WO2022022403A1 - 数据处理设备快速搜频方法和装置以及数据处理设备 - Google Patents

数据处理设备快速搜频方法和装置以及数据处理设备 Download PDF

Info

Publication number
WO2022022403A1
WO2022022403A1 PCT/CN2021/108054 CN2021108054W WO2022022403A1 WO 2022022403 A1 WO2022022403 A1 WO 2022022403A1 CN 2021108054 W CN2021108054 W CN 2021108054W WO 2022022403 A1 WO2022022403 A1 WO 2022022403A1
Authority
WO
WIPO (PCT)
Prior art keywords
frequency
data processing
processing device
output voltage
computing chip
Prior art date
Application number
PCT/CN2021/108054
Other languages
English (en)
French (fr)
Inventor
马伟彬
张宏
黄理洪
杨作兴
郭海丰
Original Assignee
深圳比特微电子科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳比特微电子科技有限公司 filed Critical 深圳比特微电子科技有限公司
Priority to US17/606,939 priority Critical patent/US11493981B1/en
Priority to CA3164095A priority patent/CA3164095C/en
Publication of WO2022022403A1 publication Critical patent/WO2022022403A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/4401—Bootstrapping
    • G06F9/4403—Processor initialisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/324—Power saving characterised by the action undertaken by lowering clock frequency
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/04—Generating or distributing clock signals or signals derived directly therefrom
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16—Constructional details or arrangements
    • G06F1/20—Cooling means
    • G06F1/206—Cooling means comprising thermal management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/266—Arrangements to supply power to external peripherals either directly from the computer or under computer control, e.g. supply of power through the communication port, computer controlled power-strips
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26—Power supply means, e.g. regulation thereof
    • G06F1/32—Means for saving power
    • G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234—Power saving characterised by the action undertaken
    • G06F1/3296—Power saving characterised by the action undertaken by lowering the supply or operating voltage
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the present application relates to the technical field of virtual currency mining machines, and in particular, to a method and device for fast frequency searching of data processing equipment and a data processing device using the same.
  • the working mode of fixed frequency and fixed voltage is adopted, that is, the computing chip in the mining machine adopts a preset fixed operating frequency, and the output voltage of the mining machine power supply adopts a preset fixed voltage.
  • the output voltage of the power supply will have a certain redundancy, and the fixed operating frequency of each computing chip is not its limit operating frequency.
  • the embodiments of the present application provide a method and device for fast frequency searching of data processing equipment and data processing equipment using the same, so as to shorten the frequency searching time of data processing equipment, make the data processing equipment enter the working state as soon as possible, and improve the work of the data processing equipment efficient.
  • a fast frequency search method for data processing equipment comprising the following steps performed after the data processing equipment is powered on:
  • the frequency of the computing chip in the data processing device is gradually increased from the initial frequency point to the target frequency through at least one pre-recorded first frequency point, wherein the computing chip is in each The duration of the first frequency point is shorter than the duration of the operation chip at each of the first frequency points in the conventional frequency search process;
  • the output voltage of the data processing device power supply and the frequency of the computing chip are adjusted to obtain the limit frequency of the computing chip and the output voltage of the data processing device power supply corresponding to the limit frequency.
  • the at least one first frequency point recorded in advance is at least one first frequency point passed through in the conventional frequency search process performed in advance by the data processing equipment;
  • the at least one output voltage value recorded in advance is at least one output voltage value used in the conventional frequency search process performed in advance by the data processing device;
  • the at least one second frequency point recorded in advance is at least one second frequency point used in the conventional frequency search process performed in advance by the data processing device.
  • the conventional frequency search process includes the following steps performed after the data processing device is powered on:
  • the frequency of the computing chip in the data processing device is gradually increased from the initial frequency point to the target frequency, and at least one of the first frequency points passed in the process of gradually increasing the frequency is recorded, wherein, when the computing chip is at each of the first frequency points, when the temperature of the data processing device is stable, the computing chip enters the next first frequency point, and then the data processing device When the computing chip reaches the target frequency, the data processing device reaches a stable working temperature;
  • the test task issued to the computing chip and the feedback from the computing chip are combined.
  • the output voltage of the data processing device power supply and the frequency of the computing chip are adjusted to obtain the limit frequency of the computing chip at the stable operating temperature and the data processing device power supply corresponding to the limit frequency
  • each of the output voltage values and each of the second frequency points used in the adjustment are recorded.
  • the number of test tasks delivered to the computing chip in the fast frequency search method of the data processing device is less than the number of test tasks delivered to the computing chip during the regular frequency search process.
  • the data is adjusted. Processing the output voltage of the device power supply and the frequency of the computing chip to obtain the limit frequency of the computing chip and the output voltage of the data processing device power supply corresponding to the limit frequency, including:
  • the power supply of the data processing device is adjusted in combination with the test task issued to the computing chip and the feedback result of the computing chip.
  • the output voltage and the frequency of the computing chip to obtain the limit frequency of the computing chip at the stable operating temperature and the output voltage of the data processing equipment power supply corresponding to the limit frequency including:
  • a fast frequency search device for data processing equipment comprising:
  • the storage module is used to store at least one first frequency point recorded in advance, at least one output voltage value recorded in advance and at least one second frequency point recorded in advance;
  • An execution module the execution module is used for:
  • the frequency of the arithmetic chip in the data processing device is gradually increased from the initial frequency point to the target frequency through the at least one first frequency point stored in the storage module, wherein the The duration of the operation chip at each of the first frequency points is shorter than the duration of the operation chip at each of the first frequency points in the conventional frequency search process;
  • the at least one output voltage value and the at least one second frequency point stored in the storage module are combined with the test issued to the arithmetic chip.
  • the task and the feedback result of the computing chip adjust the output voltage of the power supply of the data processing device and the frequency of the computing chip to obtain the limit frequency of the computing chip and the power supply of the data processing device corresponding to the limit frequency the output voltage.
  • a data processing device the data processing device is used for data processing of virtual currency, and the data processing device adopts the above-mentioned data processing device fast frequency search device.
  • a non-volatile computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to perform a data processing device fast search as described above. steps in the frequency method.
  • An electronic device comprising:
  • the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor executes the method for fast frequency searching for a data processing device as described above. A step of.
  • the fast frequency searching method for data processing equipment by reducing the number of test tasks, the time for the computing chip to perform the test tasks can be shortened.
  • a smaller number of test tasks can achieve the same or similar effect as the frequency and voltage optimization stage in the conventional frequency search process, and can greatly save the time occupied by the frequency and voltage optimization stage.
  • the method for fast frequency search of data processing equipment in the embodiment of the present application greatly shortens frequency search time, saves power consumption during frequency search, and the saved time is used in data processing calculation of data processing equipment. It also further improves the data processing efficiency.
  • the method for fast frequency search for a data processing device is more suitable for a scenario that requires frequent and repeated startup under the condition that the functions and performance of the data processing device are basically perfected.
  • the combination of the fast frequency search method for data processing equipment in the embodiment of the present application and the conventional frequency search can also enable the data processing equipment to obtain stable and reliable results by using the conventional frequency search in the case of extreme environmental changes.
  • the fast frequency search method of the data processing device in the embodiment of the present application and the conventional frequency search complement each other, and meet the frequency search requirements of the data processing device under various environmental conditions.
  • FIG. 1 is a flowchart of a method for fast frequency searching for a data processing device according to an embodiment of the application
  • FIG. 3 is a schematic diagram of a fast frequency search device for data processing equipment according to an embodiment of the present application
  • FIG. 4 is a schematic structural diagram of an electronic device in an embodiment of the present application.
  • the method for fast frequency searching of a data processing device mainly includes the following steps performed after the data processing device (for example, a virtual currency mining machine or a mining machine) is powered on:
  • the frequency of the computing chip in the data processing device is gradually increased from the initial frequency point to the target frequency through at least one pre-recorded first frequency point, wherein the computing chip is at each first frequency
  • the duration at the point is shorter than the duration when the computing chip is at each first frequency point in the conventional frequency search process
  • the arithmetic chip in the data processing device When the arithmetic chip in the data processing device reaches the target frequency, it adjusts according to the pre-recorded at least one output voltage value and the pre-recorded at least one second frequency point, combined with the test task issued to the arithmetic chip and the feedback result of the arithmetic chip.
  • the output voltage of the power supply of the data processing device and the frequency of the computing chip are used to obtain the limit frequency of the computing chip and the output voltage of the power supply of the data processing device corresponding to the limit frequency.
  • the process of the fast frequency search method is completed, and then the limit frequency is used as the operating frequency in the data processing (for example, virtual currency mining) process of the computing chip, and the power supply of the data processing equipment corresponding to the limit frequency is used.
  • the output voltage is used as the output voltage of the power supply of the data processing equipment during the working process.
  • the at least one first frequency point recorded in advance is at least one first frequency point passed through in the conventional frequency search process performed in advance by the data processing device;
  • the at least one output voltage value recorded in advance is at least one output voltage value used in the conventional frequency search process performed in advance by the data processing device;
  • the at least one second frequency point recorded in advance is at least one second frequency point used in the conventional frequency search process performed in advance by the data processing device.
  • the conventional frequency search process includes the following steps performed after the data processing device is powered on:
  • the frequency of the computing chip in the data processing device is gradually increased from the initial frequency point to the target frequency, and at least one first frequency point passed in the process of gradually increasing the frequency is recorded, wherein,
  • the computing chip is at each first frequency point, when the temperature of the data processing device is stable, the computing chip enters the next first frequency point, and then when the computing chip reaches the target frequency in the data processing device, the data processing device achieves stable operation temperature;
  • the power supply of the data processing device is adjusted in combination with the test tasks issued to the computing chip and the feedback results of the computing chip.
  • the output voltage and the frequency of the computing chip are used to obtain the limit frequency of the computing chip when the operating temperature is stable and the output voltage of the power supply of the data processing device corresponding to the limit frequency.
  • the temperature stability of the data processing device is, for example, that the temperature of the data processing device changes less than a preset threshold within a certain period of time. Keeping the stable working temperature of the data processing device basically unchanged is, for example, keeping the variation of the stable working temperature of the data processing device within a preset range.
  • the normal frequency search process needs to consider the heating rate of the data processing equipment.
  • the ambient temperature is too low, it is difficult for the computing chip to reach the set target frequency immediately when the data processing equipment is started. Therefore, it is necessary to combine the heating process of the data processing equipment. Let the computing chip gradually increase the frequency to reach the set target frequency.
  • the frequency of the arithmetic chip in the data processing device is divided into several first frequency points from the initial frequency point to the target frequency.
  • the initial frequency point may be set as the lowest frequency at which the computing chip can run when the data processing device is initially started up. See Table 1 for the relationship between the initial frequency point, the first frequency point and the target frequency.
  • Table 1 The initial frequency point, the first frequency point and the target frequency are arranged from low to high
  • the frequency of the initial frequency point is lower than the target frequency, and several first frequency points are set between the initial frequency point and the target frequency point, that is, the first frequency points a1, The first frequency point a2...the first frequency point an.
  • the computing chip In the frequency-up stage of the conventional frequency search process, that is, in the process that the computing chip passes through n first frequency points from the initial frequency point and then rises to the target frequency, it stays at each frequency point for enough time to wait for the The temperature of the data processing equipment at the frequency point increases and stabilizes. That is, when the computing chip is at the initial frequency point, it stays for a certain period of time, and when the temperature of the data processing equipment increases with the operation chip at the frequency of the initial frequency point to the point where the temperature no longer changes, the computing chip is raised to the first frequency point a1; When the computing chip is at the first frequency point a1, it stays for a certain period of time.
  • the computing chip When the temperature of the data processing device increases with the operation chip running at the frequency of the first frequency point a1 until the temperature no longer changes, the computing chip is raised to the first frequency point. a2; and so on, until the computing chip is raised to the target frequency, and the temperature of the data processing equipment is raised to the working temperature and stabilized at the working temperature.
  • a dwell time T1 is set at each first frequency point to ensure that the temperature of the data processing device can rise to a stable state within the time T1. Therefore, in the frequency-up stage of the conventional frequency search process, it can be ensured that the temperature of the data processing equipment increases as the frequency of the computing chip increases, and when the frequency of the computing chip increases to the target frequency, the temperature of the data processing equipment can increase and stabilize at the operating temperature. .
  • the above-mentioned control in the frequency-up phase of the conventional frequency search process can ensure the normal startup of the data processing equipment in extreme environments, such as low temperature environments.
  • extreme environments such as low temperature environments.
  • room temperature The temperature rises rapidly, and the stable temperature has been reached before each first frequency point has not stayed for the duration T1. Therefore, for a data processing device in a common environment, the T1 at each first frequency point in the frequency up stage of the conventional frequency search process will delay the execution time of the frequency search process because the time is too long.
  • the duration of the operation chip at each first frequency point is shorter than that of the operation chip in the conventional frequency search process. Duration at each first frequency point.
  • the duration of staying at each first frequency point is set to be t1, where t1 ⁇ T1. After this setting, it is possible that the temperature of the data processing equipment at a certain or some first frequency points has not yet reached a stable state and then rises to the next first frequency point, but because the temperature rises faster, it will not seriously affect the temperature. The increase in the frequency of the computing chip.
  • shortening the duration of the operation chip at each first frequency point can significantly shorten the frequency search time without affecting the frequency increase of the operation chip.
  • the data processing device that is already in the normal temperature state, and has performed the conventional frequency search process in the early stage and recorded each first frequency point in the conventional frequency search process, when it is restarted, it is not necessary to perform the conventional frequency search process again. It only needs to go through each first frequency point of the conventional frequency search process and shorten the corresponding stay time to achieve basically the same effect as the upscaling process in the conventional frequency search process, and can greatly save the time of the upscaling process. .
  • the frequency and voltage optimization stage is entered.
  • the data is adjusted in combination with the test tasks issued to the computing chip and the feedback results of the computing chip.
  • the output voltage of the power supply of the processing device and the frequency of the computing chip are obtained to obtain the limit frequency of the computing chip when the operating temperature is stable and the output voltage of the power supply of the data processing device corresponding to the limit frequency, wherein, after adjusting the output of the power supply of the data processing device In the process of voltage and frequency of the arithmetic chip, each output voltage value and each second frequency point used in the adjustment are recorded.
  • the process includes the following steps:
  • the number of output voltage values is, for example, m
  • the number of second frequency points is, for example, r
  • the power supply of the data processing device is the output voltage value b1 and the computing chip is the second frequency point c1
  • multiple test tasks are issued to the computing chip, and feedback results from the computing chip are received, wherein the computing chip executes the test task correctly
  • the result is known.
  • the reason for sending multiple test tasks to the computing chip for the computing chip to execute is because the operation of the computing chip may obtain wrong results after executing the test task due to factors such as interference.
  • the task is executed by the computing chip.
  • the feedback result of the computing chip may have a correct result or an incorrect result. Whether the feedback result of the computing chip is correct can be determined by comparing the feedback result of the computing chip with the known correct result.
  • the output voltage value b1 and the computing chip is the second frequency point c1, indicating that the data processing equipment is working stably.
  • the feedback results of the computing chips corresponding to all combinations in Table 2 are obtained, and the operating frequencies and A combination of power supply output voltages for data processing equipment.
  • the second frequency point and the output voltage value in the group with the highest second frequency point are used as the operation of the data processing equipment under the condition that the stable operating temperature remains unchanged.
  • the operating frequency of the chip and the output voltage of the power supply of the data processing equipment are used as the operation of the data processing equipment under the condition that the stable operating temperature remains unchanged.
  • the number of test tasks sent to the computing chip is less than that sent to the computing chip during the conventional frequency searching process. The number of test tasks issued.
  • the pre-recording refers to the recording in the regular frequency searching process before the fast frequency searching method of the data processing device is executed.
  • the process includes the following steps:
  • the fast frequency search method for data processing equipment in the embodiment of the present application shortens the waiting time of each frequency point in the up-frequency stage and reduces the frequency and voltage optimization stage according to the relevant parameters (such as the first frequency point) recorded during the conventional frequency search.
  • the number of test tasks can be increased, thereby accelerating the frequency search process as a whole and saving a lot of frequency search time.
  • Regular frequency searches take into account extreme operating environments, such as low temperature environments.
  • the data processing equipment In a low temperature environment, the data processing equipment needs to have enough time to increase the temperature of the hash board, so it is necessary to ensure that the temperature of each link is as expected.
  • the fast frequency search method for a data processing device in the embodiment of the present application does not perform too many waiting operations for the temperature, and does not affect the accuracy of the frequency search result in most non-extreme environments.
  • the conventional frequency search will repeatedly determine the output voltage point of the power supply by issuing a large number of test tasks.
  • the fast frequency search method for data processing equipment in the embodiment of the present application adopts the relevant parameters (such as output voltage value, second frequency point) recorded in the conventional frequency search performed in advance, and reduces the process of repeated confirmation by reducing the test tasks, so as to achieve The purpose of quickly regulating the output voltage of the power supply.
  • the method for fast frequency searching of a data processing device in the embodiment of the present application can greatly shorten the frequency searching time after the data processing device is turned on in a non-extreme environment.
  • the conventional frequency searching mode takes about 30 minutes.
  • the fast frequency search method of data processing equipment only needs 5 to 10 minutes to complete the frequency search process, saving at least 20 minutes of time. In the data processing calculation of the data processing device, the data processing efficiency is further improved.
  • the fast frequency search method of the data processing device in the embodiment of the present application has poor adaptability to some extreme environments, and the stability will be reduced to some extent, however, in most of the time data processing
  • the device is in a non-extreme environment, and in a non-extreme environment, the fast frequency search method for data processing devices in the embodiments of the present application is more suitable for scenarios requiring frequent and repeated startups while ensuring that functions and performance are basically perfect.
  • the combination of the fast frequency search method for mining machine data processing equipment according to the embodiment of the present invention and the conventional frequency search can also enable the mining machine data processing equipment to use the conventional frequency search to obtain stable and reliable data in the case of extreme environmental changes.
  • the fast frequency search method of the mining machine data processing equipment according to the embodiment of the present invention and the conventional frequency search complement each other, and satisfy the mining machine data processing equipment in various environmental conditions. frequency requirements below.
  • the embodiment of the present application further provides a device for fast frequency searching of data processing equipment, as shown in FIG. 3 , including a storage module 1 and an execution module 2 .
  • the storage module 1 is used for storing at least one first frequency point recorded in advance, at least one output voltage value recorded in advance and at least one second frequency point recorded in advance.
  • the execution module 2 is configured to: under the initial output voltage of the power supply of the data processing device, gradually increase the frequency of the arithmetic chip in the data processing device from the initial frequency point to the target frequency through at least one first frequency point stored in the storage module 1, wherein , the duration when the computing chip is at each first frequency point is shorter than the duration when the computing chip is at each first frequency point in the conventional frequency search process; when the computing chip in the data processing device reaches the target frequency, according to the storage module 1
  • the stored at least one output voltage value and at least one second frequency point combined with the test task issued to the computing chip and the feedback result of the computing chip, adjust the output voltage of the power supply of the data processing equipment and the frequency of the computing chip to obtain The limit frequency of the computing chip and the output voltage of the power supply of the data processing equipment corresponding to the limit frequency.
  • the embodiment of the present application also provides a data processing device, the data processing device is used for data processing of virtual currency, and the data processing device adopts the above-mentioned data processing device fast frequency search device.
  • the embodiments of the present application also provide a non-volatile computer-readable storage medium, where the non-volatile computer-readable storage medium stores instructions, which, when executed by a processor, cause the processor to execute as described in the above description Each step in the fast frequency searching method of the data processing equipment.
  • the embodiment of the present application also provides an electronic device for performing a fast frequency search method for a data processing device.
  • the electronic device includes: at least one processor 10 and a memory 20 .
  • the memory 20 is communicatively connected to the at least one processor 10, eg, the memory 20 and the at least one processor 10 are connected via a bus.
  • the memory 20 stores instructions executable by the at least one processor 10, the instructions are executed by the at least one processor 10, so that the at least one processor 10 executes the various steps in the method for fast frequency searching for a data processing device as described above .

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Human Computer Interaction (AREA)
  • Computer Hardware Design (AREA)
  • Computer Security & Cryptography (AREA)
  • Power Sources (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种数据处理设备快速搜频方法和装置以及数据处理设备，该方法包括：在数据处理设备电源的初始输出电压下，将运算芯片的频率由初始频率点经由至少一个第一频率点逐步提升至目标频率，其中，运算芯片处于每个第一频率点时的持续时长短于常规搜频过程中运算芯片处于每个第一频率点时的持续时长；当运算芯片达到目标频率后，按照至少一个输出电压值和至少一个第二频率点，并结合向运算芯片下发的测试任务和运算芯片的反馈结果，调整数据处理设备电源的输出电压和运算芯片的频率，以获得运算芯片的极限频率和与该极限频率对应的数据处理设备电源的输出电压。

Description

数据处理设备快速搜频方法和装置以及数据处理设备
本申请要求2020年7月30日提交的申请号为202010748968.8、发明名称为“矿机快速搜频方法和装置以及矿机”的中国专利申请的优先权，所述申请的内容通过引用并入本申请中。
技术领域
本申请涉及虚拟货币矿机技术领域，特别涉及一种数据处理设备快速搜频方法和装置以及采用该装置的数据处理设备。
背景技术
现有的虚拟货币矿机中，采用固定频率和固定电压的工作方式，即矿机中计算芯片采用事先设定的固定工作频率，矿机电源的输出电压采用事先设定的固定电压。采用这种方式，为了确保矿机的正常工作，电源的输出电压会有一定冗余，并且每个运算芯片的固定工作频率也并非是其极限工作频率。
发明内容
本申请实施例提供一种数据处理设备快速搜频方法和装置以及采用该装置的数据处理设备，以缩短数据处理设备搜频的时间，使得数据处理设备尽早进入工作状态，提升数据处理设备的工作效率。
本申请实施例的技术方案是这样实现的：
一种数据处理设备快速搜频方法，包括数据处理设备开机后执行的如下步骤：
在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点经由事先记录的至少一个第一频率点逐步提升至目标频率，其中，所述运算芯片处于每个所述第一频率点时的持续时长短于常规搜频过程中所述运算芯片处于每个所述第一频率点时的持续时长；
当所述数据处理设备中运算芯片达到目标频率后，按照事先记录的至少一个输出电压值和事先记录的至少一个第二频率点，并结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压。
进一步，所述事先记录的至少一个第一频率点为所述数据处理设备事先进行的所述常规 搜频过程中所经过的至少一个第一频率点；
所述事先记录的至少一个输出电压值为所述数据处理设备事先进行的所述常规搜频过程中所采用的至少一个输出电压值；
所述事先记录的至少一个第二频率点为所述数据处理设备事先进行的所述常规搜频过程中所采用的至少一个第二频率点。
进一步，所述常规搜频过程包括数据处理设备开机后执行的如下步骤：
在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点逐步提升至目标频率，并记录在频率逐步提升过程中所经过的至少一个所述第一频率点，其中，在所述运算芯片处于每个所述第一频率点时，当所述数据处理设备的温度稳定后，所述运算芯片进入下一个所述第一频率点，进而在所述数据处理设备中运算芯片达到目标频率时，所述数据处理设备达到稳定工作温度；
当所述数据处理设备中运算芯片达到目标频率后，在保持所述数据处理设备的稳定工作温度基本不变的条件下，结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得在所述稳定工作温度时的所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压，其中，在调整所述数据处理设备电源的输出电压和所述运算芯片的频率过程中，记录调整时所采用的每一个所述输出电压值和每一个所述第二频率点。
进一步，在所述数据处理设备快速搜频方法中向所述运算芯片下发的测试任务的数量少于在所述常规搜频过程中向所述运算芯片下发的测试任务的数量。
进一步，所述的按照事先记录的至少一个输出电压值和事先记录的至少一个第二频率点，并结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压，包括：
在任意一个所述输出电压值和任意一个所述第二频率点条件下，向所述运算芯片下发多次测试任务；
接收所述运算芯片的反馈结果，根据对应于所述多次测试任务的多次反馈结果，确定所述运算芯片在当前第二频率点是否能够正常工作；
遍历所有所述输出电压值和在每个所述输出电压值下的所有第二频率点，以确定出所述运算芯片能够正常工作的最高第二频率点和与该最高第二频率点对应的输出电压值，并将该最高第二频率点作为所述运算芯片的极限频率，将与该最高第二频率点对应的输出电压值作为与该极限频率对应的所述数据处理设备电源的输出电压。
进一步，所述的在保持所述数据处理设备的稳定工作温度基本不变的条件下，结合向所 述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得在所述稳定工作温度时的所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压，包括：
保持所述数据处理设备的稳定工作温度基本不变；
在任意一个所述输出电压值和任意一个所述第二频率点条件下，向所述运算芯片下发多次测试任务；
接收所述运算芯片的反馈结果，根据对应于所述多次测试任务的多次反馈结果，确定所述运算芯片在所述稳定工作温度条件下的当前第二频率点是否能够正常工作；
遍历所有所述输出电压值和在每个所述输出电压值下的所有第二频率点，以确定出所述运算芯片在所述稳定工作温度条件下能够正常工作的最高第二频率点和与该最高第二频率点对应的输出电压值，并将该最高第二频率点作为所述稳定工作温度条件下的所述运算芯片的极限频率，将与该最高第二频率点对应的输出电压值作为所述稳定工作温度条件下的与该极限频率对应的所述数据处理设备电源的输出电压。
一种数据处理设备快速搜频装置，包括：
存储模块，所述存储模块用于存储事先记录的至少一个第一频率点、事先记录的至少一个输出电压值和事先记录的至少一个第二频率点；
执行模块，所述执行模块用于：
在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点经由所述存储模块所存储的所述至少一个第一频率点逐步提升至目标频率，其中，所述运算芯片处于每个所述第一频率点时的持续时长短于常规搜频过程中所述运算芯片处于每个所述第一频率点时的持续时长；
当所述数据处理设备中运算芯片达到目标频率后，按照所述存储模块所存储的所述至少一个输出电压值和所述至少一个第二频率点，并结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压。
一种数据处理设备，所述数据处理设备用于虚拟货币的数据处理，所述数据处理设备采用如上所述的数据处理设备快速搜频装置。
一种非易失性计算机可读存储介质，所述非易失性计算机可读存储介质存储指令，所述指令在由处理器执行时使得所述处理器执行如上所述的数据处理设备快速搜频方法中的步骤。
一种电子设备，包括：
至少一个处理器；以及，
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器执行如上所述的数据处理设备快速搜频方法中的步骤。
从上述方案可以看出，本申请实施例的数据处理设备快速搜频方法中，缩短运算芯片处于每个第一频率点时的持续时长，能够显著缩短搜频时间并且不影响运算芯片频率提升的效果，对于处于非极端环境，并且已经在前期执行过常规搜频过程的数据处理设备来说，在再次启动时，只需要缩短各个第一频率点的停留时长便能够达到与常规搜频过程中的升频过程相同的效果，并且能够极大地节省升频过程的时间。本申请实施例的数据处理设备快速搜频方法中，通过减少测试任务的数量，能够缩短运算芯片执行测试任务的时间，在再次启动时，便可不必再次执行过多的测试任务，仅需要执行较少量的测试任务便能够达到与常规搜频过程中的频率电压优化阶段相同或相近似的效果，并且能够极大地节省频率电压优化阶段占用的时间。与常规搜频相比，本申请实施例的数据处理设备快速搜频方法大大的缩短了搜频时间，节省了搜频时的耗电，并且节省出的时间用在数据处理设备的数据处理计算还进一步提升了数据处理效率。与常规搜频相比，本申请实施例的数据处理设备快速搜频方法在保证数据处理设备功能与性能基本完善的情况下，更加适用于需要频繁反复启动的场景。另外，本申请实施例的数据处理设备快速搜频方法与常规搜频相结合，还能够使得数据处理设备在出现极端环境变化的情况下，利用常规搜频获得稳定可靠的结果，在非极端环境中，达到快速开机的目的，本申请实施例的数据处理设备快速搜频方法与常规搜频二者相辅相成，满足了数据处理设备在多种环境条件下的搜频要求。
附图说明
图1为本申请实施例的数据处理设备快速搜频方法流程图；
图2为本申请实施例中的常规搜频过程流程图；
图3为本申请实施例的数据处理设备快速搜频装置示意图；
图4为本申请实施例中的电子设备的结构示意图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下参照附图并举实施例，对本申请作进一步详细说明。
如前所述，现有的虚拟货币矿机中，采用固定频率和固定电压的工作方式。虽然这种设置能够确保矿机的正常工作，但是冗余的电压造成了矿机功率的浪费，提升了虚拟货币矿场 的无用功耗，并且，固定工作频率也限制了运算芯片性能的发挥，降低了挖矿效率。所以，采用固定频率和固定电压的工作方式，一方面增加了虚拟货币矿场的无用功耗，另一方面降低了挖矿效率，从而整体上大幅增加了矿机和矿场的运行成本。
放开频率和电压限制，并在开机时采用搜频方式确定出矿机中运算芯片工作时的极限频率和矿机电源的输出电压，既可以提升挖矿效率也可以避免无用功耗，从而节约大量用电消耗，降低矿机和矿场的运行成本。然而，由于矿机的工作还受到温度的影响，温度过高和过低都不利于矿机的算力的发挥，在搜频过程中还需要考虑矿机温度随运算芯片频率的增加而逐步升高，所以在搜频过程中存在着等待矿机温度同步提升的过程，该过程占用了大量的时间，使得搜频过程过长，矿机开机后的搜频过程往往需要花费30分钟左右的时间，并且在搜频过程中矿机并不能提供正常的算力，如果存在矿机频繁启动的情况，则极大地缩减了矿机正常挖矿的时间，从而在整体上降低了矿机和矿场的效率。
因此，如何缩短矿机搜频的时间，使得矿机尽早进入工作状态，提升矿机的工作效率，便成为亟待解决的问题。
如图1所示，本申请实施例的数据处理设备快速搜频方法，主要包括数据处理设备(例如，虚拟货币矿机或矿机)开机后执行的如下步骤：
在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点经由事先记录的至少一个第一频率点逐步提升至目标频率，其中，运算芯片处于每个第一频率点时的持续时长短于常规搜频过程中运算芯片处于每个第一频率点时的持续时长；
当数据处理设备中运算芯片达到目标频率后，按照事先记录的至少一个输出电压值和事先记录的至少一个第二频率点，并结合向运算芯片下发的测试任务和运算芯片的反馈结果，调整数据处理设备电源的输出电压和运算芯片的频率，以获得运算芯片的极限频率和与该极限频率对应的数据处理设备电源的输出电压。
至此，便完成了快速搜频方法的过程，之后便将该极限频率作为运算芯片的数据处理(例如，虚拟货币挖矿)过程中的工作频率，将与该极限频率对应的数据处理设备电源的输出电压作为数据处理设备电源在工作过程中的输出电压。
其中，事先记录的至少一个第一频率点为数据处理设备事先进行的常规搜频过程中所经过的至少一个第一频率点；
事先记录的至少一个输出电压值为数据处理设备事先进行的常规搜频过程中所采用的至少一个输出电压值；
事先记录的至少一个第二频率点为数据处理设备事先进行的常规搜频过程中所采用的至少一个第二频率点。
为说明本申请实施例的数据处理设备快速搜频方法，需要介绍本申请实施例中所涉及的 常规搜频过程。
如图2所示，在本申请实施例中，常规搜频过程包括数据处理设备开机后执行的如下步骤：
在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点逐步提升至目标频率，并记录在频率逐步提升过程中所经过的至少一个第一频率点，其中，在运算芯片处于每个第一频率点时，当数据处理设备的温度稳定后，运算芯片进入下一个第一频率点，进而在数据处理设备中运算芯片达到目标频率时，数据处理设备达到稳定工作温度；
当数据处理设备中运算芯片达到目标频率后，在保持数据处理设备的稳定工作温度基本不变的条件下，结合向运算芯片下发的测试任务和运算芯片的反馈结果，调整数据处理设备电源的输出电压和运算芯片的频率，以获得在稳定工作温度时的运算芯片的极限频率和与该极限频率对应的数据处理设备电源的输出电压，其中，在调整数据处理设备电源的输出电压和运算芯片的频率过程中，记录调整时所采用的每一个输出电压值和每一个第二频率点。数据处理设备的温度稳定例如为数据处理设备的温度在一定的时间内的变化小于一个预设的阈值。保持数据处理设备的稳定工作温度基本不变例如为，保持数据处理设备的稳定工作温度的变化在一个预设的范围内。
常规搜频过程需要考虑数据处理设备的升温速度，当环境温度过低时，在数据处理设备启动时的运算芯片难以立即达到所设定的目标频率，因此，需要结合数据处理设备运行的升温过程让运算芯片逐步提升频率，从而达到所设定的目标频率。
如上所述，在常规搜频过程中，将数据处理设备中运算芯片的频率从初始频率点到目标频率之间划分若干个第一频率点。其中，在一实施例中，初始频率点可以设置为在数据处理设备启动初始时运算芯片能够运行的最低频率。初始频率点、第一频率点和目标频率之间的关系可参见表1。
表1初始频率点、第一频率点和目标频率由低到高排列
Figure PCTCN2021108054-appb-000001
如表1所示，初始频率点的频率低于目标频率，在初始频率点和目标频率点之间设置若干个第一频率点，即按照频率从低到高的顺序设置第一频率点a1、第一频率点a2……第一频率点an。
在常规搜频过程的升频阶段，即运算芯片由初始频率点依次经过n个第一频率点后升至目标频率的过程中，在每一个频率点处都停留足够的时间，以等待在该频率点处的数据处理设备温度升高并稳定。即，运算芯片处于初始频率点时，停留一定时间，待数据处理设备温 度随着运算芯片以初始频率点的频率运行以升高到温度不再变化时，运算芯片提升至第一频率点a1；运算芯片处于第一频率点a1时，停留一定时间，待数据处理设备温度随着运算芯片以第一频率点a1的频率运行以升高到温度不再变化时，运算芯片提升至第一频率点a2；依此类推，直到运算芯片提升至目标频率，并且数据处理设备温度随之升高到工作温度并稳定在工作温度。
由于在升频阶段的任意一个第一频率点处，数据处理设备温度的变化可能并不明显，难以确保数据处理设备温度在任意一个第一频率点均达到稳定。所以，在一实施例中，在常规搜频过程的升频阶段，设定在每一个第一频率点处停留时长T1，确保在时长T1内数据处理设备温度能够升至稳定状态。从而，在常规搜频过程的升频阶段，能够确保数据处理设备温度随着运算芯片的频率提升而提升，并且运算芯片的频率提升至目标频率时，数据处理设备温度能够提升并稳定在工作温度。
上述在常规搜频过程的升频阶段的控制能够确保在极限环境下，如低温环境，数据处理设备的正常开机，然而，对于处于普通环境，例如室温条件中的数据处理设备而言，开机后温度上升速度较快，在每一个第一频率点处尚未停留到时长T1时，便已经达到了稳定温度。因此，对于处于普通环境中的数据处理设备来说，常规搜频过程的升频阶段中每一个第一频率点处的T1会因为时间过长而拖延搜频过程的执行时间。
因此，如上述说明中所述，本申请实施例的数据处理设备快速搜频方法中，在升频阶段，运算芯片处于每个第一频率点时的持续时长短于常规搜频过程中运算芯片处于每个第一频率点时的持续时长。例如，在本申请实施例的数据处理设备快速搜频方法中，设定在每一个第一频率点处停留时长为t1，其中t1<T1。这样设置后，有可能在某个或者某些第一频率点处的数据处理设备温度尚未升至稳定状态便提升到下一个第一频率点，但因为温度上升速度较快，不会严重影响到运算芯片频率的提升。进而在整体上来说，在数据处理设备快速搜频方法中，缩短运算芯片处于每个第一频率点时的持续时长，能够获得显著缩短搜频时间并且不影响运算芯片频率提升的效果。这对于已经处于常温状态，并且已经在前期执行过常规搜频过程并记录了常规搜频过程中的各个第一频率点的数据处理设备来说，在再次启动时，便可不必再次执行常规搜频过程，只需要经过常规搜频过程的各个第一频率点并缩短相应的停留时长便能够达到与常规搜频过程中的升频过程基本相同的效果，并且能够极大地节省升频过程的时间。
当数据处理设备中运算芯片达到目标频率并且数据处理设备达到工作温度时，便进入频率电压优化阶段。
如上所述，在常规搜频过程中的频率电压优化阶段：在保持数据处理设备的稳定工作温度基本不变的条件下，结合向运算芯片下发的测试任务和运算芯片的反馈结果，调整数据处 理设备电源的输出电压和运算芯片的频率，以获得在稳定工作温度时的运算芯片的极限频率和与该极限频率对应的数据处理设备电源的输出电压，其中，在调整数据处理设备电源的输出电压和运算芯片的频率过程中，记录调整时所采用的每一个输出电压值和每一个第二频率点。
该过程包括以下步骤：
保持数据处理设备的稳定工作温度基本不变；
在任意一个输出电压值和任意一个第二频率点条件下，向运算芯片下发多次测试任务；
接收运算芯片的反馈结果，根据对应于多次测试任务的多次反馈结果，确定运算芯片在稳定工作温度条件下的当前第二频率点是否能够正常工作；
遍历所有输出电压值和在每个输出电压值下的所有第二频率点，以确定出运算芯片在稳定工作温度条件下能够正常工作的最高第二频率点和与该最高第二频率点对应的输出电压值，并将该最高第二频率点作为稳定工作温度条件下的运算芯片的极限频率，将与该最高第二频率点对应的输出电压值作为稳定工作温度条件下的与该极限频率对应的数据处理设备电源的输出电压。
具体来说，输出电压值数量例如为m个，第二频率点数量例如为r个，那么遍历所有输出电压值和在每个输出电压值下的所有第二频率点，最多可获得表2所示的组合。
表2输出电压值和第二频率点的组合
Figure PCTCN2021108054-appb-000002
对于表2中每一个输出电压值和第二频率点的组合，均向运算芯片下发多次测试任务， 并接收运算芯片的反馈结果。
例如，在数据处理设备电源为输出电压值b1并且运算芯片为第二频率点c1时，向运算芯片下发多次测试任务，并接收运算芯片的反馈结果，其中，运算芯片执行测试任务的正确结果是已知的，向运算芯片下发多次测试任务让运算芯片执行，是因为运算芯片工作可能会因为干扰等因素而导致执行测试任务后获得了错误结果，向运算芯片下发多次测试任务让运算芯片执行，这样，运算芯片的反馈结果可能有正确结果，也可能有错误结果，运算芯片的反馈结果是否正确可以通过将运算芯片的反馈结果和已知的正确结果进行比较而确定。举例而言，运算芯片执行多次测试任务所反馈结果中，正确结果的数量达到设定数量时，则可认为输出电压值b1并且运算芯片为第二频率点c1，是数据处理设备在稳定工作温度不变的情况下的确保数据处理设备正常工作的一种运算芯片工作频率和数据处理设备电源输出电压的组合；否则，若正确结果的数量未达到设定数量，则认为输出电压值b1并且运算芯片为第二频率点c1不能保证数据处理设备的正常工作，从而可以舍弃输出电压值b1并且运算芯片为第二频率点c1的组合。
通过这种方式，获得表2中所有组合所对应的运算芯片的反馈结果，并确定出数据处理设备在稳定工作温度不变的情况下的所有能够确保数据处理设备正常工作的运算芯片工作频率和数据处理设备电源输出电压的组合。在这些能够确保数据处理设备正常工作的组合中，将第二频率点最高的那一组中的第二频率点和输出电压值，分别作为数据处理设备在稳定工作温度不变的情况下的运算芯片的工作频率和数据处理设备电源输出电压。
由于数据处理设备中的运算芯片数量众多，对于每一个运算芯片也需要下发多次测试任务，并且还需要将所有运算芯片的工作频率和各自所对应的数据处理设备电源输出电压进行综合评判，以确定出最优的电源输出电压和对应的各个运算芯片的工作频率，所以，在常规搜频过程中的频率电压优化阶段会在执行众多测试任务上消耗大量的时间。这种情况对于长期处于同一个环境中的数据处理设备来说，虽然每次开机的微环境可能稍有不同(例如环境温度、电网电压的微弱波动等)，每次开机均执行一次这样的频率电压优化，会占用大量的时间，并且最终所得到的运算芯片的工作频率和数据处理设备电源输出电压的却不会存在过大差异，这势必造成频率电压优化效率低下的问题。
因此，在一实施例中，本申请实施例的数据处理设备快速搜频方法中，在频率电压优化阶段，向运算芯片下发的测试任务的数量少于在常规搜频过程中向运算芯片下发的测试任务的数量。
这对于已经在前期执行过常规搜频过程并记录了常规搜频过程中的各个输出电压值和各个第二频率点的数据处理设备来说，在再次启动时，便可不必再次执行常规搜频过程，在频率电压优化阶段对于每一个输出电压值和第二频率点的组合减少下发的测试任务数量便能够 达到与常规搜频过程中的频率电压优化阶段相同或相近似的效果，并且能够极大地节省频率电压优化阶段的时间。
在本申请实施例中，在数据处理设备快速搜频方法中的频率电压优化阶段：按照事先记录的至少一个输出电压值和事先记录的至少一个第二频率点，并结合向运算芯片下发的测试任务和运算芯片的反馈结果，调整数据处理设备电源的输出电压和运算芯片的频率，以获得运算芯片的极限频率和与该极限频率对应的数据处理设备电源的输出电压。其中，事先记录是指执行数据处理设备快速搜频方法之前的常规搜频过程中的记录。
该过程包括以下步骤：
在任意一个输出电压值和任意一个第二频率点条件下，向运算芯片下发多次测试任务；
接收运算芯片的反馈结果，根据对应于多次测试任务的多次反馈结果，确定运算芯片在当前第二频率点是否能够正常工作；
遍历所有输出电压值和在每个输出电压值下的所有第二频率点，以确定出运算芯片能够正常工作的最高第二频率点和与该最高第二频率点对应的输出电压值，并将该最高第二频率点作为运算芯片的极限频率，将与该最高第二频率点对应的输出电压值作为与该极限频率对应的数据处理设备电源的输出电压。
本申请实施例的数据处理设备快速搜频方法，根据常规搜频时记录的相关参数(如第一频率点)，将其中升频阶段各个频率点的等待时间缩短，并减少了频率电压优化阶段的测试任务的数量，从而整体上加速了搜频过程，节省了大量的搜频时间。
常规搜频会考虑极端的运行环境，例如低温环境。低温环境下，数据处理设备需要有足够长的时间去提升算力板的温度，所以需要确保每个环节的温度是否达到预期。而本申请实施例的数据处理设备快速搜频方法不再对温度做过多的等待操作，在绝大多数的非极端环境中，并不会影响搜频结果的准确性。
为了让数据处理设备运行功耗尽量小，常规搜频会通过下发大量的测试任务来反复确定电源的输出电压点。而本申请实施例的数据处理设备快速搜频方法则采用事前执行的常规搜频所记录的相关参数(如输出电压值、第二频率点)，通过减少测试任务而减少反复确认的流程，达到快速调节电源输出电压的目的。
数据处理设备启动过程中，为了达到性能的最优，会有相对复杂的搜频流程。在实际应用的大部分场景中，并不需要考虑一些极端因素对搜频结果的影响。继而可以简化部分搜频流程，达到缩短搜频时间的目的。本申请实施例的数据处理设备快速搜频方法，在非极端环境下可极大地缩短数据处理设备开机后的搜频时间，常规搜频模式需要花费30分钟左右的时间，而本申请实施例的数据处理设备快速搜频方法只需要5至10分钟的时间便可完成搜频流程，节省了至少20分钟的时间，一方面节省了搜频时的耗电，另一方面节省出的时间用在数 据处理设备的数据处理计算中还进一步提升了数据处理效率。
虽然相对于常规搜频模式而言，本申请实施例的数据处理设备快速搜频方法对于某些极端环境的适应性比较差，稳定性也会有所降低，但是，在绝大多数时候数据处理设备是处于非极端环境中的，进而在非极端环境中，本申请实施例的数据处理设备快速搜频方法能够在保证功能与性能基本完善的情况下，更加适用于需要频繁反复启动的场景。
另外，本发明本申请实施例的矿机数据处理设备快速搜频方法与常规搜频相结合，还能够使得矿机数据处理设备在出现极端环境变化的情况下，利用常规搜频获得稳定可靠的结果，在非极端环境中，达到快速开机的目的，本发明本申请实施例的矿机数据处理设备快速搜频方法与常规搜频二者相辅相成，满足了矿机数据处理设备在多种环境条件下的搜频要求。
本申请实施例还提供了一种数据处理设备快速搜频装置，如图3所示，包括存储模块1和执行模块2。其中，存储模块1用于存储事先记录的至少一个第一频率点、事先记录的至少一个输出电压值和事先记录的至少一个第二频率点。
执行模块2用于：在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点经由存储模块1所存储的至少一个第一频率点逐步提升至目标频率，其中，运算芯片处于每个第一频率点时的持续时长短于常规搜频过程中运算芯片处于每个第一频率点时的持续时长；当数据处理设备中运算芯片达到目标频率后，按照存储模块1所存储的至少一个输出电压值和至少一个第二频率点，并结合向运算芯片下发的测试任务和运算芯片的反馈结果，调整数据处理设备电源的输出电压和运算芯片的频率，以获得运算芯片的极限频率和与该极限频率对应的数据处理设备电源的输出电压。
执行模块2的进一步执行过程说明可参见上述数据处理设备快速搜频方法实施例。
本申请实施例还提供了一种数据处理设备，该数据处理设备用于虚拟货币的数据处理，该数据处理设备采用如上所述的数据处理设备快速搜频装置。
本申请实施例还同时提供一种非易失性计算机可读存储介质，该非易失性计算机可读存储介质存储指令，该指令在由处理器执行时使得所述处理器执行如上述说明中的数据处理设备快速搜频方法中的各个步骤。
本申请实施例还同时提供一种执行数据处理设备快速搜频方法的电子设备，如图4所示，该电子设备包括：至少一个处理器10以及存储器20。存储器20和至少一个处理器10通信连接，例如存储器20和至少一个处理器10通过总线连接。存储器20存储有可被至少一个处理器10执行的指令，所述指令被至少一个处理器10执行，以使至少一个处理器10执行如上述说明中的数据处理设备快速搜频方法中的各个步骤。
以上所述仅为本申请的较佳实施例而已，并不用以限制本申请，凡在本申请的精神和原则之内，所做的任何修改、等同替换、改进等，均应包含在本申请保护的范围之内。

Claims (10)

  1. 一种数据处理设备快速搜频方法，包括数据处理设备开机后执行的如下步骤：
    在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点经由事先记录的至少一个第一频率点逐步提升至目标频率，其中，所述运算芯片处于每个所述第一频率点时的持续时长短于常规搜频过程中所述运算芯片处于每个所述第一频率点时的持续时长；
    当所述数据处理设备中运算芯片达到目标频率后，按照事先记录的至少一个输出电压值和事先记录的至少一个第二频率点，并结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压。
  2. 根据权利要求1所述的数据处理设备快速搜频方法，其中，
    所述事先记录的至少一个第一频率点为所述数据处理设备事先进行的所述常规搜频过程中所经过的至少一个第一频率点；
    所述事先记录的至少一个输出电压值为所述数据处理设备事先进行的所述常规搜频过程中所采用的至少一个输出电压值；
    所述事先记录的至少一个第二频率点为所述数据处理设备事先进行的所述常规搜频过程中所采用的至少一个第二频率点。
  3. 根据权利要求1所述的数据处理设备快速搜频方法，其中，所述常规搜频过程包括数据处理设备开机后执行的如下步骤：
    在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点逐步提升至目标频率，并记录在频率逐步提升过程中所经过的至少一个所述第一频率点，其中，在所述运算芯片处于每个所述第一频率点时，当所述数据处理设备的温度稳定后，所述运算芯片进入下一个所述第一频率点，进而在所述数据处理设备中运算芯片达到目标频率时，所述数据处理设备达到稳定工作温度；
    当所述数据处理设备中运算芯片达到目标频率后，在保持所述数据处理设备的稳定工作温度基本不变的条件下，结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得在所述稳定工作温度时的所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压，其中，在调整所述数据处理设备电源的输出电压和所述运算芯片的频率过程中，记录调整时所采用的每一个所述输出电压值和每一个所述第二频率点。
  4. 根据权利要求3所述的数据处理设备快速搜频方法，其中，
    在所述数据处理设备快速搜频方法中向所述运算芯片下发的测试任务的数量少于在所述 常规搜频过程中向所述运算芯片下发的测试任务的数量。
  5. 根据权利要求1所述的数据处理设备快速搜频方法，其中，所述的按照事先记录的至少一个输出电压值和事先记录的至少一个第二频率点，并结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压，包括：
    在任意一个所述输出电压值和任意一个所述第二频率点条件下，向所述运算芯片下发多次测试任务；
    接收所述运算芯片的反馈结果，根据对应于所述多次测试任务的多次反馈结果，确定所述运算芯片在当前第二频率点是否能够正常工作；
    遍历所有所述输出电压值和在每个所述输出电压值下的所有第二频率点，以确定出所述运算芯片能够正常工作的最高第二频率点和与该最高第二频率点对应的输出电压值，并将该最高第二频率点作为所述运算芯片的极限频率，将与该最高第二频率点对应的输出电压值作为与该极限频率对应的所述数据处理设备电源的输出电压。
  6. 根据权利要求3所述的数据处理设备快速搜频方法，其中，所述的在保持所述数据处理设备的稳定工作温度基本不变的条件下，结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得在所述稳定工作温度时的所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压，包括：
    保持所述数据处理设备的稳定工作温度基本不变；
    在任意一个所述输出电压值和任意一个所述第二频率点条件下，向所述运算芯片下发多次测试任务；
    接收所述运算芯片的反馈结果，根据对应于所述多次测试任务的多次反馈结果，确定所述运算芯片在所述稳定工作温度条件下的当前第二频率点是否能够正常工作；
    遍历所有所述输出电压值和在每个所述输出电压值下的所有第二频率点，以确定出所述运算芯片在所述稳定工作温度条件下能够正常工作的最高第二频率点和与该最高第二频率点对应的输出电压值，并将该最高第二频率点作为所述稳定工作温度条件下的所述运算芯片的极限频率，将与该最高第二频率点对应的输出电压值作为所述稳定工作温度条件下的与该极限频率对应的所述数据处理设备电源的输出电压。
  7. 一种数据处理设备快速搜频装置，包括：
    存储模块，所述存储模块用于存储事先记录的至少一个第一频率点、事先记录的至少一个输出电压值和事先记录的至少一个第二频率点；
    执行模块，所述执行模块用于：
    在数据处理设备电源的初始输出电压下，将数据处理设备中运算芯片的频率由初始频率点经由所述存储模块所存储的所述至少一个第一频率点逐步提升至目标频率，其中，所述运算芯片处于每个所述第一频率点时的持续时长短于常规搜频过程中所述运算芯片处于每个所述第一频率点时的持续时长；
    当所述数据处理设备中运算芯片达到目标频率后，按照所述存储模块所存储的所述至少一个输出电压值和所述至少一个第二频率点，并结合向所述运算芯片下发的测试任务和所述运算芯片的反馈结果，调整所述数据处理设备电源的输出电压和所述运算芯片的频率，以获得所述运算芯片的极限频率和与该极限频率对应的所述数据处理设备电源的输出电压。
  8. 一种数据处理设备，所述数据处理设备用于虚拟货币的数据处理，所述数据处理设备采用如权利要求7所述的数据处理设备快速搜频装置。
  9. 一种非易失性计算机可读存储介质，所述非易失性计算机可读存储介质存储指令，所述指令在由处理器执行时使得所述处理器执行如权利要求1至6中任一项所述的数据处理设备快速搜频方法中的步骤。
  10. 一种电子设备，包括：
    至少一个处理器；以及，
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器执行如权利要求1至6中任一项所述的数据处理设备快速搜频方法中的步骤。
PCT/CN2021/108054 2020-07-30 2021-07-23 数据处理设备快速搜频方法和装置以及数据处理设备 WO2022022403A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US17/606,939 US11493981B1 (en) 2020-07-30 2021-07-23 Rapid frequency searching method and apparatus for data processing device, and data processing device
CA3164095A CA3164095C (en) 2020-07-30 2021-07-23 Rapid frequency searching method and apparatus for data processing device, and data processing device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010748968.8A CN111966409B (zh) 2020-07-30 2020-07-30 矿机快速搜频方法和装置以及矿机
CN202010748968.8 2020-07-30

Publications (1)

Publication Number Publication Date
WO2022022403A1 true WO2022022403A1 (zh) 2022-02-03

Family

ID=73362561

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/108054 WO2022022403A1 (zh) 2020-07-30 2021-07-23 数据处理设备快速搜频方法和装置以及数据处理设备

Country Status (4)

Country Link
US (1) US11493981B1 (zh)
CN (1) CN111966409B (zh)
CA (1) CA3164095C (zh)
WO (1) WO2022022403A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115113675A (zh) * 2022-08-25 2022-09-27 深圳比特微电子科技有限公司 一种电源电压控制方法、装置、区块链服务器和存储介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111966409B (zh) 2020-07-30 2021-04-02 深圳比特微电子科技有限公司 矿机快速搜频方法和装置以及矿机
US11631138B2 (en) 2021-08-05 2023-04-18 Marc Fresa System, method and non-transitory computer-readable medium for cryptocurrency mining
CN115374164B (zh) * 2022-10-20 2023-04-28 深圳比特微电子科技有限公司 搜频调速方法、装置和数据处理设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100094572A1 (en) * 2008-10-15 2010-04-15 International Business Machines Corporation Dynamic Frequency And Voltage Scaling For A Computer Processor
US20130293269A1 (en) * 2012-04-26 2013-11-07 Huawei Technologies Co., Ltd. Method and apparatus for controlling chip performance
CN104597393A (zh) * 2015-01-29 2015-05-06 华为技术有限公司 一种芯片最高工作频率的确定方法及装置
CN109002356A (zh) * 2018-06-06 2018-12-14 北京嘉楠捷思信息技术有限公司 计算设备的芯片调频方法、装置、算力板、计算设备及存储介质
CN109240880A (zh) * 2018-08-27 2019-01-18 北京比特大陆科技有限公司 一种数字货币挖矿机的控制方法、装置及相关设备
CN111966409A (zh) * 2020-07-30 2020-11-20 深圳比特微电子科技有限公司 矿机快速搜频方法和装置以及矿机

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7941675B2 (en) * 2002-12-31 2011-05-10 Burr James B Adaptive power control
US7774627B2 (en) * 2002-10-03 2010-08-10 Via Technologies, Inc. Microprocessor capable of dynamically increasing its performance in response to varying operating temperature
CN1256762C (zh) * 2002-11-12 2006-05-17 联想(北京)有限公司 芯片频率调整的方法
US7689847B2 (en) * 2006-06-13 2010-03-30 Via Technologies, Inc. Method for increasing the data processing capability of a computer system
US7915910B2 (en) * 2009-01-28 2011-03-29 Apple Inc. Dynamic voltage and frequency management
CN102931655B (zh) * 2011-08-12 2014-12-10 珠海全志科技股份有限公司 一种动态调节电压和频率的电路控制系统和方法
CN102681450B (zh) * 2012-05-04 2014-04-02 华为技术有限公司 一种芯片控制方法及设备
US10007310B2 (en) * 2016-07-08 2018-06-26 Qualcomm Incorporated Circuits and methods providing calibration for temperature mitigation in a computing device
CN106407008B (zh) * 2016-08-31 2019-12-03 北京比特大陆科技有限公司 挖矿业务处理方法、装置和系统
CN108459934B (zh) * 2017-12-22 2021-01-29 深圳比特微电子科技有限公司 搜寻最佳频率的方法
CN108389044A (zh) * 2018-02-28 2018-08-10 北京比特大陆科技有限公司 工作量证明的计算方法和系统、电子设备、程序和介质
CN108734466A (zh) * 2018-05-16 2018-11-02 横琴密达科技有限责任公司 一种限制区块链矿场算力的方法及系统
CN108984469A (zh) * 2018-06-06 2018-12-11 北京嘉楠捷思信息技术有限公司 计算设备的芯片调频方法、装置、算力板、计算设备及存储介质
CN108693934A (zh) * 2018-06-28 2018-10-23 北京比特大陆科技有限公司 一种数字货币挖矿机以及数字货币挖矿系统
CN208298094U (zh) * 2018-06-28 2018-12-28 北京比特大陆科技有限公司 一种数字货币挖矿机以及数字货币挖矿系统
CN109032681A (zh) * 2018-08-27 2018-12-18 北京比特大陆科技有限公司 一种数字货币挖矿机的控制方法、装置及相关设备
CN109445688B (zh) * 2018-09-29 2022-04-15 上海百功半导体有限公司 一种存储控制方法、存储控制器、存储设备及存储系统
CN110687990A (zh) * 2019-10-08 2020-01-14 展讯半导体(成都)有限公司 温度校准装置及数字货币矿机

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100094572A1 (en) * 2008-10-15 2010-04-15 International Business Machines Corporation Dynamic Frequency And Voltage Scaling For A Computer Processor
US20130293269A1 (en) * 2012-04-26 2013-11-07 Huawei Technologies Co., Ltd. Method and apparatus for controlling chip performance
CN104597393A (zh) * 2015-01-29 2015-05-06 华为技术有限公司 一种芯片最高工作频率的确定方法及装置
CN109002356A (zh) * 2018-06-06 2018-12-14 北京嘉楠捷思信息技术有限公司 计算设备的芯片调频方法、装置、算力板、计算设备及存储介质
CN109240880A (zh) * 2018-08-27 2019-01-18 北京比特大陆科技有限公司 一种数字货币挖矿机的控制方法、装置及相关设备
CN111966409A (zh) * 2020-07-30 2020-11-20 深圳比特微电子科技有限公司 矿机快速搜频方法和装置以及矿机

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115113675A (zh) * 2022-08-25 2022-09-27 深圳比特微电子科技有限公司 一种电源电压控制方法、装置、区块链服务器和存储介质
CN115113675B (zh) * 2022-08-25 2022-11-18 深圳比特微电子科技有限公司 一种电源电压控制方法、装置、区块链服务器和存储介质

Also Published As

Publication number Publication date
CA3164095A1 (en) 2022-02-03
CA3164095C (en) 2024-02-27
US11493981B1 (en) 2022-11-08
CN111966409B (zh) 2021-04-02
US20220374063A1 (en) 2022-11-24
CN111966409A (zh) 2020-11-20

Similar Documents

Publication Publication Date Title
WO2022022403A1 (zh) 数据处理设备快速搜频方法和装置以及数据处理设备
US7392413B2 (en) Changing of operating voltage in semiconductor integrated circuit
US9645829B2 (en) Techniques to communicate with a controller for a non-volatile dual in-line memory module
DE112006003628B4 (de) System und Verfahren zum Betreiben von Komponenten einer integrierten Schaltung mit unabhängigen Frequenzen und/oder Spannungen
US8442697B2 (en) Method and apparatus for on-demand power management
JP2019159802A (ja) 記憶装置の電力制御方法および記憶装置
TW201339820A (zh) 使用串列介面之適應性電壓比例調整
US20140013140A1 (en) Information processing apparatus and computer program product
TWI634415B (zh) 感知熱策略方法和相應感知熱策略裝置
WO2022121767A1 (zh) 一种调压方法、装置、设备及介质
US10649519B2 (en) Computer system, method for conserving power, and computer
US11775048B1 (en) Safety control method and system for AI server
US9042196B2 (en) Memory system and method of controlling memory system
CN103150288B (zh) 一种快速开机的soc芯片及其实现方法
CN116893885B (zh) Java虚拟机调节方法、装置、电子设备及可读存储介质
CN104597983A (zh) 一种计算机风扇转速调节方法及主板系统
CN109375543B (zh) Dvs电压管理装置、系统及方法、存储介质、计算机设备
KR20230174749A (ko) 메모리의 동작 주파수 조정 방법, 스마트 단말기 및 저장 매체
WO2021259109A1 (zh) 补丁加载方法、网元及计算机可读存储介质
US11204593B2 (en) Control device and adjustment method
CN113986001A (zh) 芯片及控制方法
WO2021134216A1 (zh) 电源调节系统、方法、装置、芯片及电子设备
CN113791738A (zh) 基于固态硬盘的归一化低功耗处理方法、装置及计算机设备
EP3891594B1 (en) Memory control system with a sequence processing unit
JP2022144512A (ja) ストレージ装置、ストレージ装置制御方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21850896

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 3164095

Country of ref document: CA

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 27.06.2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21850896

Country of ref document: EP

Kind code of ref document: A1