Multi-dimensional power consumption test method, device and storage medium for solid state disk

Abstract

The application discloses a kind of multi-dimension power consumption test method, equipment and storage medium of solid state disk, the method includes: in response to test start instruction, control test environment parameter and solid state disk enters preset initial state;Based on the test matrix of multiple temperature and capacity occupation combinations being predefined, control test equipment to apply multi-dimension dynamic work load to solid state disk, and synchronously collect power consumption data, internal state data and performance data;The multi-source data collected are time-synchronized and data fusion, form associated data set;Intelligently analyze the data set, extract the power consumption characteristics of solid state disk, and generate power consumption evaluation report based on it.The application solves the problems of single test dimension, incomplete test condition coverage, ineffective fusion of multi-source data and lack of intelligent analysis in the prior art, and realizes comprehensive, accurate and automated evaluation of solid state disk power consumption.

Description

Technical Field

[0001] This invention relates to the field of solid-state drive (SSD) testing technology, and in particular to a multi-dimensional power consumption testing method, device, and storage medium for SSDs. Background Technology

[0002] With the rapid development of big data, cloud computing, and mobile computing technologies, solid-state drives (SSDs) have become the core carrier of data storage due to their advantages such as high performance, low latency, and high shock resistance. At the same time, the miniaturization and portability of devices, along with the high demands for battery life, have made SSD power consumption performance a key indicator for evaluating their overall technical level, directly affecting the total cost of ownership of data centers and the battery life of mobile devices. Currently, the industry commonly uses a combination of static and benchmark testing methods based on point measurements to test SSD power consumption. A typical procedure includes: connecting the SSD under test to a power consumption testing fixture, which is then connected to the motherboard; idling the SSD for a period of time, and then recording its static voltage and current values; subsequently, using benchmark software, performing read and write operations respectively, and recording the current and voltage data during the operation; finally, by performing simple calculations on these data, obtaining the average power consumption of the SSD or the power consumption value at a certain moment.

[0003] However, existing methods have significant limitations and cannot meet the needs of comprehensively evaluating the power consumption performance of solid-state drives (SSDs). First, existing methods rely on a single testing dimension, primarily using synthetic benchmark software for point sampling or averaging, which fails to fully reflect the dynamic power consumption characteristics of SSDs under complex loads. Second, existing methods fail to systematically consider the impact of temperature and storage capacity differences, resulting in incomplete test condition coverage and a high degree of randomness in the evaluation results, making it impossible to know the power consumption performance of SSDs under various potential operating conditions. Finally, existing methods lack intelligent analysis capabilities for test data, relying mainly on manual data interpretation and report generation, which is inefficient and prone to introducing subjective errors. Summary of the Invention

[0004] This invention provides a multi-dimensional power consumption testing method, device, and storage medium for solid-state drives, to solve the problems of existing testing methods having single dimensions, incomplete coverage of testing conditions, and lack of intelligent analysis capabilities.

[0005] Firstly, a multi-dimensional power consumption testing method for solid-state drives is provided, including:

[0006] In response to the test start command, control the test environment parameters and the solid-state drive to enter the preset initial state;

[0007] Based on a predefined test matrix, the test equipment is controlled to apply a multi-dimensional dynamic workload to the solid-state drive, and power consumption data, internal status data and performance data are collected simultaneously. The test matrix defines multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions.

[0008] Power consumption data, internal status data, and performance data are synchronized and fused in time to form a related dataset;

[0009] Intelligent analysis is performed on the associated dataset to extract the power consumption characteristics of solid-state drives, and a power consumption assessment report is generated based on the power consumption characteristics.

[0010] Secondly, a multi-dimensional power consumption testing device for solid-state drives is provided, including:

[0011] The initialization module is used to respond to the test start command and control the test environment parameters and the solid-state drive to enter the preset initial state.

[0012] The testing module is used to control the testing equipment to apply multi-dimensional dynamic workloads to the solid-state drive based on a predefined test matrix, and to simultaneously collect power consumption data, internal status data and performance data. The test matrix defines multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions.

[0013] The data association module is used to synchronize and fuse power consumption data, internal status data, and performance data in time to form an associated dataset.

[0014] The analysis module is used to perform intelligent analysis on the associated dataset to extract the power consumption characteristics of the solid-state drive and generate a power consumption assessment report based on the power consumption characteristics.

[0015] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the multi-dimensional power consumption test method for the solid-state drive described above.

[0016] Fourthly, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, it implements the steps of the above-mentioned multi-dimensional power consumption test method for solid-state drives.

[0017] The aforementioned multi-dimensional power consumption testing method, device, equipment, and storage medium for solid-state drives (SSDs) systematically apply dynamic workloads and collect data under different combinations of ambient temperature and capacity usage conditions using a predefined test matrix. This overcomes the problem of single test conditions in traditional methods and comprehensively reveals the power consumption behavior of SSDs under various potential operating conditions. Furthermore, by synchronizing and fusing the synchronously collected power consumption data, internal status data, and performance data, a correlated dataset containing multi-dimensional information is formed, laying a data foundation for in-depth power consumption analysis from phenomenon to root cause, and solving the drawbacks of isolated data and superficial analysis in traditional methods. Moreover, through intelligent analysis of the correlated dataset, power consumption characteristics are automatically extracted and evaluation reports are generated, reducing reliance on manual analysis. This not only improves testing efficiency but also ensures the objectivity and consistency of analysis results, realizing an automated closed loop from data collection to knowledge insight. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a multi-dimensional power consumption testing method for solid-state drives in one embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of a multi-dimensional power consumption testing device for solid-state drives in one embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 As shown, Figure 1 A flowchart illustrating a multi-dimensional power consumption testing method for solid-state drives provided in this embodiment of the invention includes the following steps:

[0024] Step S1: In response to the test start command, control the test environment parameters and the solid-state drive to enter the preset initial state.

[0025] Specifically, in order to establish a highly controllable and fully repeatable test starting point and ensure the consistency and comparability of all subsequent test results, this embodiment performs an initial operation on the solid-state drive using pre-set test environment parameters before conducting the test, so that the solid-state drive enters a preset initial state, and subsequent test operations are performed based on this initial state.

[0026] Furthermore, in step S1, controlling the test environment parameters and the solid-state drive to enter a preset initial state includes:

[0027] 1. Control the test system resources to meet the preset test capability benchmarks, which include the sampling rate and accuracy of the data acquisition system, the temperature control range and accuracy of the temperature control device, and the interaction capability between the test host and the solid-state drive firmware.

[0028] Specifically, before the test begins, the system will perform a comprehensive self-check process to ensure that all test equipment meets the preset stringent capability requirements:

[0029] First, a high-precision data acquisition system is configured: the system uses a digital power analyzer (such as a Keysight N6705C with an N6781A SMU module) with a sampling rate of no less than 1MS / s and an accuracy of no less than 0.1% to ensure accurate capture of microsecond-level power consumption changes of the solid-state drive during transient processes. A customized test fixture board precisely introduces the power supply lines of the solid-state drive into the power analyzer and integrates a dedicated GPIO trigger interface. This interface is connected to specific signal pins of the solid-state drive (such as DEVSLP) or the command output of the test host through a precision timing circuit, realizing nanosecond-level hardware synchronization between power consumption acquisition and the internal operation of the solid-state drive, completely eliminating time errors caused by software latency.

[0030] Secondly, it is equipped with precise temperature control and environmental simulation devices: the system uses professional high and low temperature chambers (such as the ESPEC T-242 series), whose temperature control range covers -40°C to +125°C and whose control accuracy is better than ±0.5°C, ensuring that it can accurately simulate various real-world application environments from extreme cold to high temperatures. In addition to controlling the ambient temperature of the solid-state drive, it also monitors the temperature of key hot spots on the surface of the solid-state drive controller chip and NAND flash memory chip in real time through a fine-diameter T-type thermocouple or infrared thermal imager interface, with a sampling rate of no less than 10Hz, providing accurate thermal data for subsequent power consumption-temperature correlation analysis;

[0031] Finally, a high-performance test host and internal status monitoring platform were configured: the test host is equipped with a high-performance processor (such as an Intel Core i9), ample PCIe lanes and memory to ensure it can withstand high-intensity load generation. More importantly, the host is equipped with a self-developed "Intelligent Solid State Drive Power Consumption Test Platform" software. This software interacts deeply with the solid state drive firmware through the NVMe-MI protocol, NVMeAdmin command set, or specific manufacturer-defined commands. It can read the status of the solid state drive's internal key registers in real time at a frequency of no less than 100Hz, including but not limited to: controller busy / idle percentage, NAND read / write / erase operation counters, entry and exit flags for various power-saving states (such as PS0-PS4, Slumber), temperature sensor readings, and status flags for background activities such as garbage collection and wear leveling.

[0032] 2. Control the solid-state drive to perform data preprocessing operations, and achieve multiple test combinations defined in the test matrix one by one. When each test combination is achieved, control the solid-state drive to enter an idle state. In the idle state, control the solid-state drive to cycle through all power saving states and record the steady-state power consumption and state switching delay of each power saving state to establish a basic power consumption profile corresponding to each test combination.

[0033] Specifically, in order to establish an accurate, multi-dimensional power consumption benchmark for each specific combination of test conditions, this embodiment performs a basic power consumption profile for each temperature and capacity usage test combination during the solid-state drive initialization process. This provides a baseline for the analysis of subsequent test results. The power consumption profile process can be divided into three stages:

[0034] In the first stage, the system first controls the SSD to perform a complete secure erase operation (such as the NVMe Format NVM command) to ensure it is in a completely known blank state. Then, based on the target capacity utilization rate set in the test matrix, intelligent adaptive preprocessing writes are performed. For example, for a 70% utilization target, the system does not simply perform continuous filling, but dynamically generates a "cellular" data filling pattern based on the SSD's actual physical architecture (obtained by reading the Identify command). For example, data in a mixed mode (including compressible 0x00 mode and incompressible random mode) is written at intervals in 32MB units at different logical block addresses. This distribution more realistically simulates the data fragmentation situation in actual user use, while also allowing observation of the impact of the controller's data compression operation on the power consumption of the preprocessing stage.

[0035] In the second phase, under each test combination condition (e.g., temperature set at 55°C, capacity utilization set at 95%), the control system first waits for the chamber temperature to stabilize (maintaining at least 5 minutes within the target value ±0.3°C range), and then prompts the solid-state drive (SSD) to enter a stable idle state. This stable idle state is determined using a dual standard: first, the input current fluctuation monitored by the power analyzer remains within ±2% for 30 seconds; second, the internal status monitoring platform confirms that the current power status reported by the SSD is stable at the target state (e.g., PS3) and there are no background activity indicators triggered.

[0036] In the third stage, after confirming a stable idle state, the test software actively, quickly, and orderly cycles through all power-saving states supported by the SSD using standard protocol commands (such as the NVMeSet Features command to set the Power State). This process involves switching sequentially from the highest performance PS0 to PS1, PS2, PS3, and then back to PS0. During this process, the power analyzer synchronously records, under precise GPIO triggering, the following data: the steady-state average power consumption for each power-saving state (average over 100ms after stabilization), the state transition latency (the time from sending the switch command to the power consumption stabilizing in the new state) and the exit latency (the time from sending the exit command to the power consumption returning to the PS0 level), as well as the complete voltage and current transient waveforms during the state transition (sampling rate maintained at 1ms / s). All this data, after being timestamped, constitutes a complete basic power consumption profile for this specific test combination, serving as a normalized comparison benchmark for all subsequent dynamic test results.

[0037] It is important to emphasize that the creation of the basic power consumption profile is not a one-time operation, but rather a cyclical sub-step in the testing process. The system automatically traverses each temperature and capacity utilization combination in the test matrix (for example, first at 25°C & 30% utilization, then at 25°C & 70% utilization, and so on up to 85°C & 95% utilization), creating a unique basic power consumption profile for each combination. The purpose is to ensure that subsequent dynamic power consumption tests performed under that combination have an accurate power consumption benchmark obtained under exactly the same initial conditions for comparison.

[0038] Step S2: Based on the predefined test matrix, control the test equipment to apply multi-dimensional dynamic workload to the solid-state drive, and simultaneously collect power consumption data, internal status data and performance data. The test matrix defines multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions.

[0039] Specifically, in order to fully stimulate and capture the real power consumption characteristics of solid-state drives under various complex operating conditions, this embodiment uses a test matrix to define multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions. Under each test combination, an intelligent dynamic load is applied to the solid-state drive, thereby conducting a comprehensive test of the solid-state drive.

[0040] Furthermore, step S2, which involves controlling the testing equipment to apply a multi-dimensional dynamic workload to the solid-state drive and simultaneously collecting power consumption data, internal status data, and performance data, specifically includes:

[0041] 1. For each set of ambient temperature and capacity usage conditions, the workload parameters of the solid-state drive are controlled to gradually increase from the initial value, and the power consumption data and internal status data are monitored in real time during the increase process.

[0042] Specifically, according to the definition of the test matrix, the system performs a progressive load ramp for each combination of temperature and capacity utilization (e.g., 85°C & 70% utilization). This is achieved by controlling the SSD's workload parameters, starting with lower initial values ​​(e.g., queue depth QD=1, block size 4KB, 100% random read), and gradually increasing them according to a preset stepping strategy. For example, initially fixing the block size at 4KB, the QD is increased from 1 to 2, 4, 8, 16, and 32; then, at key QD points (e.g., 8, 16, 32), the block size is gradually increased to 128KB and 1MB. Throughout this ramp-up process, the system monitors the trends of power consumption data (instantaneous power, current) and key status data read from the SSD's internal registers (e.g., controller load rate, NAND command queue depth) in real time at a frequency of at least 10Hz.

[0043] 2. Identify the power-sensitive operating point of the solid-state drive based on the changing trend.

[0044] Specifically, during the gradual load ramp-up process, the system's built-in real-time analysis engine automatically processes the monitoring data stream. When it detects an inflection point in the power consumption growth curve at a certain load parameter change step, the algorithm determines that the current workload parameter combination (e.g., QD=24, 128KB sequential read) constitutes a "power-sensitive operating point" (i.e., an "efficiency inflection point"). The system records the precise parameters at this point and automatically performs more intensive sampling tests in the vicinity of these parameters (e.g., QD=16, 20, 24, 28, 32) to accurately depict the shape and boundary of the efficiency inflection point.

[0045] Furthermore, the step of identifying the power-sensitive operating point of the solid-state drive based on the changing trend specifically includes: when the growth rate of power consumption data exceeds the first threshold and the growth rate of performance data is lower than the second threshold during the process of increasing workload parameters, it is determined that the current combination of workload parameters constitutes a power-sensitive operating point.

[0046] Specifically, the inflection point of the power consumption data can be determined by the instantaneous growth rate of the power consumption data and the growth rate of the performance data. When the instantaneous growth rate of the power consumption data exceeds a set first threshold (e.g., power growth exceeds 50%) while the growth rate of the performance data (such as bandwidth or IOPS) is lower than a set second threshold (e.g., performance growth is lower than 5%), the current workload parameters can be combined as a power-sensitive operating point.

[0047] 3. Focusing on power-sensitive operating points, control the test equipment to perform at least one in-depth test, including: transient state switching test, scenario simulation test based on real input / output trace files, or targeted burst traffic stress test.

[0048] Specifically, after detecting power-sensitive operating points, the system selectively performs one or more in-depth tests based on one or more intelligently identified power-sensitive operating points to explore the power consumption behavior of the solid-state drive from different dimensions.

[0049] Furthermore, transient state transition tests include:

[0050] 3.01. Send power state switching command to solid-state drive.

[0051] 3.02. While sending the command, trigger a high sampling rate power consumption data acquisition operation to capture the voltage and current transient waveforms of the solid-state drive at the moment of power state switching.

[0052] Specifically, under specific test conditions (e.g., 0°C and 30% utilization), the test software sends a precise sequence of power state switching commands to the SSD via NVMe commands (e.g., Set Features). This sequence is used to force a switch from Active to PS3 state, maintain the switch for 100ms, and then switch back to Active. At the same nanosecond-level moment the switching commands are sent, a hardware trigger signal is sent to the power analyzer via the GPIO interface on the fixture board. This initiates a high-sampling-rate (≥1MS / s) acquisition window (typically lasting several milliseconds), accurately capturing key phenomena occurring during the state transition. These phenomena include the instantaneous voltage sag amplitude and duration, the current inrush current peak and waveform, and any ringing frequency and attenuation characteristics. This high-precision transient waveform data is crucial for evaluating the SSD's own power integrity design (e.g., the adequacy of the capacitor network) and its impact on the system power supply.

[0053] Furthermore, scenario simulation tests based on real input / output trace files include:

[0054] 3.11 Control the solid-state drive to play back pre-recorded input / output trace files.

[0055] 3.12 During playback, power consumption data is collected, and the collected power consumption data stream is timestamped with the operation type in the input / output trace file to establish a correspondence between specific application operations and power consumption peaks.

[0056] Specifically, the control system guides the solid-state drive (SSD) to replay high-fidelity input / output trace files captured in real-world application scenarios. These trace files can originate from actual recordings of the entire Windows system startup process, stress tests of Chrome browsers simultaneously loading 50 tabs, or I / O records from MySQL databases executing TPC-C benchmark tests. During playback, the power analyzer continuously collects power consumption data, while the internal status monitoring platform synchronously records SSD register data. The test platform's post-processing module uses a precise timestamp alignment algorithm (with millisecond-level accuracy) to accurately correlate the collected power consumption data stream with each micro-operation type (such as "4KB random read" or "128KB sequential write") in the trace file. Through this correlation, specific peaks or abnormal patterns on the power consumption curve can be clearly located to specific application operations within the trace. For example, the analysis report may clearly state: "At 12.3 seconds after system startup, due to the simultaneous loading of driver files for multiple antivirus services (manifested as a series of 4-8KB random writes), the power consumption of the solid-state drive experienced a plateau period of 800ms with a peak of 4.5W, which is 250% higher than the idle power consumption."

[0057] Further, targeted burst traffic stress tests include:

[0058] 3.21. Based on the workload parameters corresponding to the power-sensitive operating point, generate a high-intensity burst traffic load that lasts for a first preset duration.

[0059] 3.22. During periods of sudden surge in power consumption, peak power consumption data and chip temperature data of the solid-state drive are collected simultaneously to analyze the correlation between heat dissipation performance and peak power consumption.

[0060] It should be noted that this first preset duration is pre-set and must meet the condition of being extremely short, for example, setting the first preset duration to 100ms. Specifically, based on the workload parameters corresponding to the power-sensitive operating point found by adaptive stress testing (for example, identifying 128KB sequential writes at QD=32 as a sensitive point), the test engine specifically generates a burst workload with an extremely short duration but extremely high intensity. For example, within a 100ms time window, 128KB sequential read requests are sent to the solid-state drive at an extreme rate, with the queue depth instantly reaching 64 or even 128. During this burst workload, the power analyzer captures the peak power consumption data of the solid-state drive at the highest sampling rate, while simultaneously acquiring its temperature rise curve at a high frequency (e.g., 1kHz) using thermocouples attached to the surface of the main controller chip. By comprehensively analyzing the absolute value of peak power consumption, duration, and instantaneous temperature rise rate (°C / second) of the chip, we can obtain the instantaneous power consumption tolerance of the solid-state drive, the heat dissipation characteristics of the package, and the stability of the controller and flash memory working together, providing key data for evaluating the performance of solid-state drives in extreme application scenarios (such as instantaneous writing of database transaction logs).

[0061] It should be understood that in this embodiment, the progressive load ramping test process under various test combinations is a global, parameterized scan. Through systematic "climbing," it plots the power consumption-performance curve of the solid-state drive across the entire load range, achieving unbiased, discovery-based testing. In contrast, deep testing precisely focuses on each "suspicious point" (power-sensitive operating point) discovered by adaptive testing. For example, for the discovered point (QD=32, 128KB Seq Read), a dedicated, more in-depth test is performed. This is a targeted, diagnostic test. In terms of their driving relationship, the output of the progressive load ramping test (the specific parameters of the power-sensitive operating point, such as queue depth, block size, and operation type) is the input of the deep test. Without the power-sensitive operating point provided by the progressive load ramping test, the deep test can only blindly select test parameters, making it difficult to guarantee that it will reach its correct power bottleneck. In other words, progressive load ramping testing is the first stage of exploring and discovering problems. Through intelligent load ramping, it automatically and efficiently locates the key points in the power consumption behavior of solid-state drives. In-depth testing is the next stage. It uses the findings of the first stage to mobilize more precise and complex testing methods to conduct in-depth pathological analysis of these key points, and ultimately achieves root cause diagnosis and comprehensive evaluation of the power consumption characteristics of solid-state drives.

[0062] Step S3: Synchronize and fuse power consumption data, internal status data, and performance data in time to form a related dataset.

[0063] Specifically, at the start of the test, the system initializes a unified hardware-software co-working timeline with microsecond-level (or even higher) precision. This timeline is based on the sampling clock of the power analyzer, and the timestamps of all other devices are synchronized to this reference after precise latency calibration. The data fusion engine realigns and interpolates all raw power consumption data (including voltage and current waveforms and calculated real-time power) from the power analyzer, status data streams from the SSD's internal registers (controller load rate, NAND operation counters, power-saving status flags, etc.), and performance data (IOPS, bandwidth, latency timing logs) from testing tools (such as FIO) based on this unified timeline. This results in a set of row-level, multi-dimensional structured data records arranged according to a unified time series. Each record in the dataset contains complete test dimension information at the same microsecond, providing a data foundation for subsequent rigorous causal and correlation analysis.

[0064] It should be noted that the time synchronization and data fusion steps also include synchronizing and fusing the data representing the solid-state drive's temperature status with the power consumption response data, internal status data, and performance data. In this embodiment, the solid-state drive's temperature status data includes ambient temperature and chip hotspot temperature.

[0065] The ambient temperature is provided by the control system of the high and low temperature chamber (such as the ESPEC T-242). This control system has a built-in high-precision platinum resistance temperature sensor (PT100), and the temperature data it collects is transmitted in real time to the main control software of the test via the chamber's communication interface (such as GPIB, Ethernet, or RS-232), with a sampling rate of no less than 1Hz and an accuracy better than ±0.3°C. This data represents the macroscopic ambient temperature of the solid-state drive.

[0066] Chip hotspot temperatures are directly monitored by precisely attaching thin-diameter T-type or K-type thermocouples (wire diameter less than 0.2mm) to the surface of the solid-state drive controller and specific NAND flash memory chips, or by non-contact measurement using an infrared thermal imager. The thermocouple voltage signals are acquired and converted from analog to digital by a high-precision data acquisition card (such as the NI PXIe-4353), with a sampling rate configurable from 10Hz to 1kHz to capture transient temperature changes. These thermocouple readings directly reflect the heat generated by the chip's own power consumption and are crucial inputs for accurate power consumption and temperature correlation analysis.

[0067] Step S4: Perform intelligent analysis on the associated dataset to extract the power consumption characteristics of the solid-state drive and generate a power consumption assessment report based on the power consumption characteristics.

[0068] Specifically, after obtaining the associated dataset, intelligent analysis is performed on the associated dataset. This can be achieved using machine learning algorithm models, preset rule strategies, and other methods to extract the power consumption characteristics of the solid-state drive and generate a power consumption assessment report based on these characteristics.

[0069] Furthermore, step S4, the step of intelligent analysis of the associated dataset, specifically includes:

[0070] 1. Clean up outlier data points in associated datasets in an automated manner.

[0071] Specifically, this embodiment employs advanced machine learning anomaly detection algorithms (such as IsolationForest or Local Outlier Factor (LOF)) to automatically clean the fused associated dataset. This algorithm can automatically learn the distribution patterns of normal power consumption data, accurately identifying and removing abnormal data points caused by external power grid interference, occasional faults in test equipment, or momentary connection interruptions without the need for preset fixed thresholds, ensuring high quality and reliability of the input data for subsequent feature extraction and analysis.

[0072] 2. Extract a set of predefined key power consumption characteristics from the cleaned data. The key power consumption characteristics include at least one of the following: steady-state power consumption value of each power supply state, energy consumed during state switching process, duration of peak power consumption, energy efficiency ratio, and correlation coefficient between power consumption and temperature.

[0073] Specifically, from the cleaned, high-quality data, the system automatically calculates and extracts a predefined set of key power consumption characteristic values ​​covering static and dynamic, steady-state and transient states, including but not limited to:

[0074] Static characteristics: Steady-state average power consumption (mW) under each power saving state (PS0 to PSx);

[0075] Dynamic energy characteristics: The total energy consumed during state switching (uJ, obtained by integrating the transient power curve within the switching time window);

[0076] Peak characteristics: the cumulative duration (ms) during which peak power consumption exceeds a specific threshold (such as 150% of average power consumption) and the frequency of peak occurrence;

[0077] Energy efficiency characteristics: Energy efficiency ratio in different load ranges, such as performance / power ratio (IOPS / mW or MB / s / mW).

[0078] Thermal coupling characteristics: statistical correlation coefficients (such as Pearson coefficients) between power consumption and chip junction temperature or ambient temperature, and the approximate slope (mW / °C) of power consumption as a function of temperature.

[0079] 3. Perform cross-level correlation analysis, including at least: analyzing the correlation between power consumption data and performance data to determine the energy efficiency point, analyzing the correlation between power consumption data and internal state data to diagnose the source of high power consumption anomalies, and analyzing the correlation between power consumption data and temperature data to assess the thermal stability of the solid-state drive.

[0080] Specifically, power consumption and performance correlation analysis: By plotting a two-dimensional scatter plot of "power consumption-bandwidth" or "power consumption-IOPS", clustering algorithms (such as DBSCAN) are used to automatically identify the "sweet spot" with optimal energy efficiency and the "black hole" with extremely poor energy efficiency, providing a target operating range for system optimization.

[0081] Power consumption and internal state correlation analysis: On a unified timeline, the specific time periods of all high power consumption were located, and the internal state data of the solid-state drive at the same moment were traced back and correlated. For example, the analysis revealed that a certain sustained high power consumption plateau always strictly corresponded to the continuous setting of the "garbage collection activity flag" in the internal register and the sharp increase of the "NAND programming operation count". This confirmed that the root cause of this high power consumption anomaly was the background garbage collection operation triggered by the firmware, rather than the user data read and write itself.

[0082] Power consumption and temperature correlation analysis: Power consumption data collected at different steady-state temperatures are subjected to linear or nonlinear regression fitting to establish an empirical model of power consumption as a function of temperature (e.g., P(T) = a + b). The thermal stability of solid-state drives (SSDs) can be quantitatively assessed using the parameters of this model (such as coefficient b). Simultaneously, by observing whether periodic power consumption drops and performance fluctuations occur due to thermal throttling during high-temperature, high-load testing, the long-term high-temperature reliability of the SSD can be evaluated.

[0083] Furthermore, based on the comprehensive analysis results described above, the system automatically populates the data into a pre-defined, structured power consumption assessment report template. The generated report is not merely a list of data and charts; it may also include the following:

[0084] Executive Summary: The "Power Consumption Diagnostic Conclusion," automatically generated in natural language, incisively points out the most significant power consumption characteristics and potential problems of this solid-state drive.

[0085] Key findings: The report presents detailed quantitative results on basic power consumption, dynamic power consumption, energy efficiency, and thermal stability across multiple dimensions.

[0086] Root cause analysis: Based on correlation analysis, the underlying reasons for high power consumption or energy efficiency inflection points are clearly identified. For example, at 55°C and 95% utilization, a 128KB sequential write load causes the main controller to be under continuous high load (>95%) and NAND programming bottleneck, resulting in an average power consumption of 4.2W, which is 3 times worse than the optimal energy efficiency point.

[0087] Specific optimization recommendations: Provide actionable, data-driven improvement suggestions from multiple levels, including firmware strategy (such as "optimizing garbage collection trigger conditions"), hardware design (such as "strengthening power filtering during PSO to Active switching"), or system application (such as "avoiding prolonged high queue depth writes at high temperatures").

[0088] Raw data archiving: Package all test configurations, load models, cleaned datasets, and analysis algorithm configurations into a complete and traceable test archive to ensure that any test can be accurately reproduced and compared in the future.

[0089] The multi-dimensional power consumption testing method for solid-state drives (SSDs) in this embodiment systematically applies dynamic workloads and collects data under different combinations of ambient temperature and capacity usage conditions through a predefined test matrix. This overcomes the problem of single test conditions in traditional methods and can comprehensively reveal the power consumption behavior of SSDs under various potential operating conditions. Furthermore, by synchronizing and fusing the synchronously collected power consumption data, internal status data, and performance data, a correlated dataset containing multi-dimensional information is formed, laying a data foundation for in-depth power consumption analysis from phenomenon to root cause, and solving the drawbacks of isolated data and superficial analysis in traditional methods. In addition, by intelligently analyzing the correlated dataset, power consumption characteristics are automatically extracted and evaluation reports are generated, reducing the reliance on manual analysis. This not only improves testing efficiency but also ensures the objectivity and consistency of analysis results, realizing an automated closed loop from data collection to knowledge insight.

[0090] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0091] In one embodiment, a multi-dimensional power consumption testing device for solid-state drives (SSDs) is provided, which corresponds one-to-one with the multi-dimensional power consumption testing method for SSDs described in the above embodiments. For example... Figure 2 As shown, the multi-dimensional power consumption testing device for the solid-state drive includes an initialization module 11, a testing module 12, a data association module 13, and an analysis module 14.

[0092] Initialization module 11 is used to respond to the test start command and control the test environment parameters and solid-state drive to enter the preset initial state;

[0093] Test module 12 is used to control the test equipment to apply multi-dimensional dynamic workload to the solid-state drive based on a predefined test matrix, and to simultaneously collect power consumption data, internal status data and performance data. The test matrix defines multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions.

[0094] Data association module 13 is used to synchronize and fuse power consumption data, internal status data and performance data in time to form an associated dataset;

[0095] Analysis module 14 is used to perform intelligent analysis on the associated dataset to extract the power consumption characteristics of the solid-state drive and generate a power consumption assessment report based on the power consumption characteristics.

[0096] Optionally, the test module 12 performs operations to control the test equipment to apply multi-dimensional dynamic workloads to the solid-state drive and simultaneously collect power consumption data, internal status data, and performance data, specifically including:

[0097] For each set of ambient temperature and capacity usage conditions, the workload parameters of the solid-state drive are gradually increased from the initial value, and the power consumption data and internal status data are monitored in real time during the increase process.

[0098] Identify the power-sensitive operating points of solid-state drives based on changing trends;

[0099] With power-sensitive operating points as the core, control the test equipment to perform at least one deep test, including: state switching transient test, scenario simulation test based on real input / output trace files, or targeted burst traffic stress test.

[0100] Optionally, the test module 12 performs an operation to identify the power-sensitive operating point of the solid-state drive based on the changing trend, specifically including:

[0101] When the power consumption data growth rate exceeds the first threshold and the performance data growth rate is lower than the second threshold during the increase of workload parameters, it is determined that the current combination of workload parameters constitutes a power-sensitive operating point.

[0102] Optionally, the test module 12 performs a state transition transient test, specifically including:

[0103] Send a power state switching command to the solid-state drive;

[0104] While sending the command, a high sampling rate power consumption data acquisition operation is triggered to capture the voltage and current transient waveforms of the solid-state drive during power state switching.

[0105] Optionally, test module 12 performs scenario simulation tests based on real input / output trace files, specifically including:

[0106] Control the solid-state drive to play back pre-recorded input / output trace files;

[0107] During playback, power consumption data is collected, and the collected power consumption data stream is timestamped with the operation type in the input / output trace file to establish a correspondence between specific application operations and power consumption peaks.

[0108] Optionally, test module 12 performs targeted burst traffic stress tests, specifically including:

[0109] Based on the workload parameters corresponding to the power-sensitive operating point, a high-intensity burst traffic load is generated for a first preset duration.

[0110] During periods of sudden surges in power consumption, peak power consumption data and chip temperature data of the solid-state drive are collected simultaneously to analyze the correlation between heat dissipation performance and peak power consumption.

[0111] Optionally, the initialization module 11 performs operations to control the test environment parameters and the solid-state drive to enter a preset initial state, specifically including:

[0112] Control the test system resources to meet the preset test capability benchmarks, which include the sampling rate and accuracy of the data acquisition system, the temperature control range and accuracy of the temperature control device, and the interaction capability between the test host and the solid-state drive firmware.

[0113] The system controls the solid-state drive (SSD) to perform data preprocessing operations, achieving multiple test combinations defined in the test matrix one by one. Upon achieving each test combination, the SSD is controlled to enter an idle state. In the idle state, the SSD is controlled to cycle through all power-saving states and records the steady-state power consumption and state switching delay of each power-saving state to establish a basic power consumption profile for each test combination.

[0114] Optionally, the analysis module 14 performs intelligent analysis on the associated dataset, specifically including:

[0115] Clean up outlier data points in correlated datasets in an automated manner;

[0116] Extract a set of predefined key power consumption characteristics from the cleaned data. The key power consumption characteristics include at least one of the following: steady-state power consumption value of each power state, energy consumed during state switching process, duration of peak power consumption, energy efficiency ratio, and correlation coefficient between power consumption and temperature.

[0117] Perform cross-level correlation analysis, including at least: analyzing the correlation between power consumption data and performance data to determine the energy efficiency point, analyzing the correlation between power consumption data and internal state data to diagnose the source of high power consumption anomalies, and analyzing the correlation between power consumption data and temperature data to assess the thermal stability of the solid-state drive.

[0118] For specific limitations regarding the multi-dimensional power consumption testing device for solid-state drives (SSDs), please refer to the limitations of the multi-dimensional power consumption testing method for SSDs mentioned above, which will not be repeated here. Each module in the aforementioned multi-dimensional power consumption testing device for SSDs can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the computer device in hardware form or independent of it, or they can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0119] In one embodiment, a computer device is provided, the internal structure of which can be shown in the following diagram. Figure 3 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it performs the following steps:

[0120] In response to the test start command, control the test environment parameters and the solid-state drive to enter the preset initial state;

[0121] Based on a predefined test matrix, the test equipment is controlled to apply a multi-dimensional dynamic workload to the solid-state drive, and power consumption data, internal status data and performance data are collected simultaneously. The test matrix defines multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions.

[0122] Power consumption data, internal status data, and performance data are synchronized and fused in time to form a related dataset;

[0123] Intelligent analysis is performed on the associated dataset to extract the power consumption characteristics of solid-state drives, and a power consumption assessment report is generated based on the power consumption characteristics.

[0124] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0125] In response to the test start command, control the test environment parameters and the solid-state drive to enter the preset initial state;

[0126] Based on a predefined test matrix, the test equipment is controlled to apply a multi-dimensional dynamic workload to the solid-state drive, and power consumption data, internal status data and performance data are collected simultaneously. The test matrix defines multiple test combinations formed by multiple ambient temperature conditions and multiple capacity occupancy conditions.

[0127] Power consumption data, internal status data, and performance data are synchronized and fused in time to form a related dataset;

[0128] Intelligent analysis is performed on the associated dataset to extract the power consumption characteristics of solid-state drives, and a power consumption assessment report is generated based on the power consumption characteristics.

[0129] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0130] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other storage media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0131] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0132] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for multi-dimension power consumption test of a solid state drive, characterized in that, It comprises: In response to the test start instruction, control the test environment parameters and the solid state disk into the preset initial state; Based on the predefined test matrix, control the test equipment to apply multi-dimensional dynamic workload to the solid state disk, and synchronously collect power consumption data, internal state data and performance data, the test matrix defines a plurality of test combinations formed by a plurality of environmental temperature conditions and a plurality of capacity occupation conditions; Time synchronization and data fusion are performed on the power consumption data, the internal state data and the performance data to form a correlation data set; Intelligent analysis is performed on the correlation data set to extract the power consumption characteristics of the solid state disk, and a power consumption evaluation report is generated based on the power consumption characteristics; Wherein, the control test equipment applies multi-dimensional dynamic workload to the solid state disk, and synchronously collects power consumption data, internal state data and performance data, comprising: For each set of environmental temperature conditions and capacity occupation conditions, control the workload parameters of the solid state disk to start increasing from the initial value, and in the increasing process, the change trend of the power consumption data and the internal state data is monitored in real time; Based on the change trend, the power consumption sensitive working point of the solid state disk is identified; Taking the power consumption sensitive working point as the core, the test equipment is controlled to perform at least one depth test, the depth test including: state switching transient test, scene simulation test based on real input / output tracking file or targeted burst traffic stress test; The power consumption sensitive working point of the solid state disk is identified based on the change trend, comprising: When the growth rate of the power consumption data exceeds the first threshold and the growth rate of the performance data is lower than the second threshold during the increasing process of the workload parameters, it is determined that the current workload parameter combination constitutes a power consumption sensitive working point; The targeted burst traffic stress test includes: Based on the workload parameters corresponding to the power consumption sensitive working point, a high-intensity burst traffic load lasting for a first preset time is generated; During the application of the burst traffic load, the peak power consumption data and the chip temperature data of the solid state disk are synchronously collected, and the chip temperature data of the solid state disk is collected at a preset high frequency to obtain a temperature rise curve, and the instantaneous rising slope of the chip temperature is calculated to analyze the packaging heat dissipation characteristics of the solid state disk in combination with the peak power consumption data. 2.The method of claim 1, wherein, The state switching transient test includes: Send a power state switching instruction to the solid state disk; At the same time of sending the instruction, trigger a high sampling rate power consumption data collection operation to capture the voltage and current transient waveform of the solid state disk at the power state switching moment. 3.The method of claim 1, wherein, The scene simulation test based on the real input / output tracking file includes: Control the solid state disk to play back the pre-recorded input / output tracking file; During playback, the power consumption data is collected, and the collected power consumption data stream is time-stamped aligned with the operation type in the input / output tracking file to establish the correspondence between the specific application operation and the power consumption peak. 4.The method of claim 1, wherein, The control test environment parameters and the solid state disk into the preset initial state, comprising: controlling test system resources to meet preset test capability benchmarks, wherein the test capability benchmarks include sampling rate and accuracy of a data acquisition system, temperature control range and accuracy of a temperature control device, and interaction capability of a test host and a solid state disk firmware; controlling the solid state disk to perform data preprocessing operations, one by one reaching a plurality of test combinations defined in the test matrix, and in each reaching a test combination, controlling the solid state disk to enter an idle state, and in the idle state, controlling the solid state disk to cyclically switch all power saving states and record steady-state power consumption and state switching delay of each power saving state, to establish a basic power consumption image corresponding to each test combination. 5.The method of claim 1, wherein, The intelligent analysis of the associated data set comprises: cleaning abnormal data points in the associated data set in an automated manner; extracting a set of pre-defined key power consumption feature values from the cleaned data, the key power consumption feature values including at least one of steady-state power consumption values of each power state, energy consumed in a state switching process, duration of peak power consumption, energy efficiency ratio, and correlation coefficient between power consumption and temperature; performing cross-level correlation analysis, including at least: analyzing the correlation between the power consumption data and the performance data to determine energy efficiency points, analyzing the correlation between the power consumption data and the internal state data to diagnose the source of high power consumption anomalies, and analyzing the correlation between the power consumption data and the temperature data to evaluate the thermal stability of the solid state disk.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the multi-dimensional power consumption test method of the solid state disk according to any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, wherein the computer program comprises the following steps of: receiving a request for a resource from a client; determining whether the client is authorized to access the resource; and if the client is authorized to access the resource, providing the resource to the client. The computer program is executed by the processor to implement the steps of the multi-dimensional power consumption test method of the solid state disk according to any one of claims 1 to 5.

Images