip 110. Hereinafter, how to consider these functional systems and application scenarios during the comprehensive performance evaluation of the vehicle-mounted chip 110 according to an embodiment of the present disclosure will be further described in detail.

[0058] Through the above Figure 4 According to the example process of an embodiment of the present disclosure shown in , multiple static metrics 430 for the onboard chip 110 can be determined. According to an embodiment of the present disclosure, the multiple static metrics 430 for the onboard chip 110 may include at least two of the following items: production process, central processing unit CPU architecture, system safety monitoring control unit MCU, graphics processing unit / visual processing unit GPU / VPU, neural network accelerator, memory, peripheral interface, camera serial interface, and image signal processor. It should be noted that the description of the multiple static metrics 430 herein is only exemplary and not restrictive, and the multiple static metrics 430 according to an embodiment of the present disclosure may also include more or fewer static metrics. These static metrics 430 are Figure 6 In the example, static metrics 430 may be schematically represented by static metrics 1, static metrics 2, ..., static metrics n. Below, these static metrics 430 will be described in further detail.

[0059] Regarding production processes, all computer chips, including CPUs and AI chips, are made up of millions or even billions of transistors. The more transistors integrated into a semiconductor of precise size, the higher the performance it will provide, that is, performance is highly correlated with the number of transistors per unit area. Node manufacturing technology and chip size are physical limitations on the number of transistors. Regarding CPU architecture, measuring the performance of a CPU is a complex task. The problem is that different CPUs are good at different jobs. In an SoC, the CPU performs arithmetic, logic, control, and input / output (I / O) operations specified by the program and schedules overall resources. Throughput is a key indicator of CPU performance. It represents the number of instructions executed at a given clock rate per unit time.

[0060] Regarding the system safety monitoring control unit MCU, it is an independent control and monitoring unit used to observe the status of the SoC when it is working. It is a mandatory part of the functional requirements of the automotive safety integrity level ASIL-B or higher ASIL level. Regarding the GPU / VPU, it is designed for graphics by optimizing matrix graphics calculations. In general GPU-based applications, the main processing power is still graphics processing and display output. Especially for the field of smart cockpits, 3D GPU accelerators are a must for optimizing the drawing of large-scale complex UI interfaces.

[0061] Regarding neural network accelerators, due to the natural opposition between general-purpose computing and specialized computing, unique algorithms running on general-purpose processors have worse execution efficiency, especially for neural network algorithms. Neural network accelerators are designed to accelerate the performance of standard machine learning tasks for specific NN models to increase speed and balance power efficiency. Regarding memory, a memory controller is a digital circuit that manages the flow of data to and from the computer's main memory. The memory controller can be separated or integrated into the SoC. The efficiency of data throughput can significantly affect SoC performance.

[0062] Regarding the peripheral interface, it interconnects with other onboard chips or electronic control units (ECUs) in electronic / electrical systems (E / E). High-speed peripheral interfaces are essential for data interaction with other ECUs and debugging tools in the E / E. Regarding the camera serial interface, the Camera Serial Interface 2 (MIPI CSI-2) is an interface between the camera and the host processor and is a specification of the Mobile Industry Processor Interface (MIPI) Alliance. The Camera Serial Interface CSI specifies the ability to process up to 2.5 billion pixels per second from various video input devices. Regarding the image signal processor, it is designed to process digital image forms by performing basic operations such as demosaicing, noise reduction, auto exposure, auto focus, auto white balance, image sharpening and complex operations, as well as complex operations such as LED flicker mitigation and HDR (high dynamic range) processing.

[0063] Figure 6 The exemplary process 600 of determining the static performance score 311 of the static metric 430 of the vehicle-mounted chip 110 according to an embodiment of the present disclosure is schematically illustrated. The hardware performance evaluation for the vehicle-mounted chip 110 according to an embodiment of the present disclosure may be performed by the hardware performance evaluation model 310 in the comprehensive performance evaluation model 300. According to an embodiment of the present disclosure, multiple static metrics 430 (such as, Figure 4, static metric n) exemplarily shown in FIG. 1 , static metric 1, static metric 2, ..., static metric n) are used to set multiple static metric weights 620. By way of example and not limitation, static metric 1 corresponds to static metric weight 1 or static metric weight 1* in multiple static metric weights 620, static metric 2 corresponds to static metric weight 2 or static metric weight 2* in multiple static metric weights 620, static metric n corresponds to static metric weight n or static metric weight n* in multiple static metric weights 620, and so on.

[0064] According to an embodiment of the present disclosure, the plurality of static metric weights 620 may be divided into a plurality of static metric weight groups (such as, Figure 6 In addition, each static metric weight group may have different static metric weights. For example, static metric weight 1 in static metric weight group 621 may be different from static metric weight 2, static metric weight 2 may be different from static metric weight n, or static metric weight 1, static metric weight 2, ..., static metric weight n are all different static metric weights, and so on.

[0065] According to an embodiment of the present disclosure, multiple static measurement weight groups may correspond to multiple scenarios (such as the above-mentioned ADAS scenario and IVI scenario). For example, static measurement weight group 621 may correspond to an ADAS system, static measurement weight group 622 may correspond to an IVI system, and so on. In other words, different static measurement weight groups may be used for different scenarios, and these static measurement weight groups are designed differently for different scenarios. In this way, it is possible to focus on the specific functions that the automotive product focuses on, thereby obtaining an accurate evaluation of the on-board chip 110.

[0066] According to an embodiment of the present disclosure, it is possible to determine which of the multiple scenarios the onboard chip 110 is to be used for (hereinafter, also referred to as the target scenario). It is possible to determine a static metric weight group corresponding to the target scenario among the multiple static metric weight groups (hereinafter, also referred to as the target static metric weight group). A static performance score 610 (such as, Figure 6 The static metric score 1, static metric score 2, ..., static metric score n) exemplarily shown in the figure and the target static metric weight group can determine the static performance score 311 of the on-board chip 110 for the target scenario.

[0067] The example process 600 of determining the static performance score 311 of the on-board chip 110 according to an embodiment of the present disclosure is described in an exemplary and non-limiting manner. It can be determined that the on-board chip 110 is to be used for the ADAS system. It can be determined that the static metric weight group 621 in the multiple static metric weight groups corresponds to the ADAS system. Based on the static performance scores 610 of multiple static metrics 430 and the static metric weights in the static metric weight group 621, the static performance score 311 of the on-board chip 110 for the target functional system can be determined. Below, Table 1 is used to exemplarily illustrate an example of multiple static metric weight groups of the on-board chip 110.

[0068]

[0069] Table 1 Static metric weights

[0070] As shown in Table 1, the plurality of static metric weights 620 of the plurality of static metrics 430 (ie, the second column shown in Table 1) are divided into a plurality of static metric weight groups. That is, Figure 1 The ADAS weights (i.e., the 3rd column shown in Table 1) and the IVI weights (i.e., the 4th column shown in Table 1) shown in . The production process weight (i.e., 5%) and the CPU architecture weight (i.e., 20%) in the ADAS weight are different. Based on multiple static metric scores 610 of multiple static metrics 430 and the corresponding static metric weight group (i.e., the target static metric weight group) in the static metric weight group 620, the static performance score 311 of the on-board chip 110 for the target functional system can be determined. The static performance score 311 of the on-board chip 110 for the target functional system can be determined by the following formula (1).

[0071]

[0072] Total_score j The static performance score of the vehicle chip 110 for the target scenario is 311, Component_Scores ij is a static performance score 610 of a plurality of static metrics 430, and Weight ij is each weight in the target static metric weight group of the plurality of static metric weight groups 620. Based on each weight in the target static metric weight group of the plurality of static metric weight groups 620 (ie, Weight shown in formula (1) ij ), the static performance scores 610 of the multiple static metrics 430 (ie, the Component_Scores shown in formula (1)) can be calculated. ij ) is weighted averaged, so as to determine the static performance score 311 of the vehicle-mounted chip 110 for the target scene (ie, the Total_score shown in formula (1)j ).

[0073] According to an embodiment of the present disclosure, each static metric in the plurality of static metrics 430 includes a plurality of static sub-metrics. Based on a plurality of static sub-metric weights 640 and a plurality of static sub-metric scores 630 of the plurality of static sub-metrics, a static metric score 610 may be determined for each static metric, wherein the plurality of static sub-metric weights 640 may be different for the plurality of static sub-metrics. Figure 6 An example process of determining a static metric score 1 among the plurality of static metric scores 610 is schematically shown. It should be noted that Figure 6 For ease of illustration, only an example process of determining static metric score 1 among multiple static metric scores 610 is shown, and static metric score 2, ..., static metric score n among multiple static metric scores 610 may also be determined in the same or similar manner.

[0074] According to an embodiment of the present disclosure, multiple static sub-metric weights 640 may be set for multiple static sub-metrics. By way of example and not limitation, static sub-metric 1 corresponds to static sub-metric weight 1 in multiple static sub-metric weights 640, static sub-metric 2 corresponds to static sub-metric weight 2 in multiple static sub-metric weights 640, static sub-metric n corresponds to static sub-metric weight n in multiple static sub-metric weights 640, and so on. Static sub-metric weight 1 in multiple static sub-metric weights 640 may be different from static sub-metric weight 2, static sub-metric weight 2 may be different from static sub-metric weight n, or static sub-metric weight 1, static sub-metric weight 2, ..., static sub-metric weight n are all different static sub-metric weights, and so on.

[0075] By way of example and not limitation, if static metric 1 is a production process, static metric 1 may include technology node and chip size as static sub-metric 1 and static sub-metric 2. A static sub-metric weight 1 of 80% may be set for static sub-metric 1, and a static sub-metric weight 2 of 20% may be set for static sub-metric 2. The score of the production process (static metric score 1) may be determined by weighted averaging the score of the technology node of the on-board chip 110 (i.e., static metric score 1) and the score of the chip size (i.e., static metric score 2). It should be understood that the description herein is exemplary and not limiting. The static metric may include more or fewer static sub-metrics, and the static metric score 610 may be determined for each of the multiple static metrics 430 in the same or similar manner.

[0076] Figure 7An example process 700 of determining a dynamic performance score 312 of a dynamic metric for an onboard chip 110 according to an embodiment of the present disclosure is schematically illustrated. In order to conduct a comprehensive and all-round test on the onboard chip 110, in addition to evaluating the hardware parameters and design layout of the onboard chip 110, the performance of the onboard chip 110 during operation can also be tested. The dynamic performance evaluation of the onboard chip 100 according to an embodiment of the present disclosure can monitor the operating performance of the onboard chip 110 when executing various functions of an automotive product.

[0077] According to an embodiment of the present disclosure, the multiple dynamic metrics for the vehicle-mounted chip 110 may include at least two of the following items: artificial intelligence AI, memory, central processing unit CPU, network input / output I / O, and storage device I / O. It should be noted that the description of the multiple dynamic metrics here is only exemplary and not restrictive, and the multiple dynamic metrics according to the embodiment of the present disclosure may also include more or fewer dynamic metrics. These dynamic metrics are Figure 7 In the example, dynamic metrics 1, dynamic metrics 2, ..., dynamic metrics n can be schematically represented. These dynamic metrics will be described in further detail below.

[0078] Regarding AI, different algorithms require different computing power. When evaluating the performance of AI SoC solutions, the hardware and software integration of autonomous AI applications is very important. Regarding memory, memory speed represents the speed of data reading / writing / copying. Regarding CPU, CPU is the central brain of SoC and even the entire hardware component. Regarding network I / O, IO port evaluation focuses on the throughput of high-speed peripheral interfaces. Regarding storage device I / O, user data is usually stored in storage devices, and the speed of storage devices can be tested through storage interfaces.

[0079] According to an embodiment of the present disclosure, multiple dynamic metric weights 720 may be set for multiple dynamic metrics. By way of example and not limitation, dynamic metric 1 corresponds to dynamic metric weight 1 among multiple dynamic metric weights 720, dynamic metric 2 corresponds to dynamic metric weight 2 among multiple dynamic metric weights 720, dynamic metric n corresponds to dynamic metric weight n among multiple dynamic metric weights 720, and so on. In addition, the multiple dynamic metric weights 720 set may be different for multiple dynamic metrics. For example, dynamic metric weight 1 may be different from dynamic metric weight 2, dynamic metric weight 2 may be different from dynamic metric weight n, or dynamic metric weight 1, dynamic metric weight 2, ..., dynamic metric weight n are all different dynamic metric weights, and so on.

[0080] In some embodiments, multiple dynamic metric weights 720 of multiple dynamic metrics can be divided into multiple dynamic metric weight groups. In an exemplary and non-limiting manner, multiple dynamic metric weight groups can be associated with multiple scenarios (such as the above-mentioned ADAS scenario and IVI scenario, etc.). Multiple dynamic metric weight groups can also correspond to multiple vehicle models (such as SUVs, sedans, sports cars, trucks, etc.). Multiple dynamic metric weight groups can also be associated with multiple environments (such as urban roads, mountainous areas, coastal areas, highways, rural and rural areas, etc.).

[0081] The example process 700 of determining the dynamic performance score 312 of the onboard chip 110 according to the embodiment of the present disclosure is described in an exemplary and non-limiting manner. Based on each dynamic metric weight in the multiple dynamic metric weights 720 of the multiple dynamic metrics, the multiple dynamic metric scores 710 of the multiple dynamic metrics can be weighted averaged, so that the dynamic performance score 312 of the onboard chip 110 can be determined. Below, Table 2 is used to exemplarily show an example of multiple dynamic metric weights 720 for multiple dynamic metrics of the onboard chip 110.

[0082] No.(i) Dynamic Metrics Weight i 1 AI 55% 2 Memory 20% 3 CPU 15% 4 Network I / O 5% 5 Storage Device I / O 5% Total 100%

[0083] Table 2 Dynamic metric weights

[0084] Table 2 schematically illustrates multiple dynamic metric weights 720 (i.e., column 3 shown in Table 2) of multiple dynamic metrics (i.e., column 2 shown in Table 2). The dynamic metric AI may be assigned a dynamic metric weight of 55%. Next are memory and CPU, which may be assigned dynamic metric weights of 20% and 15%, respectively. Next are network I / O and storage device I / O, which may be assigned dynamic metric weights of 5% and 5%, respectively. Some of the multiple dynamic metric weights in Table 2 may be different, and some of the multiple dynamic metric weights may be the same. It should be understood that this is exemplary and non-restrictive, and that these dynamic metric weights may all be the same or all different. Based on multiple dynamic metric scores 710 and multiple dynamic metric weights 720 of multiple dynamic metrics, the dynamic performance score 312 of the onboard chip 110 may be determined. The dynamic performance score 312 of the onboard chip 110 may be determined by the following formula (2).

[0085]

[0086] Total_score j The dynamic performance score of the vehicle chip 110 is 312, Component_Tested ij is a plurality of dynamic metrics scores 710 and Weights ijis a plurality of dynamic metric weights 720 of the plurality of dynamic metrics. Based on the plurality of dynamic metric weights 720 (ie, the Weights shown in equation (2) ij ), multiple dynamic metric scores 710 (i.e., Component_Tested ij ) is weighted averaged to determine the dynamic performance score 312 of the vehicle-mounted chip 110) (ie, the Total_score shown in formula (2) j ).

[0087] According to an embodiment of the present disclosure, each of the multiple dynamic variables of the vehicle chip may include multiple dynamic sub-metrics. Based on multiple dynamic sub-metric weights 740 and multiple dynamic sub-metric scores 730 of the multiple dynamic sub-metrics, a dynamic metric score 710 is determined for each dynamic metric, wherein the multiple dynamic sub-metric weights 740 are the same for the multiple dynamic sub-metrics. Figure 7 An example process of determining a dynamic metric score 1 among the plurality of dynamic metric scores 710 is schematically shown. It should be noted that Figure 7 For ease of illustration, only an example process of determining dynamic metric score 1 among multiple dynamic metric scores 710 is shown, and dynamic metric score 2, ..., dynamic metric score n among multiple dynamic metric scores 710 may also be determined in the same or similar manner.

[0088] According to an embodiment of the present disclosure, multiple dynamic sub-metric weights 740 may be set for multiple dynamic sub-metrics. By way of example and not limitation, dynamic sub-metric 1 corresponds to dynamic sub-metric weight 1 in multiple dynamic sub-metric weights 740, dynamic sub-metric 2 corresponds to dynamic sub-metric weight 2 in multiple dynamic sub-metric weights 740, dynamic sub-metric n corresponds to dynamic sub-metric weight n in multiple dynamic sub-metric weights 740, and so on. Dynamic sub-metric weight 1, dynamic sub-metric weight 2, ... dynamic sub-metric weight n in multiple dynamic sub-metric weights 740 may be the same. In other words, multiple dynamic sub-metric weights 740 are the same for multiple dynamic sub-metrics, so that in the case of insufficient knowledge for multiple dynamic sub-metrics 740, dynamic metric scores can be robustly and accurately determined. It should be understood that multiple dynamic sub-metric weights 740 may also be different for multiple dynamic sub-metrics. For example, the dynamic sub-metric weight 1 among the multiple dynamic sub-metric weights 740 may be different from the dynamic sub-metric weight 2, the dynamic sub-metric weight 2 may be different from the dynamic sub-metric weight n, or the dynamic sub-metric weight 1, the dynamic sub-metric weight 2, ..., the dynamic sub-metric weight n are all different dynamic sub-metric weights, and so on.

[0089] By way of example and not limitation, if dynamic metric 1 is AI, dynamic metric 1 may include image classification and object detection as dynamic sub-metrics. According to an embodiment of the present disclosure, different data sets and models may be used to enable the onboard chip 110 to perform image classification and object detection to determine the image classification capability and object detection capability of the onboard chip 110, and then determine the dynamic metric score of the dynamic metric AI for the onboard chip 110. Figure 8 Let’s further describe this process.

[0090] Figure 8 An example process 800 of load testing for the onboard chip 110 according to an embodiment of the present disclosure is schematically illustrated, and the example process 800 may be executed by the operating performance evaluation module 320. In order to determine the performance of the onboard chip 110 under operating conditions, a load test may be performed on the onboard chip 110 using multiple data sets and multiple models to obtain the value of the load parameter of the onboard chip 110. Figure 8 As shown in , an input 810 including multiple data sets and multiple models can be transmitted to the operation performance evaluation module 320. The preprocessing module 820 included in the operation performance evaluation module 320 can be configured to preprocess the multiple data sets before they are loaded into the on-board chip 110, such as decoding, cropping, scaling, type conversion, normalization, etc. Next, the preprocessing module 830 included in the operation performance evaluation module 320 can perform a load test using the preprocessed data, and can monitor the value of the load parameter of the on-board chip 110 during the load test. The load parameters of the on-board chip 110 may include delay, throughput, throughput of delay boundaries, etc. Based on the value of the load parameter of the on-board chip 110, the multiple dynamic sub-metric scores 730 of the multiple dynamic sub-metrics can be obtained. For example, the value of the load parameter when the on-board chip 110 performs image classification using a certain data set and a certain model can be normalized to a dynamic sub-metric score. In other words, the values ​​of load parameters when the onboard chip 110 performs a task using different data sets and models can be monitored, and the values ​​of the monitored load parameters can be normalized to obtain multiple dynamic sub-metric scores 730 of multiple dynamic sub-metrics.

[0091] According to an embodiment of the present disclosure, after obtaining multiple dynamic sub-metric scores 730, a dynamic metric score 710 may be determined for each dynamic metric based on multiple dynamic sub-metric weights 740 of the multiple dynamic sub-metrics. Below, Table 3 is used to exemplarily illustrate the dynamic sub-metric scores of the dynamic sub-metrics of the dynamic metric AI for the onboard chip 110.

[0092]

[0093] Table 3 Dynamic sub-metric scores of dynamic sub-metrics of dynamic metric AI

[0094] As shown in Table 3, the task performed by the on-board chip 110 (i.e., the second column shown in Table 3), the dataset used (i.e., the third column shown in Table 3), and the model (i.e., the fourth column shown in Table 3) can be regarded as a dynamic sub-metric, that is, the task-dataset-model entry can be regarded as a dynamic sub-metric. Taking the first entry (i.e., number (1) shown in Table 3) as an example, dynamic sub-metric 1 can indicate that the on-board chip performs image classification using the dataset ImageNet (224×224) and the model Resnet50-v1.5, and the dynamic sub-metric weight 1 with dynamic sub-metric 1 can be 20%. It should be understood that the description of other dynamic sub-metrics is similar and will not be repeated here. The dynamic metric score of the dynamic metric of the on-board chip 110 can be determined by the following formula (3).

[0095]

[0096] in is the dynamic metric score of the vehicle chip 110 710, It is multiple dynamic sub-metric scores 730 of multiple dynamic metrics, and Weight i is a plurality of dynamic metric sub-metric weights 740 of the plurality of dynamic metrics. Based on the plurality of dynamic sub-metric weights 740 (Weight in formula (3) i ), multiple dynamic sub-metric scores 730 (in formula (3) ) is weighted averaged to determine the dynamic metric score 710 of the vehicle-mounted chip 110 (in formula (3) ).

[0097] Through the above-mentioned methods and processes according to the embodiments of the present disclosure, multiple dynamic metric scores 710 can be determined based on multiple dynamic sub-metric weights 740 and multiple dynamic sub-metric scores 730, and the dynamic performance score 312 can be further determined based on multiple dynamic metric scores 710 and multiple dynamic metric weights 720.

[0098] According to an embodiment of the present disclosure, the vehicle-mounted chip 110 may be a vehicle-mounted SoC chip. The comprehensive performance score of each vehicle-mounted SoC chip in a plurality of vehicle-mounted SoC chips may be determined, and at least one vehicle-mounted SoC chip in a plurality of vehicle-mounted SoC chips may be selected based on the comprehensive performance score of each vehicle-mounted SoC chip in a plurality of vehicle-mounted SoC chips. In this way, the candidate vehicle-mounted SoC chips may be selected.

[0099] Fig. 9900 is a schematic block diagram of an example device that can be used to implement some embodiments of the present disclosure. Fig. 9 As shown in FIG. 1 , the device 900 includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) 902 or loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The CPU 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0100] A number of components in the device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a disk, an optical disk, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the device 900 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0101] The various processes and processing described above, such as method 200, may be performed by processing unit 901. For example, in some embodiments, method 200 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by CPU 901, one or more actions of method 200 described above may be performed.

[0102] The present disclosure may be a method, a device, an apparatus, a system, and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present disclosure.

[0103] Computer readable storage medium can be a tangible device that can hold and store instructions used by an instruction execution device. Computer readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (non-exhaustive lists) of computer readable storage medium include: portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), static random access memories (SRAM), portable compact disk read-only memories (CD-ROM), digital versatile disks (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or convex structures in grooves on which instructions are stored, and any suitable combination thereof. Computer readable storage medium used herein is not interpreted as a transient signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagated by waveguides or other transmission media (e.g., light pulses by optical fiber cables), or electrical signals transmitted by wires.

[0104] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0105] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0106] Various aspects of the present disclosure are described herein with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0107] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0108] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0109] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special hardware and computer instructions.

[0110] The embodiments of the present disclosure are described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A performance evaluation method for an on-board chip, comprising: Determining a static performance score of the onboard chip based on multiple static metric weights and multiple static metric scores of multiple static metrics of the onboard chip, wherein the multiple static metrics indicate hardware performance of the onboard chip in a non-operating state; Determining a dynamic performance score of the onboard chip based on a plurality of dynamic metric weights and a plurality of dynamic metric scores for a plurality of dynamic metrics of the onboard chip, wherein the plurality of dynamic metrics indicate an operating performance of the onboard chip in an operating state; as well as Based on the static performance score and the dynamic performance score, a comprehensive performance score of the vehicle-mounted chip is determined.

2. The performance evaluation method according to claim 1, further comprising: Obtaining a chip description for the vehicle-mounted chip, wherein the chip description includes parameters and usage of the vehicle-mounted chip; extracting at least a portion associated with the hardware performance of the vehicle-mounted chip from the chip description; as well as The plurality of static metrics are determined by parsing the extracted at least one portion.

3. The performance evaluation method according to claim 2, wherein extracting the at least one portion associated with the hardware performance of the vehicle-mounted chip comprises: Based on the keyword associated with the hardware capability of the on-board chip, the at least a portion associated with the hardware capability of the on-board chip is extracted from the chip description.

4. The performance evaluation method according to claim 1, wherein the multiple static measurement weights are divided into multiple static measurement weight groups, the multiple static measurement weight groups correspond to multiple scenarios, and each static measurement weight group in the multiple static measurement weight groups has a different static measurement weight.

5. The performance evaluation method according to claim 4, determining the static performance score of the vehicle-mounted chip comprises: Determining a target scenario in which the vehicle-mounted chip is to be used among the multiple scenarios; Determining a target static measurement weight group corresponding to the target scene among the plurality of static measurement weight groups; as well as Based on the static performance scores of the multiple static metrics and the target static metric weight group, the static performance score of the on-board chip for the target scenario is determined. 6 . The performance evaluation method according to claim 5 , wherein the plurality of scenarios include an advanced driver assistance (ADAS) scenario and an in-vehicle infotainment (IVI) scenario.

7. The performance evaluation method according to claim 1, wherein the multiple static metrics include at least two of the following: production process, central processing unit CPU architecture, system safety monitoring control unit MCU, graphics processing unit / visual processing unit GPU / VPU, neural network accelerator, memory, peripheral interface, camera serial interface, and image signal processor.

8. The performance evaluation method according to claim 1, wherein each of the plurality of static metrics comprises a plurality of static sub-metrics, the method further comprising: determining a static metric score of each static metric based on a plurality of static sub-metric weights and a plurality of static sub-metric scores of the plurality of static sub-metrics, The plurality of static sub-metric weights are different for the plurality of static sub-metrics.

9. The performance evaluation method according to claim 1, wherein the multiple dynamic metric weights are different for the multiple dynamic metrics, and the multiple dynamic metrics include at least two of the following: artificial intelligence AI, memory, central processing unit CPU, network input / output I / O, and storage device I / O.

10. The performance evaluation method according to claim 1, wherein each of the plurality of dynamic metrics comprises a plurality of dynamic sub-metrics, the method further comprising: determining a dynamic metric score for each dynamic metric based on a plurality of dynamic sub-metric weights and a plurality of dynamic sub-metric scores of the plurality of dynamic sub-metrics, The plurality of dynamic sub-metric weights are the same for the plurality of dynamic sub-metrics.

11. The performance evaluation method according to claim 10, further comprising: By using multiple data sets and multiple models to perform a load test on the vehicle-mounted chip, a value of a load parameter of the vehicle-mounted chip is obtained; as well as Based on the value of the load parameter of the onboard chip, the multiple dynamic sub-metric scores of the multiple dynamic sub-metrics are obtained.

12. The performance evaluation method according to claim 11, wherein: The load parameter of the vehicle chip includes delay, throughput, or throughput at a delay boundary, and The multiple data sets are preprocessed before being loaded into the on-board chip, and the preprocessing includes decoding, cropping, scaling, type conversion and normalization.

13. The performance evaluation method according to claim 1, wherein the vehicle-mounted chip comprises a vehicle-mounted system-on-chip (SoC) chip, and the method further comprises: Determine a comprehensive performance score of each of the plurality of vehicle-mounted SoC chips; as well as Based on the comprehensive performance score of each of the multiple vehicle-mounted SoC chips, at least one vehicle-mounted SoC chip among the multiple vehicle-mounted SoC chips is selected.

14. An electronic device comprising: processor, and A memory is coupled to the processor and stores instructions which, when executed by the processor, cause the device to perform the method according to claims 1 to 13.

15. A computer-readable storage medium having computer-executable instructions stored thereon, wherein when the computer-executable instructions are executed, the computer executes the steps of the performance evaluation method for an on-board chip according to any one of claims 1 to 13. 16 . A computer program product, which is tangibly stored on a non-volatile computer-readable medium and comprises machine-executable instructions, which when executed, cause a machine to perform the steps of the performance evaluation method for an on-vehicle chip according to any one of claims 1 to 13 .

Citation Information

Cited By

  • Vehicle-mounted application evaluation method and device and computer readable storage medium

    CN120104458A

  • In-vehicle Application Evaluation Method, Device, and Computer-Readable Storage Medium

    CN120104458B