Performance test method, device and system for computing equipment and electronic equipment

Abstract

The embodiment of the invention provides a performance test method, device and system for computing equipment and electronic equipment, and the method comprises the steps: obtaining test data of to-be-tested computing equipment; performance parameters of the computing device are detected according to the test data and the carbon emission parameters, the carbon emission parameters are used for indicating the carbon emission condition of resources consumed by the computing device, and the performance parameters are used for indicating the operation performance of the computing device under the unit carbon emission. Through the method and the device, the technical problem that the applicability of testing the performance of the computing equipment is relatively poor is solved, and the technical effect of improving the applicability of testing the performance of the computing equipment is achieved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computers, and more specifically, to a performance testing method, apparatus, system, and electronic device for a computing device. Background Art

[0002] Computing devices are being used more and more widely in current production and life. Currently, the testing methods for the operating performance of computing devices are relatively single-dimensional, and the computing power that can be measured by the computing devices is relatively limited, resulting in low performance applicability of the tested computing devices, which affects the design and planning of the application and deployment of computing devices.

[0003] With regard to the problems in related technologies such as poor applicability of testing the performance of computing devices, no effective solution has yet been proposed. Summary of the invention

[0004] Embodiments of the present application provide a method, apparatus, system, and electronic device for testing the performance of a computing device, so as to at least solve the problem of poor applicability of testing the performance of a computing device in the related art.

[0005] According to one embodiment of the present application, a performance testing method for a computing device is provided, comprising: obtaining test data of a computing device to be tested; detecting performance parameters of the computing device based on the test data and carbon emission parameters, wherein the carbon emission parameters are used to indicate the carbon emissions of resources consumed by the computing device, and the performance parameters are used to indicate the operating performance of the computing device under unit carbon emissions.

[0006] As an optional embodiment, detecting the performance parameter of the computing device according to the test data and the carbon emission parameter includes:

[0007] The first performance parameter of the computing device is detected according to the first performance data of the computing device, the first power consumption data of the computing device, the first test duration of the computing device, and the electric power carbon emission factor, wherein the test task for the computing device includes: executing a target workload under a target pressure value, the target pressure value is used to indicate the test pressure on the computing device, the first performance data is used to indicate the running performance of the computing device in executing the target workload under the target pressure value, the first power consumption data is used to indicate the running power consumption of the computing device in executing the target workload under the target pressure value, the electric power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the first test duration is the duration for the computing device to execute the target workload under the target pressure value, the first performance parameter is used to indicate the computing power provided by the computing device in executing the target workload under the target pressure value per unit of carbon emission, the test data includes: the first performance data, the first power consumption data, and the first test duration, the carbon emission parameter includes: the electric power carbon emission factor, and the performance parameter includes: the first performance parameter.

[0008] As an optional embodiment, detecting the first performance parameter of the computing device according to the first performance data of the computing device, the first power consumption data of the computing device, the first test duration of the computing device, and the power carbon emission factor includes:

[0009] The pressure load computing power and computing efficiency Eff of the computing device is calculated by the following formula load :

[0010]

[0011] Among them, the first performance parameter includes the pressure load computing power efficiency Eff load , Per is the first performance data, Pw is the first power consumption data, CEF is the electricity carbon emission factor, ΔT is the first test duration, ΔT=Ts-Tb, Ts is the test end time of executing the target workload under the target pressure value, and Tb is the test start time of executing the target workload under the target pressure value.

[0012] As an optional embodiment, detecting the performance parameter of the computing device according to the test data and the carbon emission parameter includes:

[0013] Detecting multiple second performance parameters of the computing device according to multiple second performance data of the computing device, multiple second power consumption data of the computing device, multiple second test durations of the computing device, and a power carbon emission factor, wherein the test task for the computing device includes: executing a reference workload under multiple pressure values, the multiple pressure values ​​are used to indicate multiple test pressures for the computing device, the second performance data are used to indicate the running performance of the computing device executing the reference workload under a single pressure value, the second power consumption data are used to indicate the running power consumption of the computing device executing the reference workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the second test duration is the duration of the computing device executing the reference workload under a single pressure value, the second performance parameter is used to indicate the computing power provided by the computing device executing the reference workload under a single pressure value under unit carbon emissions, the test data includes: multiple second performance data, multiple second power consumption data, and multiple second test durations, and the carbon emission parameter includes: the power carbon emission factor;

[0014] Convert multiple of the second performance parameters into a third performance parameter, wherein the third performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing the reference workload under unit carbon emissions, and the performance parameters include: the third performance parameter.

[0015] As an optional embodiment, converting the plurality of second performance parameters into a third performance parameter includes:

[0016] The workload computing power and computing efficiency Eff of the computing device is calculated by the following formula worklet :

[0017]

[0018] The third performance parameter includes the workload computing power efficiency Eff worklet , Eff loadi is the second performance parameter under the i-th pressure value, and n is the number of the multiple pressure values.

[0019] As an optional embodiment, detecting the performance parameter of the computing device according to the test data and the carbon emission parameter includes:

[0020] The plurality of fourth performance parameters of the computing device are detected according to the plurality of third performance data of the computing device, the plurality of third power consumption data of the computing device, the plurality of third test durations of the computing device, and the carbon emission factor of electricity, wherein the test task for the computing device comprises: executing the plurality of workloads in the target test sub-scenario under the corresponding pressure value, the corresponding pressure value is used to indicate the test pressure on the computing device in the corresponding workload, the target test sub-scenario is used to indicate the operation category allowed to run by the computing device, the plurality of workloads are used to indicate the plurality of running operations in the target test sub-scenario, and the third performance data is used to indicate the computing device under a single pressure The operating performance of executing a single workload under a single pressure value, the third power consumption data is used to indicate the operating power consumption of the computing device executing a single workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of power consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device under a single pressure value to execute a single workload per unit of carbon emission, the test data includes: a plurality of the third performance data, a plurality of the third power consumption data and a plurality of the third test durations, and the carbon emission parameters include: the power carbon emission factor;

[0021] Convert the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of the fifth performance parameters, wherein the fifth performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount;

[0022] Convert multiple fifth performance parameters into sixth performance parameters, wherein the sixth performance parameter is used to indicate the average computing power and efficiency provided by the computing device at unit carbon emissions in the target test sub-scenario, and the performance parameters include: the sixth performance parameter.

[0023] As an optional embodiment, converting the plurality of fifth performance parameters into a sixth performance parameter includes:

[0024] The scene computing power and computing efficiency Eff of the computing device is calculated by the following formula workload :

[0025]

[0026] Among them, the sixth performance parameter includes the scene computing power and efficiency Eff workload , Eff workletj is the fifth performance parameter under the j-th workload, and m is the number of the multiple workloads.

[0027] As an optional embodiment, the obtaining test data of the computing device to be tested includes:

[0028] Collecting a plurality of initial performance data of the computing device, wherein the initial performance data is used to indicate the actual operating performance of the computing device executing a single workload under a single pressure value;

[0029] The plurality of the initial performance data are normalized to obtain the plurality of the third performance data, wherein the third performance data are used to indicate the normalized operating performance of the computing device executing a single workload under a single pressure value.

[0030] As an optional embodiment, the normalizing the multiple initial performance data includes:

[0031] Extracting a target normalization factor corresponding to a single workload from workloads and normalization factors having a corresponding relationship, wherein the normalization factor is used to normalize performance data corresponding to each workload to the same data level;

[0032] The ratio of each of the initial performance data under a single workload to the target normalization factor is calculated to obtain each of the third performance data under a single workload.

[0033] As an optional embodiment, detecting the performance parameter of the computing device according to the test data and the carbon emission parameter includes:

[0034] Detecting multiple fourth performance parameters of the computing device according to multiple third performance data of the computing device, multiple third power consumption data of the computing device, multiple third test durations of the computing device, and a power carbon emission factor, wherein the test task for the computing device includes: executing corresponding workloads in multiple test sub-scenarios under corresponding pressure values, the corresponding pressure values ​​are used to indicate the test pressure on the computing device in the corresponding workload, the multiple test sub-scenarios are used to indicate multiple operation categories allowed to run by the computing device, the third performance data are used to indicate the running performance of the computing device when executing a single workload under a single pressure value, the third power consumption data are used to indicate the running power consumption of the computing device when executing a single workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device when executing a single workload under a single pressure value under a unit of carbon emission, the test data includes: multiple third performance data, multiple third power consumption data, and multiple third test durations, and the carbon emission parameter includes: the power carbon emission factor;

[0035] Convert the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of the fifth performance parameters, wherein the fifth performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount;

[0036] Converting a plurality of the fifth performance parameters in a single test sub-scenario into a seventh performance parameter to obtain a plurality of the seventh performance parameters, wherein the seventh performance parameter is used to indicate an average computing power and computing efficiency provided by the computing device under a unit carbon emission in the single test sub-scenario;

[0037] Convert multiple seventh performance parameters into an eighth performance parameter, wherein the eighth performance parameter is used to indicate the average computing power and efficiency provided by the computing device under unit carbon emissions, and the performance parameters include: the eighth performance parameter.

[0038] As an optional embodiment, converting the plurality of the seventh performance parameters into the eighth performance parameter comprises one of the following:

[0039] Calculating a weighted geometric mean of a plurality of the seventh performance parameters as the eighth performance parameter, wherein the weight corresponding to each of the test sub-scenario is determined according to the degree of influence of the running of the test sub-scenario on the computing power and efficiency of the computing device;

[0040] Calculating a weighted geometric mean of a plurality of the seventh performance parameters as the eighth performance parameter, wherein the plurality of test sub-scenarios are divided into a plurality of test scenarios, and the plurality of test scenarios include: basic computing and application performance, the basic computing is used to test the basic performance of the computing device outputting computing power, and the application performance is used to test the performance of the computing device running a user-side application, a first weight of the basic computing is less than a second weight of the application performance, the first weight is assigned to the test sub-scenario belonging to the basic computing, and the second weight is assigned to the test sub-scenario belonging to the application performance;

[0041] A geometric mean of a plurality of the seventh performance parameters is calculated as the eighth performance parameter.

[0042] As an optional embodiment, in the case where the multiple test sub-scenarios include: a basic computing class, the workload corresponding to the basic computing class includes at least one of the following: compression and decompression, encryption and decryption, hash conversion, matrix operation, linear equation, sorting algorithm, concurrent operation, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential read and write, storage random read and write, network bandwidth test, network delay test;

[0043] In the case where the plurality of test sub-scenarios include: big data, the workload corresponding to the big data includes at least one of the following: big data reading, big data writing, and sorting calculation;

[0044] In the case where the plurality of test sub-scenarios include: artificial intelligence AI, the workload corresponding to the artificial intelligence AI includes at least one of the following: graphic recognition, natural language processing;

[0045] In the case where the plurality of test sub-scenarios include: virtualization, the workload corresponding to the virtualization includes at least one of the following: virtualization platform basic performance, virtualization database performance;

[0046] In the case where the plurality of test sub-scenarios include: a database, the workload corresponding to the database includes: database performance.

[0047] As an optional embodiment, converting the plurality of the seventh performance parameters into an eighth performance parameter includes:

[0048] The computing power and computing efficiency (CPE) of the computing device is calculated by the following formula:

[0049] CPE=exp(k1×ln(Eff compute )+k2×ln(Eff bd )+k3×ln(Eff AI )+k4×ln(Eff VM )+k5×ln(Eff SQL ));

[0050] The eighth performance parameter includes the computing power and efficiency of the computing device CPE, and the multiple test sub-scenarios include: the basic computing class, the big data, the artificial intelligence AI, the virtualization and the database, Eff compute is the seventh performance parameter under the basic computing class, Eff bd is the seventh performance parameter under the big data, Eff AI is the seventh performance parameter under the artificial intelligence AI, Eff VM is the seventh performance parameter under the virtualization, Eff SQL is the seventh performance parameter under the database, and k1, k2, k3, k4, and k5 are weights corresponding to the respective test sub-scenarios.

[0051] As an optional embodiment, detecting the performance parameter of the computing device according to the test data and the carbon emission parameter includes:

[0052] The initial performance parameters of the computing device are calculated according to the target performance data, the target power consumption data, the target test duration and the electric power carbon emission factor, wherein the target performance data is used to indicate the operating performance of the computing device under the test task, the target power consumption data is used to indicate the operating power consumption of the computing device under the test task, the electric power carbon emission factor is used to indicate the carbon emission level of the unit power consumed by the computing device, the target test duration is the duration for the computing device to perform the test task, the initial performance parameters are used to indicate the computing power provided by the computing device under the test task, the test data includes: the target performance data, the target power consumption data and the target test duration, and the carbon emission parameters include: the electric power carbon emission factor;

[0053] Searching for a target parameter range into which the initial performance parameter falls from a plurality of parameter ranges, wherein the performance parameter values ​​of the computing device in the test task are divided into the plurality of parameter ranges to obtain parameter ranges and performance levels having corresponding relationships;

[0054] The target performance level corresponding to the target parameter range is determined as the computing power and efficiency level of the computing device, wherein the performance parameters include: the target performance level.

[0055] As an optional embodiment, the obtaining test data of the computing device to be tested includes:

[0056] Receive a first test request, wherein the first test request is used to request to perform a test of a first test task on the computing device;

[0057] In response to the first test request, storing first test data corresponding to the first test task to the computing device, wherein the first test data is used to be called by the computing device to implement the test of the first test task;

[0058] Initiating a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the first test task;

[0059] In the process of the computing device responding to the test instruction to execute the first test task, the performance data of the computing device, the power consumption data of the computing device and the test duration of the computing device are obtained as the test data, wherein the performance data is used to indicate the operating performance of the computing device in the first test task, the power consumption data is used to indicate the operating power consumption of the computing device in the first test task, and the test duration is the duration for the computing device to execute the first test task.

[0060] As an optional embodiment, the obtaining test data of the computing device to be tested includes:

[0061] receiving a second test request, wherein the second test request is used to request to perform a second test task on the computing device;

[0062] In response to the second test request, initiating a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the second test task;

[0063] Before the computing device executes the candidate workload in the second test task in response to the test instruction, storing second test data corresponding to the candidate workload in the computing device, wherein the second test data is used to be called by the computing device to implement the test of the candidate workload;

[0064] During the process of the computing device executing the candidate workload, performance data of the computing device, power consumption data of the computing device and test duration of the computing device are obtained as the test data, wherein the performance data is used to indicate the operating performance of the computing device under the candidate workload, the power consumption data is used to indicate the operating power consumption of the computing device under the candidate workload, and the test duration is the duration of the computing device executing the candidate workload.

[0065] According to another embodiment of the present application, a performance testing system for a computing device is provided, comprising: a controller, a collector and a power supply, wherein the controller and the collector are connected, and the controller, the collector and the power supply are all used to connect to the computing device to be tested; the power supply is used to supply power to the computing device; the collector is used to collect test data of the computing device; and the controller is used to execute the steps of the performance testing method for the computing device described in any of the preceding items.

[0066] Optionally, the collector includes: a target processor and a power meter, the power meter is connected between the power supply and the computing device, the controller and the target processor are deployed on a control machine, and the power meter is connected to the control machine;

[0067] The target processor is used to collect performance data and test duration in the test data, wherein the performance data is used to indicate the operating performance of the computing device in the target test task, and the test duration is the duration of the computing device executing the target test task;

[0068] The power meter is used to collect power consumption data in the test data, wherein the power consumption data is used to indicate the operating power consumption of the computing device in the target test task.

[0069] Optionally, in the case where the computing device includes a liquid-cooled computing device, the computing device includes: a test machine and a liquid cooling device, the power output end of the power supply is connected to the power input end of the power meter, and the power output end of the power meter is connected to the power input end of the test machine and the power input end of the liquid cooling device;

[0070] The power consumption data in the test data collected by the power meter includes power consumption parameters of the test machine and power consumption parameters of the liquid cooling device.

[0071] Optionally, the system further comprises: a thermometer, the thermometer is connected to the controller, and the computing device is deployed within a temperature detection range of the thermometer;

[0072] The thermometer is used to detect the ambient temperature of the computing device;

[0073] The controller is used to test the computing device when the ambient temperature falls within a target temperature range.

[0074] According to another embodiment of the present application, a performance testing device for a computing device is provided, comprising: an acquisition module for acquiring test data of a computing device to be tested; and a detection module for detecting performance parameters of the computing device based on the test data and carbon emission parameters, wherein the carbon emission parameters are used to indicate the carbon emissions of resources consumed by the computing device, and the performance parameters are used to indicate the operating performance of the computing device under unit carbon emissions.

[0075] According to another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when run.

[0076] According to another embodiment of the present application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0077] According to another embodiment of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0078] Through this application, test data is obtained from the computing device to be tested, and then the performance parameters of the computing device are detected by combining the test data and the carbon emission parameters. The carbon emission parameters can indicate the carbon emission of the resources consumed by the computing device. The obtained performance parameters can measure the operating performance of the computing device under the unit carbon emission, so that the operating performance of the computing device can be evaluated while also measuring the performance of the computing device in terms of carbon emissions, and then the impact of the computing device on the environment can be compared. Therefore, the technical problem of poor applicability of testing the performance of the computing device can be solved, and the technical effect of improving the applicability of testing the performance of the computing device can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a hardware structure block diagram of a server device of a performance testing method of a computing device according to an embodiment of the present application;

[0080] Figure 2 is a flow chart of a performance testing method for a computing device according to an embodiment of the present application;

[0081] Figure 3 is a schematic diagram of a performance testing system for a computing device according to an embodiment of the present application;

[0082] Figure 4 is an optional air-cooled server test environment framework diagram according to an embodiment of the present application;

[0083] Figure 5 is an optional liquid cooling server test environment framework diagram according to an embodiment of the present application;

[0084] Figure 6 It is a structural block diagram of a performance testing device for a computing device according to an embodiment of the present application. DETAILED DESCRIPTION

[0085] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0086] It should be noted that the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0087] The method embodiments provided in the embodiments of the present application can be executed in a server device or a similar computing device. Taking running on a server device as an example, Figure 1 1 is a hardware structure block diagram of a server device of a performance testing method for a computing device according to an embodiment of the present application. Figure 1 As shown, the server device may include one or more ( Figure 1Only one is shown in the figure) processor 102 (processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the server device may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above server device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0088] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the performance test method of the computing device in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the server device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0089] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the server device. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0090] For the convenience of describing the following terms involved in this application, the following terms and definitions established in GB / T 9813.3--2017 apply to this application:

[0091] Computing power: The comprehensive computing capability of a server when processing various types of computing tasks.

[0092] Computing power and efficiency: The energy consumption of a server when performing a specific computing task, specifically expressed as the amount of computing that can be completed per unit of carbon emissions.

[0093] The following abbreviations apply to this application:

[0094] CEF Carbon Emission Factor.

[0095] CPE computing power and efficiency (Computing Power and Efficiency).

[0096] In this embodiment, a performance testing method for a computing device is provided. Figure 2 is a flow chart of a method for testing the performance of a computing device according to an embodiment of the present application. Figure 2 As shown, the process includes the following steps:

[0097] Step S202, obtaining test data of the computing device to be tested;

[0098] Step S204, detecting the performance parameters of the computing device according to the test data and the carbon emission parameters, wherein the carbon emission parameters are used to indicate the carbon emission of the resources consumed by the computing device, and the performance parameters are used to indicate the operating performance of the computing device under unit carbon emission.

[0099] Through the above steps, test data is obtained from the computing device to be tested, and then the performance parameters of the computing device are detected by combining the test data and the carbon emission parameters. The carbon emission parameters can indicate the carbon emission of the resources consumed by the computing device. The obtained performance parameters can measure the operating performance of the computing device under the unit carbon emission, so that the operating performance of the computing device can be evaluated while also measuring the performance of the computing device in terms of carbon emissions, and then the impact of the computing device on the environment can be compared. Therefore, the technical problem of poor applicability of testing the performance of the computing device can be solved, and the technical effect of improving the applicability of testing the performance of the computing device can be achieved.

[0100] In the embodiment provided in the above step S202, the computing device is an electronic device capable of performing computing tasks, and the computing tasks may be, but are not limited to, mathematical calculations, logical judgments, storage, information processing, and other data processing-related tasks. In the process of performing computing tasks, the computing device can receive relevant data, and automatically process the data according to program settings and then output relevant computing results.

[0101] Optionally, in the embodiment of the present application, the resources consumed by the computing device may refer to any resources that can be used to measure the carbon emissions of the computing device while providing energy for the computing device, such as electrical energy and the like.

[0102] Optionally, in the embodiments of the present application, the computing device may include, but is not limited to, general-purpose computers (including personal computers, laptops, etc., which are widely used in the daily work of individuals and enterprises, such as document processing, data analysis and web browsing, etc.), mobile devices (such as smart phones, tablets, smart wearable devices, which are easy to carry and powerful, and can be used for mobile office, mobile computing and communication), servers (high-performance computers specially designed to provide computing services on the network, such as Web servers, database servers, etc., which can be used to perform complex computing tasks such as scientific computing, weather forecasting, data simulation, etc.), embedded devices (such devices are usually embedded in other products and equipment, and have specific computing functions and application scenarios), etc. With the development of science and technology, computing devices are constantly changing in performance, use and complexity, so this application does not limit the specific type of computing devices.

[0103] Optionally, in an embodiment of the present application, the test data is data collected during the process of controlling the computing device to run according to the test scenario indicated by the test task. The test data is used to indicate the device operation capability of the computing device in the corresponding test scenario. Therefore, in this embodiment, the test data may include but is not limited to performance data characterizing the business processing performance of the computing device in the corresponding test scenario, power consumption data characterizing the power consumption of the computing device in the corresponding test scenario, and test duration data characterizing the time consumption of the computing device in the corresponding test scenario.

[0104] In the embodiment provided in the above step S204, the carbon emission parameters are used to quantify the impact of the energy consumption of the computing device during operation on the environment. The carbon emission parameters may include but are not limited to carbon emission factors, total carbon emissions, etc., where the carbon emission factor is used to indicate the carbon emission level caused by the unit energy consumption of the computing device, and the total carbon emissions are the carbon emissions generated by all the energy consumed by the computing device to complete the test project.

[0105] Optionally, in an embodiment of the present application, the performance parameter is used to reflect the operating capability of the computing device under the test requirements indicated by the test task through the carbon emission factor, and the test task is used to indicate the complexity of the test items that need to be tested on the computing device and the test pressure required for the test items. The test items may include basic scenario projects and application scenario projects. The basic scenario projects are used to test the basic business functions of the computing device (the basic business functions are the business functions required to ensure the basic operation of the server). This type of project covers the testing of the computing performance, storage, network and other basic business functions of the computing device; the application scenario projects are used to test the business functions that the computing device has to meet the application scenario requirements of the user. This category covers special application scenarios such as big data distributed computing, massive data processing, AI (Artificial Intelligence) graphic recognition, natural language processing, AIGC (Artificial Intelligence Generated Content) large models, virtualization, cloud hosts, etc.

[0106] As an optional embodiment, detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes:

[0107] The first performance parameter of the computing device is detected according to the first performance data of the computing device, the first power consumption data of the computing device, the first test duration of the computing device and the electric power carbon emission factor, wherein the test task for the computing device includes: executing the target workload under the target pressure value, the target pressure value is used to indicate the test pressure on the computing device, the first performance data is used to indicate the running performance of the computing device in executing the target workload under the target pressure value, the first power consumption data is used to indicate the running power consumption of the computing device in executing the target workload under the target pressure value, the electric power carbon emission factor is used to indicate the carbon emission level per unit of electricity consumed by the computing device, the first test duration is the duration for the computing device to execute the target workload under the target pressure value, the first performance parameter is used to indicate the computing power provided by the computing device in executing the target workload under the target pressure value at the unit carbon emission amount, the test data includes: the first performance data, the first power consumption data and the first test duration, the carbon emission parameters include: the electric power carbon emission factor, and the performance parameters include: the first performance parameter.

[0108] Optionally, in an embodiment of the present application, the target workload is used to indicate the test operation of a business function to be tested among the business functions of the computing device. The computing device may have business functions of multiple business dimensions, and different target workloads are used to test business functions of different business dimensions. The target workload may include but is not limited to: compression and decompression, encryption and decryption, hash conversion, matrix operations, linear equations, sorting algorithms, concurrent operations, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential reading and writing, storage random reading and writing, network bandwidth test, network delay test, big data reading, big data writing, sorting calculation, graphic recognition, natural language processing, virtualization platform basic performance, virtualization database performance, and database performance.

[0109] Optionally, in an embodiment of the present application, during the test process, one or more test pressure values ​​may be configured for each target workload according to test requirements. The test pressure value may be, but is not limited to, any value greater than 0% and less than or equal to 100%. The target pressure value is the test pressure value currently used among the one or more test pressure values. Table 1 is an optional test pressure value setting table according to an embodiment of the present application, as shown in Table 1:

[0110] Table 1

[0111]

[0112] Table 1 exemplarily lists the test pressure values ​​configured for the projects corresponding to different target workloads. In addition to the test pressure values ​​listed in the above table, the pressure value corresponding to the target workload can be set according to the test requirements during the actual test process. This embodiment does not specifically limit the value of the test pressure value corresponding to the target workload.

[0113] Optionally, in an embodiment of the present application, the first performance data is used to indicate the operating performance of the computing device under the target workload, and the operating performance indicated by the first performance data corresponding to different workloads has different meanings. For example, when the target workload is concurrent operation, the operating performance of the first performance data indicates the number of concurrent operations per unit time of the computing device; when the target workload is graphic recognition, the operating performance of the first performance data indicates the number of images processed per unit time by the computing device; when the target workload is a network bandwidth test, the operating performance of the first performance data indicates the number of bytes transmitted per unit time by the computing device. This scheme does not limit this.

[0114] Optionally, in an embodiment of the present application, the value of the electricity carbon emission factor may refer to the value set in the relevant standards, for example, the value may be 0.5810 tCO2 / MWh.

[0115] Optionally, in an embodiment of the present application, the first performance parameter can be obtained by, but is not limited to, detecting first performance data, first power consumption data, first test duration, and electricity carbon emission factor using a target detection model, wherein the target detection model records the conversion relationship between performance data, power consumption data, test duration, electricity carbon emission factor and performance parameters.

[0116] Through the above content, by executing the target workload under the test pressure indicated by the target pressure value on the computing device, the first performance data, first power consumption data and first test duration of the computing device are calculated, and then the first performance data, first power consumption data, first test duration and electricity carbon emission factor are used to detect the first performance parameter indicating the computing power provided by the computing device to execute the target workload under the target pressure value, so as to quantify the computing power and computing efficiency provided by the computing device to execute a specific workload under a specific pressure value through the carbon emission factor.

[0117] As an optional embodiment, detecting a first performance parameter of the computing device according to first performance data of the computing device, first power consumption data of the computing device, a first test duration of the computing device, and a power carbon emission factor includes:

[0118] The pressure load computing power and computing efficiency Eff of the computing device is calculated by the following formula load :

[0119]

[0120] Among them, the first performance parameter includes pressure load computing power efficiency Eff load , Per is the first performance data, Pw is the first power consumption data, CEF is the electricity carbon emission factor, ΔT is the first test duration, ΔT=Ts-Tb, Ts is the test end time of executing the target workload under the target pressure value, and Tb is the test start time of executing the target workload under the target pressure value.

[0121] Optionally, in an embodiment of the present application, the pressure load computing power and computing efficiency indicate the volume of computing business indicated by the target workload that can be completed by the computing device under a single pressure value when executing a single target workload at a unit carbon emission.

[0122] As an optional embodiment, detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes:

[0123] Detect multiple second performance parameters of the computing device according to multiple second performance data of the computing device, multiple second power consumption data of the computing device, multiple second test durations of the computing device, and the power carbon emission factor, wherein the test task for the computing device includes: executing a reference workload under multiple pressure values, the multiple pressure values ​​are used to indicate multiple test pressures for the computing device, the second performance data are used to indicate the operating performance of the computing device when executing the reference workload under a single pressure value, the second power consumption data are used to indicate the operating power consumption of the computing device when executing the reference workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the second test duration is the duration of the computing device executing the reference workload under a single pressure value, the second performance parameter is used to indicate the computing power provided by the computing device when executing the reference workload under a single pressure value under unit carbon emissions, the test data includes: multiple second performance data, multiple second power consumption data, and multiple second test durations, and the carbon emission parameters include: power carbon emission factor;

[0124] Convert multiple second performance parameters into third performance parameters, wherein the third performance parameters are used to indicate the average computing power and efficiency provided by the computing device when executing a reference workload under unit carbon emissions, and the performance parameters include: the third performance parameter.

[0125] Optionally, in an embodiment of the present application, a reference workload is used to indicate the test operation of a business function to be tested among the business functions of the computing device. The computing device may have business functions of multiple business dimensions, and different reference workloads are used to test business functions of different business dimensions. The reference workload may include but is not limited to: compression and decompression, encryption and decryption, hash conversion, matrix operations, linear equations, sorting algorithms, concurrent operations, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential read and write, storage random read and write, network bandwidth test, network delay test, big data read, big data write, sorting calculation, graphic recognition, natural language processing, virtualization platform basic performance, virtualization database performance, and database performance.

[0126] Optionally, in an embodiment of the present application, one or more test stress values ​​may be configured for each reference workload according to test requirements during the test process. When it is necessary to test the average computing power and efficiency (i.e., the third performance parameter) of the computing device under multiple stress values ​​under a reference workload, the reference workload may be configured with multiple test stress values ​​according to the test scenario requirements. Table 2 is a table of stress value quantity configuration according to an embodiment of the present application, as shown in Table 2:

[0127] Table 2

[0128]

[0129] In Table 2, n is the number of pressure value configurations of the corresponding reference workload. The pressure value of each reference workload can be configured with reference to the test pressure value configuration example of the workload in Table 1, or other pressure values ​​can be set according to actual test requirements. This embodiment does not specifically limit the number and value of the test pressure values ​​corresponding to each reference workload.

[0130] Optionally, in the embodiment of the present application, the computing efficiency Eff of the computing device indicated by the second performance parameter executing the reference workload at the i-th pressure value is calculated by the following formula: loadi Perform the calculation:

[0131]

[0132] Among them, Peri is the second performance data under the i-th pressure value of the reference workload, Pwi is the second power consumption data under the i-th pressure value of the reference workload, CEF is the electricity carbon emission factor, and ΔTi is the second test duration under the i-th pressure value of the reference workload.

[0133] Optionally, in an embodiment of the present application, the third performance parameter may be, but is not limited to, obtained by calculating the weighted geometric mean of multiple second performance parameters, and the weight corresponding to each pressure value may be determined based on the degree of influence of the pressure value on the computing power and efficiency of the computing device, or the third performance parameter may also be obtained by calculating the geometric mean of multiple second performance parameters.

[0134] Through the above content, by testing the computing device under multiple pressure values ​​of the reference workload, multiple second performance data, multiple second power consumption data and multiple second test durations are obtained, and then the second performance data, second power consumption data, second test duration and electricity carbon emission factor are used to detect the second performance parameter indicating the computing power provided by the computing device to execute the reference workload under a single pressure value, so as to quantify the computing power and efficiency provided by the computing device to execute the reference workload under a single pressure value through the carbon emission factor, and then the average computing power and efficiency of the computing device to execute the reference workload under multiple pressure values ​​can be quantified through the carbon emission factor.

[0135] As an optional embodiment, converting the plurality of second performance parameters into a third performance parameter includes:

[0136] The workload computing power and computing efficiency Eff of the computing device is calculated by the following formula worklet :

[0137]

[0138] Among them, the third performance parameter includes workload computing power and efficiency Eff worklet , Effloadi is the second performance parameter under the i-th pressure value, and n is the number of the multiple pressure values.

[0139] As an optional embodiment, detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes:

[0140] Detect multiple fourth performance parameters of the computing device according to multiple third performance data of the computing device, multiple third power consumption data of the computing device, multiple third test durations of the computing device, and the electric power carbon emission factor, wherein the test task for the computing device includes: executing multiple workloads in the target test sub-scenario under the corresponding pressure value, the corresponding pressure value is used to indicate the test pressure on the computing device in the corresponding workload, the target test sub-scenario is used to indicate the operation category allowed to run by the computing device, and the multiple workloads are used to indicate multiple running operations in the target test sub-scenario, the third performance data is used to indicate the running performance of the computing device when executing a single workload under a single pressure value, the third power consumption data is used to indicate the running power consumption of the computing device when executing a single workload under a single pressure value, the electric power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device when executing a single workload under a single pressure value under unit carbon emission, the test data includes: multiple third performance data, multiple third power consumption data, and multiple third test durations, and the carbon emission parameters include: electric power carbon emission factor;

[0141] Converting the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of fifth performance parameters, wherein the fifth performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount;

[0142] Convert multiple fifth performance parameters into sixth performance parameters, wherein the sixth performance parameter is used to indicate the average computing power and efficiency provided by the computing device at unit carbon emissions in the target test sub-scenario, and the performance parameters include: the sixth performance parameter.

[0143] Optionally, in an embodiment of the present application, the workload is used to indicate the test operation of the business function to be tested among the business functions of the computing device. The computing device may have business functions of multiple business dimensions, and different workloads are used to test business functions of different business dimensions. The workload may include but is not limited to: compression and decompression, encryption and decryption, hash conversion, matrix operations, linear equations, sorting algorithms, concurrent operations, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential reading and writing, storage random reading and writing, network bandwidth test, network delay test, big data reading, big data writing, sorting calculation, graphic recognition, natural language processing, virtualization platform basic performance, virtualization database performance, database performance.

[0144] Optionally, in an embodiment of the present application, the workload may be divided into a plurality of test sub-scenarios according to the operation category of the workload. Table 3 is an optional test sub-scenario and workload mapping table according to the present embodiment, as shown in Table 3;

[0145] Table 3

[0146]

[0147] As shown in Table 3, the test sub-scenario can be divided into basic computing, big data, artificial intelligence AI, virtualization and database according to the operation category of the workload. The sub-scenario categories divided in this embodiment are only optional examples. In addition to the test sub-scenario categories listed in the table, other test sub-scenario categories can also be included. This solution does not limit the specific types and quantities of sub-scenario categories. In actual testing, the types and quantities of workloads included in each sub-scenario can be configured according to test needs. The m value in the table is the number of workloads configured for each test sub-scenario. The types and quantities of workloads configured for each test sub-scenario in this embodiment are only optional examples. In actual testing, the types and specific quantities of workloads included can also be flexibly configured for each test sub-scenario.

[0148] Optionally, in the embodiment of the present application, the computing power and computing efficiency Eff provided by the computing device indicated by the fourth performance parameter when executing a single workload under a single pressure value is calculated by the following formula: load Perform the calculation:

[0149]

[0150] Among them, Per is the third performance data, Pw is the third power consumption data, CEF is the electricity carbon emission factor, ΔT is the third test duration, ΔT=Ts-Tb, Ts is the test end time of executing a single workload under a single pressure value, and Tb is the test start time of executing a single workload under a single pressure value.

[0151] Optionally, in the embodiment of the present application, the computing power and computing efficiency Eff provided by the computing device indicated by the fifth performance parameter under a single workload is calculated by the following formula: worklet Perform calculations;

[0152]

[0153] Among them, Eff loadi is the four performance parameters under the i-th pressure value of executing a single pressure load, and n is the number of the multiple pressure values.

[0154] Optionally, in an embodiment of the present application, the sixth performance parameter may be, but is not limited to, obtained by calculating the weighted geometric mean of multiple fifth performance parameters. The weight corresponding to each workload may be determined based on the degree of influence of the workload on the computing power and efficiency of the computing device in the test sub-scenario, or may also be obtained by calculating the geometric mean of multiple fifth performance parameters.

[0155] Through the above content, by combining the performance data (third performance data and third power consumption data) and test duration (third test duration) of the computing device, as well as the electricity carbon emission factor, this method can comprehensively evaluate the performance and energy efficiency of the computing device under different workloads. By converting the fourth performance parameter (i.e., computing power per unit carbon emission), the performance of the computing device in terms of environmental impact can be more accurately measured. Further converting multiple fifth performance parameters into sixth performance parameters can evaluate the overall energy efficiency and carbon emission efficiency of the computing device in a specific test sub-scenario, which helps to optimize the energy consumption of the computing device and reduce carbon emissions.

[0156] As an optional embodiment, converting a plurality of fifth performance parameters into a sixth performance parameter includes:

[0157] The scene computing power and computing efficiency Eff of the computing device is calculated by the following formula workload ;

[0158]

[0159] Among them, the sixth performance parameter includes scene computing power and efficiency Eff workload , Eff workletj is the fifth performance parameter under the j-th workload, and m is the number of the multiple workloads.

[0160] As an optional embodiment, obtaining test data of a computing device to be tested includes:

[0161] Collecting a plurality of initial performance data of the computing device, wherein the initial performance data is used to indicate the actual operating performance of the computing device executing a single workload under a single pressure value;

[0162] The multiple initial performance data are normalized to obtain multiple third performance data, wherein the third performance data is used to indicate the normalized operating performance of the computing device executing a single workload under a single pressure value.

[0163] Optionally, in an embodiment of the present application, multiple initial performance data represent the operating performance of a computing device under different workloads. Different workloads indicate different operating operations of the computing device in a test scenario. Therefore, the operating performance of different workloads has different meanings. Therefore, multiple initial performance data belong to different data magnitudes. Then, the multiple initial performance data can be normalized so that the normalized multiple third performance data belong to the same data magnitude, and the indicated operating performance has a unified meaning.

[0164] Through the above content, by performing normalization operations on multiple initial performance data of different magnitudes, multiple third performance data after normalization are made to belong to the same magnitude, the indicated operating performance has a unified meaning, and the performance values ​​under different workloads are roughly pulled to a closer scale, thereby ensuring the accuracy of the energy efficiency calculation results.

[0165] As an optional embodiment, normalizing the multiple initial performance data includes:

[0166] Extracting a target normalization factor corresponding to a single workload from workloads and normalization factors having a corresponding relationship, wherein the normalization factor is used to normalize performance data corresponding to each workload to the same data level;

[0167] The ratio of each initial performance data under a single workload to the target normalization factor is calculated to obtain each third performance data under the single workload.

[0168] Optionally, in an embodiment of the present application, different workloads correspond to different normalization factors. Table 4 is an optional normalization factor configuration table according to this embodiment, as shown in Table 4:

[0169] Table 4

[0170] Workload Pressure value Initial performance parameters Normalization factor Compression and decompression 1 00％ 1 75390.75 10 Encryption and Decryption 1 00％ 2142967.888 1 00 Memory cache test 1 00％ 4673471 1.078 1 000 Concurrent Operations 1 00％ 8989469.464 200

[0171] As shown in Table 4, the quantity levels of the initial performance parameters obtained by executing different workloads under the same pressure value are different. Therefore, different workloads are configured with different normalization factors, so that the performance data corresponding to each workload is normalized to the same data level.

[0172] As an optional embodiment, detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes:

[0173] Detect multiple fourth performance parameters of the computing device according to multiple third performance data of the computing device, multiple third power consumption data of the computing device, multiple third test durations of the computing device, and the electric power carbon emission factor, wherein the test task for the computing device includes: executing corresponding workloads in multiple test sub-scenarios under corresponding pressure values, the corresponding pressure values ​​are used to indicate the test pressure on the computing device in the corresponding workload, the multiple test sub-scenarios are used to indicate multiple operation categories allowed to run by the computing device, the third performance data are used to indicate the running performance of the computing device when executing a single workload under a single pressure value, the third power consumption data are used to indicate the running power consumption of the computing device when executing a single workload under a single pressure value, the electric power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device when executing a single workload under a single pressure value under unit carbon emission, the test data includes: multiple third performance data, multiple third power consumption data, and multiple third test durations, and the carbon emission parameters include: electric power carbon emission factor;

[0174] Converting the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of fifth performance parameters, wherein the fifth performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount;

[0175] Converting multiple fifth performance parameters in a single test sub-scenario into seventh performance parameters to obtain multiple seventh performance parameters, wherein the seventh performance parameter is used to indicate the average computing power and computing efficiency provided by the computing device under unit carbon emissions in the single test sub-scenario;

[0176] Convert multiple seventh performance parameters into an eighth performance parameter, wherein the eighth performance parameter is used to indicate the average computing power and efficiency provided by the computing device under unit carbon emissions, and the performance parameters include: the eighth performance parameter.

[0177] Optionally, in the embodiment of the present application, the eighth performance parameter may be obtained by calculating a weighted geometric mean of multiple seventh performance parameters. For example, the eighth performance parameter CPE may be obtained by calculating multiple seventh performance parameters using the following formula:

[0178] CPE = exp(0.2 × ln(Eff cpmpute )+0.20×ln(Eff BD )+0.20×ln(Eff AI )+0.20×ln(EffVM )+0.20×ln(Eff SQL ))

[0179] Among them, Eff cpmpute The seventh performance parameter corresponding to the basic computing class, Eff IO Eff is the seventh performance parameter corresponding to big data. AI Eff is the seventh performance parameter corresponding to artificial intelligence AI. VM The seventh performance parameter corresponding to virtualization, Eff SQL The seventh performance parameter corresponding to the database, in this embodiment, the weight value of the seventh performance parameter corresponding to each test sub-scenario is configured to be 0.2, and in actual applications, the weight can also be configured according to the degree of influence of the operation of each test sub-scenario on the computing power and efficiency of the computing device. This solution does not limit this.

[0180] Optionally, in the embodiment of the present application, a geometric mean value of multiple seventh performance parameters may be calculated by the following formula to obtain an eighth performance parameter:

[0181] CPE=exp(ln(Eff cpmpute )+ln(Eff BD )+ln(Eff AI )+ln(Eff VM )+ln(Eff SQL ))

[0182] As an optional embodiment, converting a plurality of seventh performance parameters into an eighth performance parameter includes one of the following:

[0183] Calculate a weighted geometric mean of the plurality of seventh performance parameters as an eighth performance parameter, wherein the weight corresponding to each test sub-scenario is determined according to the degree of influence of the operation of the test sub-scenario on the computing power and efficiency of the computing device;

[0184] Calculating a weighted geometric mean of multiple seventh performance parameters as an eighth performance parameter, wherein the multiple test sub-scenarios are divided into multiple test scenarios, and the multiple test scenarios include: basic computing and application performance, the basic computing is used to test the basic performance of the computing device output computing power, and the application performance is used to test the performance of the computing device running a user-side application, the first weight of the basic computing is less than the second weight of the application performance, the first weight is assigned to the test sub-scenario belonging to the basic computing, and the second weight is assigned to the test sub-scenario belonging to the application performance;

[0185] A geometric mean of the plurality of seventh performance parameters is calculated as an eighth performance parameter.

[0186] Optionally, in the embodiment of the present application, the purpose of the performance test of the computing device is to enable the computing device to have better performance in the real application scenario to meet the computing requirements in different application scenarios. Therefore, the multiple test sub-scenarios of the computing device can be divided into multiple test scenarios according to the requirements of the application scenario, namely basic computing and application performance. Basic computing can include but is not limited to the basic performance of computing devices such as computing performance, storage, and network. Application performance can include but is not limited to the performance of common applications on the current user side such as big data distributed computing, massive data processing, AI graphics recognition, natural language processing, AIGC large model, virtualization, and cloud hosting. Table 5 is an optional test scenario configuration table according to the embodiment of the present application, as shown in Table 5:

[0187] Table 5

[0188]

[0189] Table 5 exemplarily lists the optional attribution relationships between test scenarios, test sub-scenarios, and workloads.

[0190] As an optional embodiment, in the case where the multiple test sub-scenarios include: a basic computing class, the workload corresponding to the basic computing class includes at least one of the following: compression and decompression, encryption and decryption, hash conversion, matrix operation, linear equation, sorting algorithm, concurrent operation, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential read and write, storage random read and write, network bandwidth test, network delay test;

[0191] In the case where the multiple test sub-scenarios include: In the case of big data, the workload corresponding to the big data includes at least one of the following: big data reading, big data writing, and sorting calculation;

[0192] In the case where the multiple test sub-scenarios include: artificial intelligence AI, the workload corresponding to the artificial intelligence AI includes at least one of the following: graphic recognition, natural language processing;

[0193] The multiple test sub-scenarios include: in the case of virtualization, the workload corresponding to the virtualization includes at least one of the following: virtualization platform basic performance, virtualization database performance;

[0194] In the case where the multiple test sub-scenarios include: database, the workload corresponding to the database includes: database performance.

[0195] Optionally, in the embodiments of the present application, with the diversification of computing needs, modern computing equipment (in this embodiment, server distance) not only needs to obtain good test data in traditional performance evaluation, but also needs to have good performance in the user's real application scenarios to meet the computing needs in different application scenarios. The embodiments of the present application cover the following two main types of computing power:

[0196] Basic performance computing power: mainly includes the basic performance of server output computing power such as computing performance, storage, and network;

[0197] Application performance computing power: covers the computing power performance of common user-side applications such as big data distributed computing, massive data processing, AI graphic recognition, natural language processing, AIGC large models, virtualization, and cloud hosting.

[0198] With the rise of green computing and the increase in energy costs, computing efficiency has become an important indicator for server performance evaluation. This document measures the computing efficiency of servers through the following aspects:

[0199] Standby power consumption: the energy consumption of the server in idle state;

[0200] Working power consumption: the energy consumption performance of the server under a specific workload;

[0201] Energy efficiency ratio: The ratio of a server's computing power to its energy consumption when performing a specific task, usually measured by performance indicators per watt (W) (such as the number of floating-point operations per watt).

[0202] As an optional embodiment, converting a plurality of seventh performance parameters into an eighth performance parameter includes:

[0203] The computing power and computing efficiency (CPE) of the computing device is calculated using the following formula:

[0204] CPE=exp(k1×ln(Eff compute )+k2×ln(Eff bd )+k3×ln(Eff AI )+k4×ln(Eff VM )+k5×ln(Eff SQL ));

[0205] Among them, the eighth performance parameter includes computing power and efficiency of computing equipment CPE, and multiple test sub-scenarios include: basic computing, big data, artificial intelligence AI, virtualization and database, Eff compute The seventh performance parameter under the basic computing category, Eff bd Eff is the seventh performance parameter under big data. AI Eff is the seventh performance parameter of artificial intelligence AI. VM It is the seventh performance parameter under virtualization, Eff SQL is the seventh performance parameter under the database, and k1, k2, k3, k4, and k5 are the weights corresponding to each test sub-scenario.

[0206] Optionally, in an embodiment of the present application, the weight value corresponding to each test sub-scenario can be determined based on the degree of influence of the operation of the test sub-scenario on the computing power and efficiency of the computing device, or the weight value of each test sub-scenario can be set to 1, that is, the geometric mean of multiple seventh performance parameters is calculated.

[0207] Through the above content, by performing weighted geometric mean calculation on multiple seventh performance parameters, the impact of each test scenario on the computing power and efficiency of the computing device can be truly reflected, thereby improving the accuracy of the eighth performance parameter.

[0208] As an optional embodiment, detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes:

[0209] Calculate the initial performance parameters of the computing device according to the target performance data, the target power consumption data, the target test time and the electric power carbon emission factor, wherein the target performance data is used to indicate the operating performance of the computing device under the test task, the target power consumption data is used to indicate the operating power consumption of the computing device under the test task, the electric power carbon emission factor is used to indicate the carbon emission level per unit of electricity consumed by the computing device, the target test time is the time for the computing device to perform the test task, the initial performance parameters are used to indicate the computing power provided by the computing device under the test task, the test data includes: the target performance data, the target power consumption data and the target test time, and the carbon emission parameters include: the electric power carbon emission factor;

[0210] Searching for a target parameter range into which the initial performance parameter falls from a plurality of parameter ranges, wherein the performance parameter values ​​of the computing device in the test task are divided into a plurality of parameter ranges to obtain parameter ranges and performance levels having a corresponding relationship;

[0211] The target performance level corresponding to the target parameter range is determined as the computing power and efficiency level of the computing device, wherein the performance parameters include: the target performance level.

[0212] Optionally, in an embodiment of the present application, the performance parameter values ​​of the computing device in the test task are divided into multiple parameter ranges, so that the performance level of the computing device can be determined according to the logical parameter range of the performance parameter of the computing device. Table 6 is an optional table of computing power and efficiency levels according to this embodiment, as shown in Table 6:

[0213] Table 6

[0214] Computing power and efficiency level Level 1 Level 2 Level 3 Level 4 Level 5 CPE 50 40 30 20 1 0

[0215] The computing power and efficiency level of a computing device can be divided into, but is not limited to, 5 levels, with level 1 being the highest.

[0216] As an optional embodiment, obtaining test data of a computing device to be tested includes:

[0217] Receiving a first test request, wherein the first test request is used to request a computing device to be tested for a first test task;

[0218] In response to the first test request, storing first test data corresponding to the first test task in the computing device, wherein the first test data is used to be called by the computing device to implement the test of the first test task;

[0219] Initiating a test instruction to a computing device, wherein the test instruction is used to instruct the computing device to execute a first test task;

[0220] During the process of the computing device responding to the test instruction to execute the first test task, the performance data of the computing device, the power consumption data of the computing device and the test duration of the computing device are obtained as test data, wherein the performance data is used to indicate the operating performance of the computing device in the first test task, the power consumption data is used to indicate the operating power consumption of the computing device in the first test task, and the test duration is the duration for the computing device to execute the first test task.

[0221] Optionally, in an embodiment of the present application, the first test data is the test data required for the computing device to execute the first test task. If the workload indicated by the first test task is different, the first test data used will be different. For example, when the workload is encryption and decryption, the first test data is a data file to be encrypted or decrypted. When the workload is natural speech processing, the first test data is a natural language processing model to be run and sample data recognized by the natural language model. This solution does not specifically limit the data type of the first test data.

[0222] Through the embodiment of the present application, after receiving the first test request, the first test data corresponding to the first test task is stored in the computing device in response to the first test request, so as to realize the pre-storage of the first test data corresponding to the first test task to be executed in the computing device, and then after receiving the test instruction subsequently, the first test data can be directly retrieved to execute the first test task, thereby improving the execution efficiency of the first test task and avoiding test waiting caused by data problems.

[0223] As an optional embodiment, obtaining test data of a computing device to be tested includes:

[0224] receiving a second test request, wherein the second test request is used to request a second test task to be performed on the computing device;

[0225] In response to the second test request, initiating a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the second test task;

[0226] Before the computing device executes the candidate workload in the second test task in response to the test instruction, storing second test data corresponding to the candidate workload in the computing device, wherein the second test data is used to be called by the computing device to implement the test of the candidate workload;

[0227] During the process of the computing device executing the candidate workload, the performance data of the computing device, the power consumption data of the computing device and the test duration of the computing device are obtained as test data, wherein the performance data is used to indicate the operating performance of the computing device under the candidate workload, the power consumption data is used to indicate the operating power consumption of the computing device under the candidate workload, and the test duration is the duration of the computing device executing the candidate workload.

[0228] Optionally, in an embodiment of the present application, the second test data is the test data required for the computing device to perform the second test task. Different candidate workloads use different second test data. For example, when the workload is encryption and decryption, the second test data is a data file to be encrypted or decrypted. When the workload is natural speech processing, the second test data is a natural language processing model to be run and sample data recognized by the natural language model. This solution does not specifically limit the data type of the first test data.

[0229] Through the above content, the second test data is stored on the computing device after receiving the execution instruction instructing to execute the second test task, thereby avoiding the computing device from storing the test data used for the test task to be executed in advance, thereby saving the storage space of the computing device and avoiding energy consumption problems caused by storing test data in advance.

[0230] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0231] In this embodiment, a performance testing system for a computing device is also provided. Figure 3 is a schematic diagram of a performance testing system for a computing device according to an embodiment of the present application, such as Figure 3 As shown, the system includes:

[0232] It includes: a controller, a collector and a power supply, wherein the controller and the collector are connected, and the controller, the collector and the power supply are all used to connect to the computing device to be tested;

[0233] Power supplies, used to power computing devices;

[0234] A collector, used for collecting test data of computing devices;

[0235] A controller is used to execute the steps of the performance testing method of the computing device in any of the aforementioned embodiments.

[0236] Through the above content, test data is obtained from the computing device to be tested, and then the performance parameters of the computing device are detected by combining the test data and the carbon emission parameters. The carbon emission parameters can indicate the carbon emission of the resources consumed by the computing device. The obtained performance parameters can measure the operating performance of the computing device under the unit carbon emission, so that the operating performance of the computing device can be evaluated while also measuring the performance of the computing device in terms of carbon emissions, and then the impact of the computing device on the environment can be compared. Therefore, the technical problem of poor applicability of testing the performance of the computing device can be solved, and the technical effect of improving the applicability of testing the performance of the computing device can be achieved.

[0237] As an optional embodiment, the collector includes: a target processor and a power meter, the power meter is connected between the power supply and the computing device, the controller and the target processor are deployed on a control machine, and the power meter is connected to the control machine;

[0238] A target processor, used to collect performance data and test duration in the test data, wherein the performance data is used to indicate the operating performance of the computing device in the target test task, and the test duration is the duration of the computing device executing the target test task;

[0239] A power meter is used to collect power consumption data in the test data, wherein the power consumption data is used to indicate the operating power consumption of the computing device in the target test task.

[0240] As an optional embodiment, in the case where the computing device includes a liquid-cooled computing device, the computing device includes: a test machine and a liquid cooling device, the power output end of the power supply is connected to the power input end of the power meter, and the power output end of the power meter is connected to the power input end of the test machine and the power input end of the liquid cooling device;

[0241] The power consumption data in the test data collected by the power meter includes the power consumption parameters of the test machine and the power consumption parameters of the liquid cooling device.

[0242] As an optional embodiment, the system further includes: a thermometer, the thermometer is connected to the controller, and the computing device is deployed within a temperature detection range of the thermometer;

[0243] A thermometer, used to detect the ambient temperature of the computing device;

[0244] A controller is used to test a computing device when an ambient temperature falls within a target temperature range.

[0245] In view of this, the present application proposes a server computing power and efficiency evaluation test method. This embodiment takes the computing device as a server device as an example, comprehensively measures the energy efficiency performance of the server in basic performance and application scenarios, and combines the carbon emission factor to evaluate the computing power provided by the server per unit carbon emissions, so as to accurately and objectively evaluate its computing power and efficiency.

[0246] To solve the above technical problems, the evaluation test process proposed in this application includes the following steps: 1) prepare the test environment; 2) prepare the test instrument; 3) set up the server; 4) prepare the test framework; 5) calibrate the equipment; 6) select the test load; 7) run the load and collect data; 8) calculate the server computing power and efficiency value;

[0247] When selecting the test load, this application comprehensively considers the basic performance and application performance of the server. It not only covers basic scenarios such as computing performance, storage capacity, and network performance, but also includes the computing power and efficiency of various application scenarios such as big data distributed computing, AI image recognition, natural language processing, virtualization, and databases. Through the comprehensive evaluation of these diverse test scenarios, the server computing power and efficiency value is calculated.

[0248] In the step of calculating the server computing power and computing efficiency, this application redefines the concept of computing power and computing efficiency. By introducing the carbon emission factor, computing power and computing efficiency are redefined as the comprehensive performance under unit carbon emissions, thereby more comprehensively reflecting the impact of the server on environmental sustainability.

[0249] 1) Prepare the test environment:

[0250] The test environment requirements are as follows:

[0251] The ambient temperature should be maintained at (25±5)℃, the relative humidity should be 45%~75%, and the atmospheric pressure should be 86kPa~106kPa;

[0252] For equipment with a nominal power less than 1.5kW, the test power supply should be AC ​​(220±1%) V; for equipment with a nominal power greater than 1.5kW, the test power supply should be AC ​​(220±4%) V;

[0253] For equipment with a nominal power less than 1.5kW, the total harmonic distortion of the test power supply should not exceed 2%; for equipment with a nominal power greater than 1.5kW, the total harmonic distortion of the test power supply should not exceed 5%;

[0254] The power input frequency should be (50±1%) Hz;

[0255] The test environment should avoid interference factors such as strong magnetic fields and strong vibrations.

[0256] 2) Prepare the test equipment:

[0257] The test equipment requirements are as follows:

[0258] The power meter should provide power measurements with an overall accuracy of not less than 1% and a minimum frequency response of 3.0kHz;

[0259] The accuracy of the temperature sensor should be ±0.5℃ or better;

[0260] The power meter should be directly connected between the AC power supply and the server under test, and uninterruptible power supply (UPS) equipment should not be used;

[0261] The data output interfaces of the power meter and temperature sensor should be connected to the corresponding input interfaces of the control machine, and the power data and temperature data should be automatically recorded by the control machine.

[0262] 3) Set up the server:

[0263] The server setup requirements are as follows:

[0264] The server under test should include CPU, memory, hard disk and operating system;

[0265] The server should use a standard power supply and be connected to an AC power source during the test;

[0266] The server and the control machine should be connected directly with a network cable or through a network switch to achieve network communication;

[0267] All software settings should be in their default state, and power management and power saving functions should remain enabled by default;

[0268] Testing should be performed using the operating system declared by the manufacturer;

[0269] 4) Prepare the test framework:

[0270] Figure 4 is an optional air-cooled server test environment framework diagram according to an embodiment of the present application, such as Figure 4 As shown in the figure, the test environment of the air-cooled server consists of a control machine, a test machine, a power meter, and a thermometer. The control machine is connected to the test machine through the network, and the power meter is connected to the external power input and supplies power to the test machine through its power output. At the same time, the control machine is connected to the power meter and thermometer through the USB port to collect test data.

[0271] Figure 5is an optional liquid cooling server test environment framework diagram according to an embodiment of the present application, such as Figure 5 As shown in the figure, the specific configuration is: the control machine is connected to the test machine through the network, the power meter is connected to the external power input, and the power supplies of the test machine and the liquid cooling device are connected to the power output of the power meter. The power consumption of the liquid cooling device should be included in the test power consumption data. The control machine is also connected to the power meter and the thermometer through the USB port for data collection.

[0272] 5) Calibration equipment:

[0273] Before use, all test equipment must be strictly calibrated to ensure the accuracy of the test data. Calibration should comply with relevant standards. Before each test, the calibration status of the equipment must be checked and recalibrated if necessary.

[0274] 6) Select the test load:

[0275] As an optional method, the workload corresponding to each test scenario is shown in Table 7.

[0276] Table 7

[0277]

[0278]

[0279] 7) Run load and collect data:

[0280] The server is tested under the load pressure set in Table 7, and the power consumption is measured using a calibrated power meter. During the test, standard test equipment should be used and environmental conditions should be recorded. Each test is repeated at least three times to ensure the reliability of the data.

[0281] The process of calculating the server computing power and computing efficiency value may include but is not limited to the following processes:

[0282] First, calculate the unit carbon emission computing power under a single workload and a single pressure value;

[0283] In the server computing power and efficiency test, except for the standby power consumption test, all loads will combine the performance value, power consumption value, carbon emission factor and test duration to calculate the unit carbon emission computing power Effload, calculated according to formula (1):

[0284]

[0285] Where: Per is the normalized performance data of each load at each pressure value; Pw is the power consumption of the whole machine at this pressure value, in watts (W); CEF is the carbon emission factor, which indicates the carbon emission level per unit of electricity, for example, it can be 0.581 0 tons of carbon dioxide / MWh (tCO2 / MWh); ΔT is the test time of a single load and a single pressure value, calculated as ΔT=T s -T b , where T b and Ts are the test start and end time, respectively.

[0286] Second, calculate the average computing power and efficiency under multiple pressure values ​​for a single workload;

[0287] The computing efficiency of each workload, Effworklet, is calculated by the geometric mean of the unit carbon emission computing power under each pressure value, according to formula (2):

[0288]

[0289] Wherein: i represents the pressure value corresponding to the workload; n is the number of pressure values ​​under the workload, and the values ​​of the number of pressure values ​​n corresponding to each workload are shown in Table 8.

[0290] Table 8

[0291]

[0292] Third, calculate the computing power and efficiency of a single test scenario;

[0293] Computing efficiency Eff for each test scenario workload It is calculated by taking the geometric mean of the computing power and efficiency of each workload, according to formula (3):

[0294]

[0295] Wherein: i represents the i-th workload in the test scenario (equivalent to j above); m is the number of workloads in the test scenario, and the values ​​of the number of workloads m corresponding to each test scenario are shown in Table 9.

[0296] Table 9

[0297]

[0298] Fourth, calculate the total computing power and efficiency of the server;

[0299] The server computing power and computing efficiency CPE is calculated by the weighted geometric mean of each test scenario according to formula (4):

[0300] CPE = exp(0.2 × ln(Eff cpmpute)+0.20×ln(Eff BD )+0.20×ln(Eff AI )+0.20×ln(Eff VM )+0.20×ln(Eff SQL )) (4)

[0301] The unit of CPE is defined as computing power divided by carbon emissions. The higher the relative value of the data, the higher the computing power that the test machine can output under unit carbon emissions.

[0302] Alternatively, the server computing power and computing efficiency (CPE) can also be calculated by the geometric mean of each test scenario:

[0303] CPE=exp(ln(Eff cpmpute )+ln(Eff BD )+ln(Eff AI )+ln(Eff VM )+ln(Eff SQL ))

[0304] Through the above-mentioned embodiments, at least the following technical effects can be achieved: First, by comprehensively considering a variety of basic performance and application scenarios, this embodiment provides a comprehensive computing power and computing efficiency evaluation framework, so that the evaluation results are more practical and guiding, especially in application fields such as big data and artificial intelligence. Secondly, the introduction of carbon emission factors makes the definition of computing power and computing efficiency more in line with the current development direction of energy saving and low carbon of servers. In addition, promoting the evaluation of server computing power and computing efficiency is conducive to the full-stack optimization of server hardware and software, creating energy-saving and low-carbon servers, and guiding users to choose more environmentally friendly servers. Moreover, through standardized testing procedures and strict test environment control, the present invention ensures the scientificity and objectivity of the evaluation results and improves the user's trust in the evaluation results. In addition, this method is not only applicable to traditional servers, but can also be flexibly applied to the development direction of current computing platforms, such as big data and virtualized environments, making it more applicable.

[0305] The server computing power and efficiency evaluation method proposed in the above embodiment can be extended to any device that relies on the conversion of electrical energy into mathematical calculations, such as high-performance computing (HPC) clusters, edge computing devices, deep learning accelerators, etc. This means that whether it is a large data center or a small computing device, this method can be used to evaluate its computing power and energy efficiency, thereby optimizing resource allocation and improving environmental sustainability. The server computing power and efficiency evaluation method proposed in this embodiment is not limited to traditional server environments, but can also be applied to other work scenarios, such as cloud computing, virtualized environments, data analysis, financial model calculations, etc. This applicability enables enterprises and institutions in different fields to evaluate and optimize the efficiency of their computing resources according to their own needs. The server computing power and efficiency evaluation method proposed in this embodiment can also be applied to different load intensities. This flexibility makes the evaluation process more adaptable and can reflect actual usage.

[0306] In the present embodiment, a performance test device for a computing device is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made are omitted. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0307] Figure 6 is a structural block diagram of a performance testing device for a computing device according to an embodiment of the present application, such as Figure 6 As shown, the device comprises:

[0308] An acquisition module, used for acquiring test data of a computing device to be tested;

[0309] The detection module is used to detect the performance parameters of the computing device based on the test data and the carbon emission parameters, wherein the carbon emission parameters are used to indicate the carbon emissions of the resources consumed by the computing device, and the performance parameters are used to indicate the operating performance of the computing device under unit carbon emissions.

[0310] Through the above content, test data is obtained from the computing device to be tested, and then the performance parameters of the computing device are detected by combining the test data and the carbon emission parameters. The carbon emission parameters can indicate the carbon emission of the resources consumed by the computing device. The obtained performance parameters can measure the operating performance of the computing device under the unit carbon emission, so that the operating performance of the computing device can be evaluated while also measuring the performance of the computing device in terms of carbon emissions, and then the impact of the computing device on the environment can be compared. Therefore, the technical problem of poor applicability of testing the performance of the computing device can be solved, and the technical effect of improving the applicability of testing the performance of the computing device can be achieved.

[0311] Optionally, the detection module includes:

[0312] A first detection unit is used to detect a first performance parameter of the computing device according to first performance data of the computing device, first power consumption data of the computing device, first test duration of the computing device, and a carbon emission factor of electricity, wherein the test task for the computing device includes: executing a target workload under a target pressure value, the target pressure value is used to indicate the test pressure on the computing device, the first performance data is used to indicate the running performance of the computing device when executing the target workload under the target pressure value, the first power consumption data is used to indicate the running power consumption of the computing device when executing the target workload under the target pressure value, the carbon emission factor of electricity is used to indicate the carbon emission level of unit electricity consumed by the computing device, the first test duration is the duration of the computing device executing the target workload under the target pressure value, the first performance parameter is used to indicate the computing power provided by the computing device when executing the target workload under the target pressure value under unit carbon emission, the test data includes: the first performance data, the first power consumption data, and the first test duration, the carbon emission parameter includes: the carbon emission factor of electricity, and the performance parameter includes: the first performance parameter.

[0313] Optionally, the first detection unit is used to calculate the pressure load computing power efficiency Eff of the computing device by the following formula: load :

[0314]

[0315] Among them, the first performance parameter includes the pressure load computing power efficiency Eff load , Per is the first performance data, Pw is the first power consumption data, CEF is the electricity carbon emission factor, ΔT is the first test duration, ΔT=Ts-Tb, Ts is the test end time of executing the target workload under the target pressure value, and Tb is the test start time of executing the target workload under the target pressure value.

[0316] Optionally, the detection module includes:

[0317] a second detection unit, configured to detect a plurality of second performance parameters of the computing device according to a plurality of second performance data of the computing device, a plurality of second power consumption data of the computing device, a plurality of second test durations of the computing device, and a power carbon emission factor, wherein the test task for the computing device comprises: executing a reference workload under a plurality of pressure values, the plurality of pressure values ​​being used to indicate a plurality of test pressures for the computing device, the second performance data being used to indicate the running performance of the computing device when executing the reference workload under a single pressure value, the second power consumption data being used to indicate the running power consumption of the computing device when executing the reference workload under a single pressure value, the power carbon emission factor being used to indicate the carbon emission level per unit of electricity consumed by the computing device, the second test duration being the duration for the computing device to execute the reference workload under a single pressure value, the second performance parameter being used to indicate the computing power provided by the computing device when executing the reference workload under a single pressure value under a unit of carbon emission, the test data comprising: a plurality of the second performance data, a plurality of the second power consumption data, and a plurality of the second test durations, and the carbon emission parameter comprising: the power carbon emission factor;

[0318] A first conversion unit is used to convert multiple of the second performance parameters into a third performance parameter, wherein the third performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing the reference workload under unit carbon emissions, and the performance parameters include: the third performance parameter.

[0319] Optionally, the first conversion unit is used to calculate the workload computing power and computing efficiency Eff of the computing device by the following formula: worklet :

[0320]

[0321] The third performance parameter includes the workload computing power efficiency Eff worklet , Eff loadi is the second performance parameter under the i-th pressure value, and n is the number of the multiple pressure values.

[0322] Optionally, the detection module includes:

[0323] A third detection unit is used to detect multiple fourth performance parameters of the computing device according to multiple third performance data of the computing device, multiple third power consumption data of the computing device, multiple third test durations of the computing device, and a carbon emission factor of electricity, wherein the test task for the computing device includes: executing multiple workloads in a target test sub-scenario under a corresponding stress value, the corresponding stress value is used to indicate the test stress of the computing device in the corresponding workload, the target test sub-scenario is used to indicate the operation category allowed to run by the computing device, the multiple workloads are used to indicate multiple running operations in the target test sub-scenario, and the third performance data is used to indicate the computing device The operating performance of executing a single workload under a single pressure value, the third power consumption data is used to indicate the operating power consumption of the computing device executing a single workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of power consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device under a single pressure value for executing a single workload per unit of carbon emission, the test data includes: a plurality of the third performance data, a plurality of the third power consumption data and a plurality of the third test durations, and the carbon emission parameters include: the power carbon emission factor;

[0324] A second conversion unit is used to convert the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of the fifth performance parameters, wherein the fifth performance parameter is used to indicate an average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount;

[0325] A third conversion unit is used to convert the multiple fifth performance parameters into a sixth performance parameter, wherein the sixth performance parameter is used to indicate the average computing power and efficiency provided by the computing device at a unit carbon emission level in the target test sub-scenario, and the performance parameters include: the sixth performance parameter.

[0326] Optionally, the third conversion unit is used to calculate the scene computing power and computing efficiency Eff of the computing device by the following formula: workload :

[0327]

[0328] Among them, the sixth performance parameter includes the scene computing power and efficiency Eff workload , Eff workletj is the fifth performance parameter under the j-th workload, and m is the number of the multiple workloads.

[0329] Optionally, the acquisition module includes:

[0330] A collection unit, configured to collect a plurality of initial performance data of the computing device, wherein the initial performance data is used to indicate the actual operating performance of the computing device executing a single workload under a single pressure value;

[0331] A processing unit is used to normalize the multiple initial performance data to obtain the multiple third performance data, wherein the third performance data is used to indicate the normalized operating performance of the computing device executing a single workload under a single pressure value.

[0332] Optionally, the processing unit is used to:

[0333] Extracting a target normalization factor corresponding to a single workload from workloads and normalization factors having a corresponding relationship, wherein the normalization factor is used to normalize performance data corresponding to each workload to the same data level;

[0334] The ratio of each of the initial performance data under a single workload to the target normalization factor is calculated to obtain each of the third performance data under a single workload.

[0335] Optionally, the detection module includes:

[0336] a fourth detection unit, configured to detect a plurality of fourth performance parameters of the computing device according to a plurality of third performance data of the computing device, a plurality of third power consumption data of the computing device, a plurality of third test durations of the computing device, and a carbon emission factor of electricity, wherein the test task for the computing device comprises: executing corresponding workloads in a plurality of test sub-scenarios under corresponding pressure values, the corresponding pressure values ​​being used to indicate the test pressure on the computing device in the corresponding workload, the plurality of test sub-scenarios being used to indicate a plurality of operation categories allowed to be run by the computing device, and the third performance data being used to indicate the computing device executing a single workload under a single pressure value; The third power consumption data is used to indicate the operating power consumption of the computing device when executing a single workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of power consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device when executing a single workload under a single pressure value per unit of carbon emission, the test data includes: a plurality of the third performance data, a plurality of the third power consumption data and a plurality of the third test durations, and the carbon emission parameters include: the power carbon emission factor;

[0337] a fourth conversion unit, configured to convert the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of the fifth performance parameters, wherein the fifth performance parameter is used to indicate an average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount;

[0338] a fifth conversion unit, configured to convert a plurality of the fifth performance parameters under a single test sub-scenario into a seventh performance parameter, to obtain a plurality of the seventh performance parameters, wherein the seventh performance parameter is used to indicate an average computing power and computing efficiency provided by the computing device under a unit of carbon emissions in the single test sub-scenario;

[0339] The sixth conversion unit is used to convert multiple seventh performance parameters into an eighth performance parameter, wherein the eighth performance parameter is used to indicate the average computing power and efficiency provided by the computing device under unit carbon emissions, and the performance parameters include: the eighth performance parameter.

[0340] Optionally, the sixth conversion unit is used to perform one of the following operations:

[0341] Calculating a weighted geometric mean of a plurality of the seventh performance parameters as the eighth performance parameter, wherein the weight corresponding to each of the test sub-scenario is determined according to the degree of influence of the running of the test sub-scenario on the computing power and efficiency of the computing device;

[0342] Calculating a weighted geometric mean of a plurality of the seventh performance parameters as the eighth performance parameter, wherein the plurality of test sub-scenarios are divided into a plurality of test scenarios, and the plurality of test scenarios include: basic computing and application performance, the basic computing is used to test the basic performance of the computing device outputting computing power, and the application performance is used to test the performance of the computing device running a user-side application, a first weight of the basic computing is less than a second weight of the application performance, the first weight is assigned to the test sub-scenario belonging to the basic computing, and the second weight is assigned to the test sub-scenario belonging to the application performance;

[0343] A geometric mean of a plurality of the seventh performance parameters is calculated as the eighth performance parameter.

[0344] Optionally, in the case where the multiple test sub-scenarios include: a basic computing class, the workload corresponding to the basic computing class includes at least one of the following: compression and decompression, encryption and decryption, hash conversion, matrix operations, linear equations, sorting algorithms, concurrent operations, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential read and write, storage random read and write, network bandwidth test, network delay test;

[0345] In the case where the plurality of test sub-scenarios include: big data, the workload corresponding to the big data includes at least one of the following: big data reading, big data writing, and sorting calculation;

[0346] In the case where the plurality of test sub-scenarios include: artificial intelligence AI, the workload corresponding to the artificial intelligence AI includes at least one of the following: graphic recognition, natural language processing;

[0347] In the case where the plurality of test sub-scenarios include: virtualization, the workload corresponding to the virtualization includes at least one of the following: virtualization platform basic performance, virtualization database performance;

[0348] In the case where the plurality of test sub-scenarios include: a database, the workload corresponding to the database includes: database performance.

[0349] Optionally, the sixth conversion unit is used to calculate the computing power efficiency CPE of the computing device by using the following formula:

[0350] CPE=exp(k1×ln(Eff compute )+k2×ln(Eff bd )+k3×ln(Eff AI )+k4×ln(Eff VM )+k5×ln(Eff SQL ));

[0351] The eighth performance parameter includes the computing power and efficiency of the computing device CPE, and the multiple test sub-scenarios include: the basic computing class, the big data, the artificial intelligence AI, the virtualization and the database, Eff compute is the seventh performance parameter under the basic computing class, Eff bd is the seventh performance parameter under the big data, Eff AI is the seventh performance parameter under the artificial intelligence AI, Eff VM is the seventh performance parameter under the virtualization, Eff SQL is the seventh performance parameter under the database, and k1, k2, k3, k4, and k5 are weights corresponding to the respective test sub-scenarios.

[0352] Optionally, the detection module includes:

[0353] a fifth detection unit, configured to calculate the initial performance parameters of the computing device according to the target performance data, the target power consumption data, the target test duration, and the electric power carbon emission factor, wherein the target performance data is used to indicate the operating performance of the computing device under the test task, the target power consumption data is used to indicate the operating power consumption of the computing device under the test task, the electric power carbon emission factor is used to indicate the carbon emission level per unit of electric power consumed by the computing device, the target test duration is the duration for the computing device to perform the test task, the initial performance parameters are used to indicate the computing power provided by the computing device under the test task, the test data includes: the target performance data, the target power consumption data, and the target test duration, and the carbon emission parameters include: the electric power carbon emission factor;

[0354] A searching unit, configured to search a target parameter range into which the initial performance parameter falls from a plurality of parameter ranges, wherein the performance parameter values ​​of the computing device in the test task are divided into the plurality of parameter ranges to obtain parameter ranges and performance levels having corresponding relationships;

[0355] A determination unit is used to determine the target performance level corresponding to the target parameter range as the computing power and efficiency level of the computing device, wherein the performance parameters include: the target performance level.

[0356] Optionally, the acquisition module includes:

[0357] A first receiving unit, configured to receive a first test request, wherein the first test request is used to request that a first test task be performed on the computing device;

[0358] A first storage unit, configured to respond to the first test request and store first test data corresponding to the first test task in the computing device, wherein the first test data is used to be called by the computing device to implement the test of the first test task;

[0359] A first sending unit, configured to initiate a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the first test task;

[0360] A first acquisition unit is used to acquire performance data of the computing device, power consumption data of the computing device and test duration of the computing device as the test data during the process of the computing device responding to the test instruction to execute the first test task, wherein the performance data is used to indicate the running performance of the computing device in the first test task, the power consumption data is used to indicate the running power consumption of the computing device in the first test task, and the test duration is the duration for the computing device to execute the first test task.

[0361] Optionally, the acquisition module includes:

[0362] A second receiving unit, configured to receive a second test request, wherein the second test request is used to request that a second test task be performed on the computing device;

[0363] A second sending unit, configured to respond to the second test request and initiate a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the second test task;

[0364] A second storage unit is used to store second test data corresponding to the candidate workload in the computing device before the computing device responds to the test instruction to execute the candidate workload in the second test task, wherein the second test data is used to be called by the computing device to implement the test of the candidate workload;

[0365] A second acquisition unit is used to acquire performance data of the computing device, power consumption data of the computing device and test duration of the computing device as the test data during the process of the computing device executing the candidate workload, wherein the performance data is used to indicate the running performance of the computing device under the candidate workload, the power consumption data is used to indicate the running power consumption of the computing device under the candidate workload, and the test duration is the duration of the computing device executing the candidate workload.

[0366] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0367] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0368] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0369] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0370] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0371] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0372] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any one of the above method embodiments are implemented.

[0373] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0374] An embodiment of the present application also provides a computer program, which includes computer instructions, which are stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps in any one of the above method embodiments.

[0375] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0376] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0377] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A performance testing method for a computing device, characterized in that: include: Obtaining test data of the computing device to be tested; The performance parameters of the computing device are detected according to the test data and the carbon emission parameters, wherein the carbon emission parameters are used to indicate the carbon emissions of the resources consumed by the computing device, and the performance parameters are used to indicate the operating performance of the computing device under unit carbon emissions.

2. The method according to claim 1, characterized in that The detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes: The first performance parameter of the computing device is detected according to the first performance data of the computing device, the first power consumption data of the computing device, the first test duration of the computing device, and the electric power carbon emission factor, wherein the test task for the computing device includes: executing a target workload under a target pressure value, the target pressure value is used to indicate the test pressure on the computing device, the first performance data is used to indicate the running performance of the computing device in executing the target workload under the target pressure value, the first power consumption data is used to indicate the running power consumption of the computing device in executing the target workload under the target pressure value, the electric power carbon emission factor is used to indicate the carbon emission level per unit of electricity consumed by the computing device, the first test duration is the duration for the computing device to execute the target workload under the target pressure value, the first performance parameter is used to indicate the computing power provided by the computing device in executing the target workload under the target pressure value under unit carbon emission, the test data includes: the first performance data, the first power consumption data and the first test duration, the carbon emission parameter includes: the electric power carbon emission factor, and the performance parameter includes: the first performance parameter.

3. The method according to claim 2, characterized in that The detecting a first performance parameter of the computing device according to the first performance data of the computing device, the first power consumption data of the computing device, the first test duration of the computing device, and the power carbon emission factor includes: The pressure load computing power and computing efficiency Eff of the computing device is calculated by the following formula load : Among them, the first performance parameter includes the pressure load computing power efficiency Eff load , Per is the first performance data, Pw is the first power consumption data, CEF is the electricity carbon emission factor, ΔT is the first test duration, ΔT=Ts-Tb, Ts is the test end time of executing the target workload under the target pressure value, and Tb is the test start time of executing the target workload under the target pressure value.

4. The method according to claim 1, characterized in that: The detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes: Detecting multiple second performance parameters of the computing device according to multiple second performance data of the computing device, multiple second power consumption data of the computing device, multiple second test durations of the computing device, and a power carbon emission factor, wherein the test task for the computing device includes: executing a reference workload under multiple pressure values, the multiple pressure values ​​are used to indicate multiple test pressures for the computing device, the second performance data are used to indicate the running performance of the computing device executing the reference workload under a single pressure value, the second power consumption data are used to indicate the running power consumption of the computing device executing the reference workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the second test duration is the duration of the computing device executing the reference workload under a single pressure value, the second performance parameter is used to indicate the computing power provided by the computing device executing the reference workload under a single pressure value under unit carbon emissions, the test data includes: multiple second performance data, multiple second power consumption data, and multiple second test durations, and the carbon emission parameter includes: the power carbon emission factor; Convert multiple of the second performance parameters into a third performance parameter, wherein the third performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing the reference workload under unit carbon emissions, and the performance parameters include: the third performance parameter.

5. The method according to claim 4, characterized in that The converting the plurality of the second performance parameters into a third performance parameter comprises: The workload computing power and computing efficiency Eff of the computing device is calculated by the following formula workler : The third performance parameter includes the workload computing power efficiency Eff workler , Eff loadi is the second performance parameter under the i-th pressure value, and n is the number of the multiple pressure values.

6. The method according to claim 1, characterized in that The detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes: The plurality of fourth performance parameters of the computing device are detected according to the plurality of third performance data of the computing device, the plurality of third power consumption data of the computing device, the plurality of third test durations of the computing device, and the carbon emission factor of electricity, wherein the test task for the computing device comprises: executing the plurality of workloads in the target test sub-scenario under the corresponding pressure value, the corresponding pressure value is used to indicate the test pressure on the computing device in the corresponding workload, the target test sub-scenario is used to indicate the operation category allowed to run by the computing device, the plurality of workloads are used to indicate the plurality of running operations in the target test sub-scenario, and the third performance data is used to indicate the computing device under a single pressure The operating performance of executing a single workload under a single pressure value, the third power consumption data is used to indicate the operating power consumption of the computing device executing a single workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of power consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device under a single pressure value to execute a single workload per unit of carbon emission, the test data includes: a plurality of the third performance data, a plurality of the third power consumption data and a plurality of the third test durations, and the carbon emission parameters include: the power carbon emission factor; Convert the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of the fifth performance parameters, wherein the fifth performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount; Convert multiple fifth performance parameters into sixth performance parameters, wherein the sixth performance parameter is used to indicate the average computing power and efficiency provided by the computing device at unit carbon emissions in the target test sub-scenario, and the performance parameters include: the sixth performance parameter.

7. The method according to claim 6, characterized in that The converting the plurality of fifth performance parameters into a sixth performance parameter comprises: The scene computing power and computing efficiency Eff of the computing device is calculated by the following formula workload : Among them, the sixth performance parameter includes the scene computing power and efficiency Eff workload , Eff workletj is the fifth performance parameter under the j-th workload, and m is the number of the multiple workloads.

8. The method according to claim 6, characterized in that The step of obtaining test data of the computing device to be tested includes: Collecting a plurality of initial performance data of the computing device, wherein the initial performance data is used to indicate the actual operating performance of the computing device executing a single workload under a single pressure value; The plurality of the initial performance data are normalized to obtain the plurality of the third performance data, wherein the third performance data are used to indicate the normalized operating performance of the computing device executing a single workload under a single pressure value.

9. The method according to claim 8, characterized in that The normalizing the plurality of initial performance data comprises: Extracting a target normalization factor corresponding to a single workload from workloads and normalization factors having a corresponding relationship, wherein the normalization factor is used to normalize performance data corresponding to each workload to the same data level; The ratio of each of the initial performance data under a single workload to the target normalization factor is calculated to obtain each of the third performance data under a single workload.

10. The method according to claim 1, characterized in that The detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes: Detecting multiple fourth performance parameters of the computing device according to multiple third performance data of the computing device, multiple third power consumption data of the computing device, multiple third test durations of the computing device, and a power carbon emission factor, wherein the test task for the computing device includes: executing corresponding workloads in multiple test sub-scenarios under corresponding pressure values, the corresponding pressure values ​​are used to indicate the test pressure on the computing device in the corresponding workload, the multiple test sub-scenarios are used to indicate multiple operation categories allowed to run by the computing device, the third performance data are used to indicate the running performance of the computing device when executing a single workload under a single pressure value, the third power consumption data are used to indicate the running power consumption of the computing device when executing a single workload under a single pressure value, the power carbon emission factor is used to indicate the carbon emission level of the computing device per unit of electricity consumed, the third test duration is the duration of the computing device executing a single workload under a single pressure value, the fourth performance parameter is used to indicate the computing power provided by the computing device when executing a single workload under a single pressure value under a unit of carbon emission, the test data includes: multiple third performance data, multiple third power consumption data, and multiple third test durations, and the carbon emission parameter includes: the power carbon emission factor; Convert the fourth performance parameter under a single workload into a fifth performance parameter to obtain a plurality of the fifth performance parameters, wherein the fifth performance parameter is used to indicate the average computing power and efficiency provided by the computing device when executing a single workload under a unit carbon emission amount; Converting a plurality of the fifth performance parameters in a single test sub-scenario into a seventh performance parameter to obtain a plurality of the seventh performance parameters, wherein the seventh performance parameter is used to indicate an average computing power and computing efficiency provided by the computing device under a unit carbon emission in the single test sub-scenario; Convert multiple seventh performance parameters into an eighth performance parameter, wherein the eighth performance parameter is used to indicate the average computing power and efficiency provided by the computing device under unit carbon emissions, and the performance parameters include: the eighth performance parameter.

11. The method according to claim 10, characterized in that The converting the plurality of the seventh performance parameters into the eighth performance parameter comprises one of the following: Calculating a weighted geometric mean of a plurality of the seventh performance parameters as the eighth performance parameter, wherein the weight corresponding to each of the test sub-scenario is determined according to the degree of influence of the running of the test sub-scenario on the computing power and efficiency of the computing device; Calculating a weighted geometric mean of a plurality of the seventh performance parameters as the eighth performance parameter, wherein the plurality of test sub-scenarios are divided into a plurality of test scenarios, and the plurality of test scenarios include: basic computing and application performance, the basic computing is used to test the basic performance of the computing device outputting computing power, and the application performance is used to test the performance of the computing device running a user-side application, a first weight of the basic computing is less than a second weight of the application performance, the first weight is assigned to the test sub-scenario belonging to the basic computing, and the second weight is assigned to the test sub-scenario belonging to the application performance; A geometric mean of a plurality of the seventh performance parameters is calculated as the eighth performance parameter.

12. The method according to claim 10, characterized in that In the case where the multiple test sub-scenarios include: a basic computing class, the workload corresponding to the basic computing class includes at least one of the following: compression and decompression, encryption and decryption, hash conversion, matrix operation, linear equation, sorting algorithm, concurrent operation, code performance, memory bandwidth test, memory cache test, memory delay test, storage sequential read and write, storage random read and write, network bandwidth test, network delay test; In the case where the plurality of test sub-scenarios include: big data, the workload corresponding to the big data includes at least one of the following: big data reading, big data writing, and sorting calculation; In the case where the plurality of test sub-scenarios include: artificial intelligence A1, the workload corresponding to the artificial intelligence A1 includes at least one of the following: graphic recognition, natural language processing; In the case where the plurality of test sub-scenarios include: virtualization, the workload corresponding to the virtualization includes at least one of the following: virtualization platform basic performance, virtualization database performance; In the case where the plurality of test sub-scenarios include: a database, the workload corresponding to the database includes: database performance.

13. The method according to claim 12, characterized in that The converting the plurality of the seventh performance parameters into the eighth performance parameter comprises: The computing power and computing efficiency (CPE) of the computing device is calculated by the following formula: CPE=exp(k1×ln(Eff compute )+k2×ln(Eff bd )+k3×ln(Eff AI )+k4×ln(Eff VM )+k5×ln(Eff SQL )); The eighth performance parameter includes the computing power and efficiency of the computing device CPE, and the multiple test sub-scenarios include: the basic computing class, the big data, the artificial intelligence AI, the virtualization and the database, Efff compute is the seventh performance parameter under the basic computing class, Eff bd is the seventh performance parameter under the big data, Eff AI is the seventh performance parameter under the artificial intelligence AI, Eff VM is the seventh performance parameter under the virtualization, Eff SQL is the seventh performance parameter under the database, and k1, k2, k3, k4, and k5 are weights corresponding to the respective test sub-scenarios.

14. The method according to claim 1, characterized in that The detecting the performance parameters of the computing device according to the test data and the carbon emission parameters includes: The initial performance parameters of the computing device are calculated according to the target performance data, the target power consumption data, the target test duration and the electric power carbon emission factor, wherein the target performance data is used to indicate the operating performance of the computing device under the test task, the target power consumption data is used to indicate the operating power consumption of the computing device under the test task, the electric power carbon emission factor is used to indicate the carbon emission level of the unit power consumed by the computing device, the target test duration is the duration for the computing device to perform the test task, the initial performance parameters are used to indicate the computing power provided by the computing device under the test task, the test data includes: the target performance data, the target power consumption data and the target test duration, and the carbon emission parameters include: the electric power carbon emission factor; Searching for a target parameter range into which the initial performance parameter falls from a plurality of parameter ranges, wherein the performance parameter values ​​of the computing device in the test task are divided into the plurality of parameter ranges to obtain parameter ranges and performance levels having corresponding relationships; The target performance level corresponding to the target parameter range is determined as the computing power and efficiency level of the computing device, wherein the performance parameters include: the target performance level.

15. The method according to claim 1, characterized in that The step of obtaining test data of the computing device to be tested includes: Receive a first test request, wherein the first test request is used to request to perform a test of a first test task on the computing device; In response to the first test request, storing first test data corresponding to the first test task to the computing device, wherein the first test data is used to be called by the computing device to implement the test of the first test task; Initiating a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the first test task; In the process of the computing device responding to the test instruction to execute the first test task, the performance data of the computing device, the power consumption data of the computing device and the test duration of the computing device are obtained as the test data, wherein the performance data is used to indicate the operating performance of the computing device in the first test task, the power consumption data is used to indicate the operating power consumption of the computing device in the first test task, and the test duration is the duration for the computing device to execute the first test task.

16. The method according to claim 1, characterized in that The step of obtaining test data of the computing device to be tested includes: receiving a second test request, wherein the second test request is used to request to perform a second test task on the computing device; In response to the second test request, initiating a test instruction to the computing device, wherein the test instruction is used to instruct the computing device to execute the second test task; Before the computing device executes the candidate workload in the second test task in response to the test instruction, storing second test data corresponding to the candidate workload in the computing device, wherein the second test data is used to be called by the computing device to implement the test of the candidate workload; During the process of the computing device executing the candidate workload, performance data of the computing device, power consumption data of the computing device and test duration of the computing device are obtained as the test data, wherein the performance data is used to indicate the operating performance of the computing device under the candidate workload, the power consumption data is used to indicate the operating power consumption of the computing device under the candidate workload, and the test duration is the duration of the computing device executing the candidate workload.

17. A performance testing system for a computing device, It is characterized in that It includes: a controller, a collector and a power supply, wherein the controller is connected to the collector, and the controller, the collector and the power supply are all used to connect to a computing device to be tested; The power supply is used to supply power to the computing device; The collector is used to collect test data of the computing device; The controller is used to execute the steps of the method described in any one of claims 1 to 16.

18. The system according to claim 17, characterized in that The collector comprises: a target processor and a power meter, wherein the power meter is connected between the power supply and the computing device, the controller and the target processor are deployed on a control machine, and the power meter is connected to the control machine; The target processor is used to collect performance data and test duration in the test data, wherein the performance data is used to indicate the operating performance of the computing device in the target test task, and the test duration is the duration of the computing device executing the target test task; The power meter is used to collect power consumption data in the test data, wherein the power consumption data is used to indicate the operating power consumption of the computing device in the target test task.

19. The system according to claim 18, characterized in that In the case where the computing device includes a liquid-cooled computing device, the computing device includes: a test machine and a liquid cooling device, the power output terminal of the power supply is connected to the power input terminal of the power meter, and the power output terminal of the power meter is connected to the power input terminal of the test machine and the power input terminal of the liquid cooling device; The power consumption data in the test data collected by the power meter includes power consumption parameters of the test machine and power consumption parameters of the liquid cooling device.

20. The system according to claim 17, characterized in that The system further comprises: a thermometer connected to the controller, wherein the computing device is disposed within a temperature detection range of the thermometer; The thermometer is used to detect the ambient temperature of the computing device; The controller is used to test the computing device when the ambient temperature falls within a target temperature range.

21. A performance testing device for a computing device, characterized in that: include: An acquisition module, used for acquiring test data of a computing device to be tested; A detection module is used to detect the performance parameters of the computing device based on the test data and the carbon emission parameters, wherein the carbon emission parameters are used to indicate the carbon emissions of the resources consumed by the computing device, and the performance parameters are used to indicate the operating performance of the computing device under unit carbon emissions.

22. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the method described in any one of claims 1 to 16 when executed by a processor.

23. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method described in any one of claims 1 to 16 are implemented.

24. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 16 are implemented.

Images