WO2022105664A1

WO2022105664A1 - Computing system and operation method therefor, and electronic device and computer-readable medium

Info

Publication number: WO2022105664A1
Application number: PCT/CN2021/130002
Authority: WO
Inventors: 沈杨书; 何伟; 祝夭龙
Original assignee: 北京灵汐科技有限公司
Priority date: 2020-11-20
Filing date: 2021-11-11
Publication date: 2022-05-27

Abstract

A computing system and an operation method therefor, and an electronic device and a computer-readable medium, wherein the computing system comprises a computing unit. The operation method of the computing system comprises: acquiring the temperature of a computing unit (S101); and adjusting a working state of the computing unit according to the temperature of the computing unit (S102).

Description

Computing system, method for operating the same, electronic device, and computer-readable medium

technical field

The present disclosure relates to the field of computer technology, and in particular, to a computing system, a method for running the same, an electronic device, and a computer-readable medium.

Background technique

Computing systems (such as many-core systems, on-chip structures, etc.) are devices that can be used to perform various complete computing tasks. The computing system may include one or more relatively independent computing units (eg, processing cores), and each computing unit may perform all or part of the operations in the overall computing task.

However, the performance of the computing unit in the computing system is actually related to its temperature, and the influence of the temperature of the computing unit is not considered in the related art, so that the overall performance of the computing system is easily adversely affected because the temperature of the computing unit is not suitable.

SUMMARY OF THE INVENTION

The present disclosure provides a computing system, a method for operating the same, an electronic device, and a computer-readable medium.

In a first aspect, an embodiment of the present disclosure provides a method for operating a computing system, where the computing system includes a computing unit; the method includes: acquiring a temperature of the computing unit; and adjusting a temperature of the computing unit according to the temperature of the computing unit working status.

In some embodiments, the computing system includes a plurality of computing units, and the computing units are processing cores; the adjusting the working state of the computing units according to the temperature of the computing units includes: according to the temperature of the processing cores , to assign tasks to the processing cores.

In some embodiments, the obtaining the temperature of the computing unit includes: obtaining the temperature of the processing core in response to a new task waiting to be allocated.

In some embodiments, the acquiring the temperature of the computing unit includes: in response to the existence of an idle processing core, acquiring the temperature of the processing core.

In some embodiments, the acquiring the temperature of the computing unit includes: acquiring a plurality of temperatures of the processing cores; and assigning tasks to the processing cores according to the temperatures of the processing cores includes: The temperature of the processing cores, and task assignment is performed on at least one of the processing cores.

In some embodiments, assigning tasks to the processing cores according to the temperature of the processing cores includes at least one of the following: in response to the temperature of the first processing cores being lower than a preset first temperature threshold, assigning tasks to the first processing cores Allocating a computing-intensive task to the processing core; in response to the temperature of the second processing core being higher than or equal to a first temperature threshold and lower than or equal to a preset second temperature threshold, allocating a data-intensive task to the second processing core ; in response to the temperature of the third processing core being higher than the second temperature threshold, stop allocating tasks to the third processing core; wherein, the second temperature threshold is higher than the first temperature threshold, when the processing core is running computationally intensive tasks is greater than when running data-intensive tasks.

In some embodiments, after the stopping allocating tasks to the third processing core, the method further includes: moving at least part of the tasks processed by the third processing core out of the third processing core.

In some embodiments, the computing unit is an analog computing unit, and the computing system further includes a temperature sensor and a circuit unit corresponding to the analog computing unit, and the circuit unit is arranged within the first range of the corresponding analog computing unit. , the analog computing unit is set within the second range of its corresponding temperature sensor; the acquiring the temperature of the computing unit includes: acquiring the temperature collected by the temperature sensor, according to the temperature corresponding to the analog computing unit The temperature collected by the sensor determines the temperature of the analog computing unit; the adjusting the working state of the computing unit according to the temperature of the computing unit includes: adjusting the analog computing unit according to the comparison result between the temperature of the analog computing unit and a preset temperature range The operating state of the computing unit and its corresponding circuit unit.

In some embodiments, the adjusting the operating state of the analog computing unit and its corresponding circuit unit according to the comparison result between the temperature of the analog computing unit and the preset temperature range includes: responding to the temperature of the analog computing unit Within the preset temperature range, maintain the current operating state of the analog computing unit and its corresponding circuit unit; in response to the temperature of the analog computing unit being higher than the highest temperature in the preset temperature range, control the analog computing unit and its corresponding circuit The unit stops operating; in response to the temperature of the analog computing unit being lower than the lowest temperature in the preset temperature range, the operating state of the analog computing unit and its corresponding circuit unit is controlled according to whether the analog computing unit needs to start running.

In some embodiments, the controlling the operation state of the simulation computing unit and its corresponding circuit unit according to whether the simulation computing unit needs to start running includes: judging whether the simulation computing unit needs to start running; responding to the simulation computing The unit needs to start running, control the simulation computing unit to start running, and control its corresponding circuit unit to run; in response to the simulation computing unit not needing to start running, maintain the current running of the simulation computing unit and its corresponding circuit unit state.

In some embodiments, the number of the simulation computing units is multiple: the method further includes: determining the history of the simulation computing unit according to a plurality of temperatures collected by a temperature sensor corresponding to the simulation computing unit at multiple times temperature; according to the historical temperature of the analog computing unit, determine the recommended number of circuit units corresponding to the analog computing unit.

In some embodiments, the number of the analog computing units is multiple; at least some of the temperature sensors correspond to multiple analog computing units.

In some embodiments, the determining the temperature of the analog computing unit according to the temperature collected by the temperature sensor corresponding to the analog computing unit includes: in the case that the analog computing unit corresponds to only one temperature sensor, determining the temperature of the analog computing unit The temperature is the temperature collected by its corresponding temperature sensor; when the analog computing unit corresponds to multiple temperature sensors, the temperature of the analog computing unit is determined to be the average value of the multiple temperatures collected by the corresponding multiple temperature sensors .

In a second aspect, an embodiment of the present disclosure provides a method for laying out a computing system, wherein a simulation computing unit has been determined to be arranged at a first target position of the computing system; the method includes: determining and simulating according to the temperature of the simulation computing unit The circuit unit corresponding to the computing unit is located in the first range of the analog computing unit; it is determined that the temperature sensor corresponding to each analog computing unit is located in the second range of the analog computing unit.

In some embodiments, the number of the analog computing unit is multiple; the determining that the circuit unit corresponding to the analog computing unit is located within the first range of the analog computing unit according to the temperature of the analog computing unit includes: Running a preset computing system, the preset computing system includes an analog computing unit set at the first target position, and a temperature sensor corresponding to the analog computing unit; acquiring multiple temperatures collected by the temperature sensor at multiple times, Determine the historical temperature of the analog computing unit according to the multiple temperatures collected by the temperature sensor corresponding to the analog computing unit at multiple times; determine the recommended number of circuit units corresponding to the analog computing unit according to the historical temperature of the analog computing unit; determine The number of circuit units corresponding to the analog computing unit is a recommended number, and is located within the first range of the analog computing unit.

In a third aspect, an embodiment of the present disclosure provides a computing system, including: a computing unit and a controller; the controller is configured to execute any one of the operating methods of the computing system of the embodiments of the present disclosure.

In some embodiments, the computing unit is an analog computing unit; the computing system further includes: a temperature sensor and a circuit unit corresponding to the analog computing unit.

In a fourth aspect, embodiments of the present disclosure provide an electronic device, including: one or more processors; one or more memories, on which one or more programs are stored, when the one or more programs are One or more processors execute, so that the one or more processors can implement any one of the operating methods of the computing system of the embodiments of the present disclosure, or implement any one of the layout methods of the computing systems of the embodiments of the present disclosure.

In a fifth aspect, an embodiment of the present disclosure provides a computer-readable medium on which a computer program is stored, and when the computer program is executed by a processor, implements any method for running a computing system in the embodiment of the present disclosure, or implements A layout method of any computing system according to the embodiments of the present disclosure.

In the embodiment of the present disclosure, the temperature of the computing unit can be obtained when the computing system is running, and its working state can be adjusted according to the temperature of the computing unit, thereby ensuring that the temperature of the computing unit is always in a suitable range and improving the performance of the computing system.

It should be understood that what is described in this section is not intended to identify key or critical features of embodiments of the disclosure, nor is it intended to limit the scope of the disclosure. Other features of the present disclosure will become readily understood from the following description.

Description of drawings

The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification, and together with the detailed embodiments, they are used to explain the present disclosure and do not constitute a limitation on the present disclosure. The above and other features and advantages will become more apparent to those skilled in the art by describing detailed embodiments with reference to the accompanying drawings, in which:

FIG. 1 is a flowchart of a method for running a computing system according to an embodiment of the present disclosure;

FIG. 2 is a flowchart of another operating method of a computing system provided by an embodiment of the present disclosure;

FIG. 3 is a flowchart of another operating method of a computing system provided by an embodiment of the present disclosure;

4 is a block diagram of a computing system (many-core system) provided by an embodiment of the present disclosure;

FIG. 5 is a flowchart of another operating method of a computing system provided by an embodiment of the present disclosure;

FIG. 6 is a flowchart of another operating method of a computing system provided by an embodiment of the present disclosure;

Fig. 7 is the volt-ampere characteristic curve of the diode at different temperatures in the related art;

FIG. 8 is a schematic structural diagram of another computing system (on-chip structure) provided by an embodiment of the present disclosure;

FIG. 9 is a schematic structural diagram of another computing system (on-chip structure) provided by an embodiment of the present disclosure;

10 is a schematic structural diagram of another computing system (on-chip structure) provided by an embodiment of the present disclosure;

FIG. 11 is a schematic structural diagram of another computing system (on-chip structure) provided by an embodiment of the present disclosure;

12 is a schematic structural diagram of another computing system (on-chip structure) provided by an embodiment of the present disclosure;

FIG. 13 is a schematic structural diagram of another computing system (on-chip structure) provided by an embodiment of the present disclosure;

14 is a flowchart of a layout method of a computing system provided by an embodiment of the present disclosure;

FIG. 15 is a block diagram of the composition of a computing system provided by an embodiment of the present disclosure;

FIG. 16 is a block diagram of the composition of an electronic device according to an embodiment of the present disclosure;

FIG. 17 is a block diagram of the composition of a computer-readable medium provided by an embodiment of the present disclosure.

Detailed ways

In order for those skilled in the art to better understand the technical solutions of the present disclosure, the computing system, its operating method, electronic device, and computer-readable medium provided by the present disclosure will be described in detail below with reference to the accompanying drawings.

The present disclosure will be described more fully hereinafter with reference to the accompanying drawings, but the illustrated embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth in this disclosure. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Various embodiments of the present disclosure and various features of the embodiments may be combined with each other without conflict.

In the present disclosure, the terms "first" and "second" are for descriptive purposes only, and should not be construed as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature delimited with "first", "second" may expressly or implicitly include one or more of that feature. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

The terminology used in this disclosure is used to describe particular embodiments only, and is not intended to limit the disclosure. As used in this disclosure, the term "and/or" includes any and all combinations of one or more of the associated listed items. As used in this disclosure, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. The terms "comprising", "made of", as used in this disclosure, specify the presence of stated features, integers, steps, operations, elements and/or components, but do not preclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and/or groups thereof.

Unless otherwise defined, all terms (including technical and scientific terms) used in this disclosure have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in common dictionaries should be construed as having meanings consistent with their meanings in the context of the related art and the present disclosure, and will not be construed as having idealized or over-formal meanings, Unless this disclosure expressly so limited.

The present disclosure is not limited to the embodiments shown in the drawings, but includes modifications of configurations formed based on manufacturing processes. Thus, the regions illustrated in the figures have schematic properties and the shapes of regions illustrated in the figures are illustrative of the specific shapes of regions of elements and are not intended to be limiting.

In a first aspect, referring to FIG. 1 to FIG. 13 , an embodiment of the present disclosure provides a method for operating a computing system, where the computing system includes a computing unit.

The embodiments of the present disclosure are used to run a computing system, which is a system with a certain data processing capability, which includes one or more computing units; and each computing unit is a relatively independent structure and can perform certain operations.

Referring to FIG. 1 , the operation method of the computing system according to the embodiment of the present disclosure includes:

S101. Acquire the temperature of the computing unit.

S102. Adjust the working state of the computing unit according to the temperature of the computing unit.

In the embodiment of the present disclosure, when the computing system is running (at least part of the computing units is running), the temperature of at least some of the computing units is acquired, and the working state of the corresponding computing unit is adjusted according to the temperature of the computing unit.

The temperature of the computing unit refers to a parameter that can characterize the temperature characteristics of the computing unit, which may also be a temperature value at a certain position on the computing unit, or a temperature value at a specific position away from the computing unit, which will not be described in detail here.

Among them, the temperature of the computing unit can be acquired through a temperature-sensing device located inside or outside the computing unit (such as a temperature sensing device built into the computing unit, or a subsequent temperature sensor, or an infrared thermometer, etc.), or by acquiring and parsing a temperature database Or the temperature data corresponds to the acquisition of the temperature data in the cache, which will not be described in detail here.

In some embodiments, the computing system includes a plurality of computing units, the computing units being processing cores.

In some embodiments, referring to FIG. 2 , obtaining the temperature of the computing unit ( S101 ) includes:

S201. Obtain the temperature of the processing core.

And according to the temperature of the computing unit, the working state of the computing unit is adjusted (S102), including:

S202: Assign tasks to the processing cores according to the temperature of the processing cores.

As one way of implementing embodiments of the present disclosure, referring to FIG. 4 , each computing unit may be a "processing core". A processing core, also known as a core or a core, is the smallest processing unit that can be independently scheduled and has complete computing capabilities, and may specifically be a chip (IC) or a core (Core) within a chip.

Therefore, the computing system of the embodiment of the present disclosure is a “many-core system” including multiple processing cores as a whole. The multiple processing cores of the many-core system are connected through the on-chip network to form a certain topology structure, wherein each processing core can process a certain number of processing cores. Through the joint work of multiple processing cores, the overall program can be quickly run, and the ability of multi-task parallel processing is provided.

Generally speaking, the more tasks processed in the processing core, the higher the heat generation and the higher the temperature of the processing core (the temperature of the computing unit). Thus, the temperature of the processing core can be used to characterize its utilization and load to a certain extent. For example, when the temperature of the processing core is within the predetermined range, it means that the processing core is in the working range of normal operation, and when the temperature of the processing core is outside the predetermined range, it means that the processing core is overloaded or is idle.

In some related technologies, the influence of the temperature of the task processing cores is not considered, so that the temperature distribution of different processing cores is not uniform, the performance of each processing core affected by temperature is greatly different, and the amount of tasks processed in each processing core is not equal. Evenly, the processing progress of each processing core is uneven, which affects the overall performance of the many-core system.

However, in the embodiment of the present disclosure, tasks are allocated according to the temperature of each processing core, so as to ensure that the temperature of each processing core is uniform, the amount of tasks processed is relatively uniform, and the processing progress of each processing core is similar, which improves the many-core system. overall performance.

In some embodiments, acquiring the temperature of the computing unit (S101) includes:

S2011. Acquire the temperature of the processing core in response to the existence of a new task waiting to be allocated.

As a way of the embodiment of the present disclosure, when there is a new task to be assigned, the temperature of the processing core may be acquired to determine which processing core should be assigned the task.

S2012. Acquire the temperature of the processing core in response to the existence of an idle processing core.

When there is at least one idle processing core in the many-core system, it indicates that the idle processing core can receive tasks (such as new tasks, or tasks migrated from other processing cores), so as another way of the embodiment of the present disclosure, It is also possible to obtain the temperature of a processing core (not necessarily the idle processing core, but also other processing cores) when there is at least one idle processing core to determine whether to assign tasks to the idle processing core, or to determine which one should be The tasks of the processing cores are migrated to the idle processing cores.

S2013. Acquire the temperatures of the multiple processing cores.

According to the temperature of the processing core, tasks are allocated to the processing core (S202), including:

S2021. Perform task assignment on at least one of the processing cores according to the temperature of the multiple processing cores.

As a method of the embodiment of the present disclosure, the temperature of multiple processing cores may be acquired at one time, and based on this, it is determined to perform task assignment on at least some of the processing cores. For example, when there are new tasks waiting to be assigned, the temperature of multiple processing cores can be obtained, so as to decide which processing core to assign the task to.

Of course, there are various specific ways to obtain the temperature of the processing core. For example, when there is a new task waiting to be allocated, the temperature of a processing core can also be obtained first, and if the temperature indicates that a new task can be allocated to the processing core, the Allocation, if not, then obtain the temperature of another processing core.

In some embodiments, referring to FIG. 3 , according to the temperature of the processing core, task assignment is performed on the processing core (S202), including at least one of the following:

S20221. In response to the temperature of the first processing core being lower than a preset first temperature threshold, assign a computation-intensive task to the first processing core.

S20222. In response to the temperature of the second processing core being higher than or equal to the first temperature threshold and lower than or equal to a preset second temperature threshold, assign a data-intensive task to the second processing core.

S20223. In response to the temperature of the third processing core being higher than the second temperature threshold, stop allocating tasks to the third processing core.

Wherein, the second temperature threshold is higher than the first temperature threshold, and the calorific value of the processing core when running computation-intensive tasks is greater than the calorific value when running data-intensive tasks.

As a manner of the embodiment of the present disclosure, a predetermined interval from a first temperature threshold to a second temperature threshold may be set, and tasks may be divided into computationally intensive tasks and data-intensive tasks.

Among them, computing-intensive tasks have the characteristics of large amount of calculation, which will make the processing cores performing the task generate more heat, such as tasks related to convolution operations, etc.; while data-intensive tasks have the characteristics of large amount of data and small amount of computation. , which will make the processing cores that perform tasks generate less heat, such as fully connected tasks.

Alternatively, the first temperature threshold and the second temperature threshold may also have a certain mapping relationship with computation-intensive tasks and data-intensive tasks.

Therefore, according to the relationship between the temperature of the processing core and the predetermined interval, the processing core can be assigned tasks in different ways:

(1) For the first processing core whose temperature is lower than the first temperature threshold, since it has a large heating space, it can undertake a task with a large amount of heat (computation-intensive task).

(2) For the second processing core whose temperature is between the first temperature threshold and the second temperature threshold, its temperature is still in a suitable working range, but there is not much room for heating, so it is not suitable to undertake the task of generating a large amount of heat , but it can still be assigned tasks that generate less heat (data-intensive tasks).

(3) For the third processing core whose temperature is higher than the second temperature threshold, it has been overheated, and in order to prevent its temperature from continuing to rise, it should at least wait for it to cool down before assigning tasks.

Among them, the above "overheating" is only used to describe that the temperature has exceeded the appropriate working range, and it does not mean that the temperature has exceeded the upper limit of the temperature limit and failure occurs.

The above first processing core, second processing core, and third processing core are only used to represent that the processing cores are at different temperatures, and do not represent specific processing cores.

In some embodiments, referring to FIG. 3 , after the assignment of tasks to the third processing core is stopped (S20223), the method further includes:

S202231. Move at least part of the tasks processed by the third processing core out of the third processing core.

As a manner of the embodiment of the present disclosure, after the assignment of tasks to the overheated third processing core is stopped, at least some of the existing tasks may be further migrated out.

Among them, how many tasks are transferred out from the third processing core and which tasks are transferred out can be determined according to its temperature. For example, from the viewpoint of effectively lowering the temperature of the third processing core, computationally intensive tasks can be "prioritized" out of it.

Among them, the tasks transferred out of the third processing core may be temporarily "suspended" to stop processing, or may be allocated to other processing cores (ie, migrated to other processing cores) for processing.

Among them, the above task migration will occupy a certain amount of routing bandwidth, so the task can also be reassigned through the task distribution module in the system.

The processing core that receives the tasks migrated from the third processing core can be selected according to the above method, that is, the migrated computing-intensive tasks can be allocated to the first processing core whose temperature is lower than the first temperature threshold, and the The migrated data-intensive tasks are distributed to the second processing core whose temperature is between the first temperature threshold and the second temperature threshold, while other third processing cores whose temperature is higher than the second temperature threshold should no longer receive the migrated tasks.

The specific manner of assigning tasks according to temperature is not limited to this. For example, although the first processing core receives computing-intensive tasks preferentially, it can also receive data-intensive tasks; for another example, when the temperature of each processing core exceeds the first temperature threshold, if there are computing-intensive tasks that must be allocated , the computationally intensive task can still be assigned to one of the processing cores (of course, the relatively low temperature).

In some embodiments, the computing unit is an analog computing unit, and the computing system further includes a temperature sensor and a circuit unit corresponding to the analog computing unit, the circuit unit is arranged in the first range of the corresponding analog computing unit, and the analog computing unit is arranged in the within the second range of its corresponding temperature sensor.

In some embodiments, referring to FIG. 5 , acquiring the temperature of the computing unit ( S101 ) includes:

S301. Acquire the temperature collected by the temperature sensor, and determine the temperature of the analog computing unit according to the temperature collected by the temperature sensor corresponding to the analog computing unit.

According to the temperature of the computing unit, adjust the working state of the computing unit (S102), including;

S302. Adjust the operating state of the analog computing unit and its corresponding circuit unit according to the comparison result between the temperature of the analog computing unit and the preset temperature range.

Among them, the analog device refers to the device in which at least part of the processed signal is an analog signal (non-digital signal), which can specifically be a resistor, a capacitor, an inductor, a diode, a transistor, an analog amplifier, a D/A conversion circuit, an A/D Conversion circuits, analog signal conditioners, integrated voltage regulator circuits, sensors, audio and video circuits, etc.

Correspondingly, a computing unit including at least one analog device is called an analog computing unit. For example, the analog computing unit may be the above processing core or chip, or may be a part of circuit modules in the processing core or chip.

The research found that the temperature has a great influence on the characteristics of the simulated device. For example, typically, as the temperature increases, the kinetic energy of the electrons increases, which results in changes in the characteristics of the analog device. For example, Fig. 7 is a volt-ampere (IV) characteristic curve of a diode at different temperatures. It can be seen that when the temperature is different, the volt-ampere characteristic of the diode (analog device) changes significantly.

Since the analog device has the above-mentioned temperature characteristics, for a single analog device, if the temperature difference at different times is large, the operating results at different times will be greatly different; for multiple analog devices, if it is in the The temperature difference of the analog devices at different positions is large, which will lead to deviations in the operation result calculation of the analog devices at different positions. In a word, the above differences will cause the operation results generated by an analog device at one time to be unusable at another time, or the operation results generated by a certain analog device cannot be used by other analog devices, resulting in unreliable data finally generated by the analog device. .

Therefore, in the related art, for an analog computing unit including an analog device, the operation result is also affected by the temperature change, which may cause the operation result to be unreliable.

As another way of the embodiment of the present disclosure, when the computing unit is an analog computing unit, referring to FIG. 8 to FIG. 13 , a corresponding temperature sensor and a circuit unit are set for the analog computing unit, and the analog computing unit is set with a temperature corresponding to The distance between the sensor and the circuit unit does not exceed the preset range (first range, second range).

The number of analog computing units may be one referring to FIG. 8, or multiple referring to FIG. 9 to FIG. 13, but regardless of the number of analog computing units, each analog computing unit must correspond to one or more circuits unit, and one or more temperature sensors; and regardless of the above correspondence, the circuit unit and the temperature sensor must be located within the corresponding range of the corresponding analog computing unit.

For example, referring to FIG. 8 , the computing system (on-chip structure) may include 1 analog computing unit, and the analog computing unit corresponds to 1 temperature sensor; for another example, referring to FIG. 9 , the computing system may include 8 analog computing units, and each analog computing unit may include 8 analog computing units. Each temperature sensor corresponds to 1 analog computing unit; for another example, referring to FIG. 10 , the computing system may include 8 analog computing units, and every 2 analog computing units corresponds to the same temperature sensor; for another example, referring to FIG. 11 , each The temperature sensor can correspond to the four surrounding analog computing units, so that the two analog computing units on the left in the virtual frame correspond to two temperature sensors (the first temperature sensor and the second temperature sensor), and the 2 analog computing units on the right in the virtual frame correspond to two temperature sensors (the first temperature sensor and the second temperature sensor). Each analog computing unit also corresponds to 2 temperature sensors (the second temperature sensor and the third temperature sensor); for another example, the number of circuit units corresponding to different analog computing units may be the same with reference to FIG. 9 to FIG. 11 , or refer to 12 and 13, the number of circuit units corresponding to different analog computing units is different.

Wherein, the structure of the computing system in each drawing is just an example, and in other embodiments, the structure of the computing system can be determined according to the actual situation.

Therefore, the computing system as a whole may be an "on-chip structure", that is, an analog computing unit, a circuit unit, and an analog computing unit provided on a chip or a circuit board.

Among them, in the computing system (on-chip structure), the analog computing unit is the above unit for performing operations, and the circuit unit is a circuit structure arranged in the computing system for implementing the "non-operation" function, such as for analog Computing unit power supply, for processing analog computing unit signals, for transmitting analog computing unit signals, etc. For example, the circuit unit may be a functional circuit unit with the above practical functions, or may be a flip circuit unit that can only generate heat (only used for temperature control).

In the embodiment of the present disclosure, the temperature of the corresponding analog computing unit is determined according to the temperature detected by the temperature sensor, and the operating states of the corresponding analog computing unit and the circuit unit are controlled according to the comparison result between the temperature of the analog computing unit and the preset temperature range , in order to realize the temperature control of the computing system (on-chip structure), which can prevent the temperature difference of the analog computing unit from being too large (including the temperature difference of the analog computing unit being too large at different times, and the temperature difference between different analog computing units being too large), Therefore, the accuracy of the operation result of the analog computing unit is improved, and the problems in the related art that due to the large temperature difference of the analog device, the error of the operation result is large or the operation result is unreliable is solved.

The computing system of the embodiment of the present disclosure may further include a controller, and the above steps S301 and S302 are performed by the controller, that is, the temperature of the computing system may be controlled by the controller.

In some embodiments, referring to FIG. 6 , according to the comparison result between the temperature of the analog computing unit and the preset temperature range, adjusting the operating state of the analog computing unit and its corresponding circuit unit (S302), including:

S3021. In response to the temperature of the analog computing unit being within a preset temperature range, maintain the current operating state of the analog computing unit and its corresponding circuit unit.

S3022 , in response to the temperature of the analog computing unit being higher than the highest temperature in the preset temperature range, control the analog computing unit and its corresponding circuit unit to stop running.

S3023 . In response to the temperature of the analog computing unit being lower than the lowest temperature in the preset temperature range, control the operating state of the analog computing unit and its corresponding circuit unit according to whether the analog computing unit needs to start running.

Obviously, the operation of various units will generate heat and cause the temperature to rise. Therefore, as a method of the embodiment of the present disclosure, when the temperature of the analog computing unit is too high (exceeding the maximum temperature of the preset temperature range), the The analog computing unit and the corresponding circuit unit stop running to prevent the temperature from rising further; and when the temperature of the analog computing unit is appropriate (within the preset temperature range), the current operating state of the analog computing unit and the corresponding circuit unit is maintained, To maintain the temperature of the analog computing unit; when the temperature of the analog computing unit is too low (lower than the minimum temperature of the preset temperature range), try to start the analog computing unit and its corresponding circuit unit to increase the temperature of the analog computing unit .

The specific value of the above preset temperature range can be adjusted according to the actual situation. For example, the preset temperature range may be set according to the temperature corresponding to the best working state of the simulation computing unit, or the preset temperature range may be set by the staff based on experience, which will not be described in detail here.

Among them, the control of the analog computing unit and the circuit unit should not affect the overall operation of the computing system. For example, when the analog computing unit and circuit unit stop running, the work performed by it can be processed by other analog computing units and circuit units (such as task migration); for another example, when the analog computing unit and circuit unit are started, they should be assigned corresponding (such as assigning tasks); another example, when the circuit unit has no actual work that can be performed, it can "flip" to run, that is, it only generates heat and does not process the actual content.

Therefore, through the above method, the temperature difference of the analog computing unit can be reduced, and the uniformity of its temperature and performance can be improved.

In some embodiments, according to whether the analog computing unit needs to start running, controlling the running state of the analog computing unit and its corresponding circuit unit (S30213) includes:

S30231. Determine whether the simulation computing unit needs to start running.

S30232. In response to the need to start the operation of the simulation computing unit, control the simulation computing unit to start the operation, and control the corresponding circuit unit to operate.

S30233. In response to the simulation computing unit not needing to start running, maintain the current operating state of the simulation computing unit and its corresponding circuit unit.

As a mode of the embodiment of the present disclosure, when the temperature of the simulation computing unit is lower than the lowest temperature of the preset temperature range, it may be determined whether the simulation computing unit needs to start running (for example, whether there is a new task to be allocated), and then based on the Determines whether the simulation computing unit should start running.

In some embodiments, the number of analog computing units is multiple. The method of the embodiment of the present disclosure further includes:

S304: Determine the historical temperature of the analog computing unit according to multiple temperatures collected by the temperature sensor corresponding to the analog computing unit at multiple times.

S305. Determine the recommended number of circuit units corresponding to the analog computing unit according to the historical temperature of the analog computing unit.

As an embodiment of the present disclosure, when there are multiple analog computing units, the total number of corresponding circuit units is constant, but these circuit units may be distributed near different analog computing units, that is, corresponding to different Analog computing unit.

Therefore, a temperature sensor can be used to collect multiple temperatures at different times, so as to determine the historical temperature of the corresponding analog computing unit (that is, the long-term temperature distribution law of the analog computing unit) according to these temperatures. Therefore, the recommended number of circuit units corresponding to the analog computing unit can be determined according to the historical temperature (that is, when the analog computing unit corresponds to the recommended number of circuit units, the temperature can be within a relatively reasonable range with a high probability), and used as Determines the basis for the layout of the layout computing system (structure-on-chip).

In some embodiments, the number of analog computing units is multiple; at least some of the temperature sensors correspond to multiple analog computing units.

As an embodiment of the present disclosure, when there are multiple analog computing units, referring to FIG. 10 to FIG. 13 , there may be temperature sensors corresponding to multiple analog computing units, that is, the temperature measured by the temperature sensor may be used to determine multiple analog computing units Calculate the temperature of the cell.

In some embodiments, determining the temperature of the analog computing unit according to the temperature collected by the temperature sensor corresponding to the analog computing unit (S302) includes:

S3024. In the case that the analog computing unit corresponds to only one temperature sensor, determine the temperature of the analog computing unit as the temperature collected by the corresponding temperature sensor.

S3025 , in the case that the analog computing unit corresponds to multiple temperature sensors, determine that the temperature of the analog computing unit is an average value of multiple temperatures collected by the corresponding multiple temperature sensors.

As a way of the embodiment of the present disclosure, referring to FIGS. 8 , 9 , 10 , and outside the dashed-line frame in FIG. 11 , FIGS. 12 , and 13 , if the analog computing unit corresponds to only one temperature sensor, its temperature is the temperature of the temperature sensor. However, referring to the dotted box in FIG. 11 , when the analog computing unit only corresponds to multiple temperature sensors, its temperature is the average value of the temperatures collected by multiple temperature sensors.

It should be understood that the number of temperature sensors corresponding to each analog computing unit is not necessarily related to the number of analog computing units corresponding to each temperature sensor.

In the second aspect, referring to FIG. 14 , an embodiment of the present disclosure provides a layout method of a computing system, and the simulation computing unit has been determined to be arranged on a first target position of the computing system.

The layout method of the computing system according to the embodiment of the present disclosure includes:

S401. According to the temperature of the analog computing unit, determine that the circuit unit corresponding to the analog computing unit is located within the first range of the analog computing unit.

S402. Determine that the temperature sensor corresponding to each analog computing unit is located within the second range of the analog computing unit.

When laying out (or "designing") the above computing system (on-chip structure), usually the number of analog computing units and their positions in the computing system have been predetermined, so they can be determined according to the location and temperature of each analog computing unit How to set the corresponding circuit unit in its first range, and how to set the corresponding temperature sensor in its second range.

In the embodiment of the present disclosure, the corresponding circuit unit and temperature sensor are set according to the temperature of the analog computing unit, so that a reasonable layout of the computing system can be realized, and it is ensured that the analog computing unit in the computing system according to the layout is in a suitable state for most of the time during operation. temperature range, improving the performance of the computing system.

In some embodiments, the number of analog computing units is multiple; according to the temperature of the analog computing unit, it is determined that the circuit unit corresponding to the analog computing unit is located within the first range of the analog computing unit (S401), including:

S4011. Run a preset computing system, where the preset computing system includes an analog computing unit disposed at the first target position, and a temperature sensor corresponding to the analog computing unit.

S4012: Acquire multiple temperatures collected by the temperature sensor at multiple times, and determine the historical temperature of the analog computing unit according to the multiple temperatures collected by the temperature sensor corresponding to the analog computing unit at multiple times.

S4013. Determine the recommended number of circuit units corresponding to the analog computing unit according to the historical temperature of the analog computing unit.

S4014. Determine that the number of circuit units corresponding to the analog computing unit is the recommended number and is located within the first range of the analog computing unit.

As a mode of the embodiment of the present disclosure, when there are multiple simulation computing units, a "preset computing system" with the same layout of the simulation computing units may be pre-manufactured, and the preset computing system may be run to measure each simulation computing unit. Based on the historical temperature of the unit, the recommended number of circuit units corresponding to each analog computing unit is determined, and the corresponding recommended number of circuit units are actually laid out.

The temperature of the simulation computing unit in the embodiment of the present disclosure is not limited to the above historical temperature, for example, it may also be a "predicted temperature" obtained through model simulation operation or experience.

In a third aspect, referring to FIG. 15 , an embodiment of the present disclosure provides a computing system 40 , including: a computing unit and a controller 48 .

The controller 48 is configured to execute any one of the operating methods of the computing system 40 in the embodiments of the present disclosure.

The computing system 40 of the embodiment of the present disclosure may include the above computing unit, and a controller 48 for controlling the computing unit to work according to the above operating method of the computing system.

In some embodiments, referring to FIGS. 8 to 13 , the computing unit is an analog computing unit 42 ; the computing system further includes: a temperature sensor 46 and a circuit unit 44 corresponding to the analog computing unit 42 .

When the computing system is the above on-chip structure, the computing unit is the above analog computing unit 42 , and the computing system also includes the above temperature sensor 46 and the circuit unit 44 .

In a fourth aspect, referring to FIG. 16 , an embodiment of the present disclosure provides an electronic device 50, including:

one or more processors 51;

One or more memories 52 having one or more programs stored thereon, when the one or more programs are executed by the one or more processors 51, enable the one or more processors 51 to implement any one of the embodiments of the present disclosure A method for running a computing system, or a method for implementing any layout of a computing system in the embodiments of the present disclosure.

In some embodiments, the electronic device 50 may further include one or more I/O interfaces (represented by double-sided arrows in the figure), connected between the processor 51 and the memory 52, and configured to implement the communication between the processor 51 and the memory 52. Information exchange.

Wherein, the processor is a device with data processing capability, which includes but is not limited to a central processing unit (CPU), etc.; the memory is a device with data storage capability, which includes but is not limited to random access memory (RAM, more specifically such as SDRAM) , DDR, etc.), read-only memory (ROM), electrified erasable programmable read-only memory (EEPROM), flash memory (FLASH); I/O interface (read and write interface) is connected between the processor and the memory, which can realize the memory and the memory. The information exchange of the processor, which includes but is not limited to the data bus (Bus) and the like.

It should be noted that the electronic devices in the embodiments of the present disclosure include mobile electronic devices and non-mobile electronic devices.

In a fifth aspect, referring to FIG. 17 , an embodiment of the present disclosure provides a computer-readable medium 60 on which a computer program is stored, and when the computer program is executed by a processor, implements any method of running a computing system in the embodiment of the present disclosure , or implement any of the layout methods of the computing system in the embodiments of the present disclosure.

The processor may be the processor in the electronic device in the above embodiment. The computer-readable medium of the embodiment of the present disclosure includes a computer read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and the like.

Those of ordinary skill in the art can understand that all or some of the steps, systems, and functional modules/units in the apparatus disclosed above can be implemented as software, firmware, hardware, and appropriate combinations thereof.

In a hardware implementation, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be composed of several physical components Components execute cooperatively.

Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit (CPU), digital signal processor or microprocessor, or as hardware, or as an integrated circuit such as Application-specific integrated circuits. Such software may be distributed on computer-readable media, which may include computer-readable media (or non-transitory media) and communication media (or transitory media). As is known to those of ordinary skill in the art, the term computer-readable medium includes both volatile and non-transitory media implemented in any method or technology for storage of information, such as computer-readable instructions, data structures, program modules or other data. volatile, removable and non-removable media. Computer readable media include, but are not limited to, random access memory (RAM, more specifically SDRAM, DDR, etc.), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory (FLASH), or other magnetic disks memory; compact disk read only (CD-ROM), digital versatile disk (DVD), or other optical disk storage; magnetic cartridge, tape, magnetic disk storage, or other magnetic storage; any other storage that can be used to store desired information and that can be accessed by a computer medium. In addition, communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and can include any information delivery media, as is well known to those of ordinary skill in the art .

This disclosure has disclosed example embodiments, and although specific terms are employed, they are used and should only be construed in a general descriptive sense and not for purposes of limitation. In some instances, it will be apparent to those skilled in the art that features, characteristics and/or elements described in connection with a particular embodiment may be used alone or in combination with other embodiments, unless expressly stated otherwise. Features and/or elements are used in combination. Accordingly, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the scope of the present disclosure as set forth in the appended claims.

Claims

A method for operating a computing system, wherein the computing system includes a computing unit; the method includes:

obtain the temperature of the computing unit;

According to the temperature of the computing unit, the working state of the computing unit is adjusted.
The method according to claim 1, wherein the computing system includes a plurality of computing units, and the computing units are processing cores; the adjusting the working state of the computing units according to the temperature of the computing units includes:

The processing cores are assigned tasks according to the temperature of the processing cores.
The method according to claim 2, wherein the acquiring the temperature of the computing unit comprises:

The temperature of the processing core is obtained in response to a new task awaiting assignment.
The method according to claim 2, wherein the acquiring the temperature of the computing unit comprises:

In response to the existence of an idle processing core, the temperature of the processing core is obtained.
The method of claim 2, wherein:

The acquiring the temperature of the computing unit includes: acquiring the temperature of a plurality of the processing cores;

The performing task assignment on the processing cores according to the temperature of the processing cores includes: performing task assignment on at least one of the processing cores according to the temperatures of the plurality of processing cores.
The method according to claim 2, wherein the task assignment to the processing cores according to the temperature of the processing cores comprises at least one of the following:

In response to the temperature of the first processing core being lower than the preset first temperature threshold, assigning a computationally intensive task to the first processing core;

assigning a data-intensive task to the second processing core in response to the temperature of the second processing core being higher than or equal to the first temperature threshold and lower than or equal to a second preset temperature threshold;

in response to the temperature of the third processing core being above the second temperature threshold, ceasing the assignment of tasks to the third processing core;

Wherein, the second temperature threshold is higher than the first temperature threshold, and the calorific value of the processing core when running computation-intensive tasks is greater than the calorific value when running data-intensive tasks.
The method according to claim 6, wherein after the stopping of allocating tasks to the third processing core, the method further comprises:

At least some of the tasks processed by the third processing core are migrated out of the third processing core.
The method according to claim 1, wherein the computing unit is an analog computing unit, the computing system further comprises a temperature sensor and a circuit unit corresponding to the analog computing unit, and the circuit unit is set in the corresponding analog computing unit. Within the first range of the computing unit, the analog computing unit is set within the second range of its corresponding temperature sensor;

The acquiring the temperature of the computing unit includes: acquiring the temperature collected by the temperature sensor, and determining the temperature of the analog computing unit according to the temperature collected by the temperature sensor corresponding to the analog computing unit;

The adjusting the operating state of the computing unit according to the temperature of the computing unit includes: adjusting the operating state of the analog computing unit and its corresponding circuit unit according to a comparison result between the temperature of the analog computing unit and a preset temperature range.
method according to claim 8, is characterized in that, described according to the comparison result of the temperature of described simulation calculation unit and preset temperature range, adjust the operation state of simulation calculation unit and its corresponding circuit unit, including:

In response to the temperature of the analog computing unit being within a preset temperature range, maintaining the current operating state of the analog computing unit and its corresponding circuit unit;

In response to the temperature of the analog computing unit being higher than the maximum temperature of the preset temperature range, controlling the analog computing unit and its corresponding circuit unit to stop running;

In response to the temperature of the analog computing unit being lower than the lowest temperature of the preset temperature range, the operation state of the analog computing unit and its corresponding circuit unit is controlled according to whether the analog computing unit needs to start running.
The method according to claim 9, wherein the controlling the operation state of the analog computing unit and its corresponding circuit unit according to whether the analog computing unit needs to start operation, comprising:

judging whether the simulation computing unit needs to start running;

In response to the simulation computing unit needing to start running, controlling the simulation computing unit to start running, and controlling its corresponding circuit unit to operate;

In response to the simulation computing unit not needing to start running, the current operating state of the simulation computing unit and its corresponding circuit unit is maintained.
The method according to claim 8, wherein the number of the simulation computing units is multiple: the method further comprises:

Determine the historical temperature of the analog computing unit according to the multiple temperatures collected by the temperature sensor corresponding to the analog computing unit at multiple times;

According to the historical temperature of the analog computing unit, the recommended number of circuit units corresponding to the analog computing unit is determined.
A computing system, comprising: a computing unit and a controller;

The controller is adapted to perform the method of any one of claims 1 to 11.
The computing system according to claim 12, wherein the computing unit is an analog computing unit; the computing system further comprises:

A temperature sensor and a circuit unit corresponding to the analog computing unit.
An electronic device, comprising:

one or more processors;

one or more memories having stored thereon one or more programs which, when executed by the one or more processors, enable the one or more processors to implement the method as claimed in claim 1 to any of 11.
A computer-readable medium on which a computer program is stored, characterized in that, when the computer program is executed by a processor, any one of claims 1 to 11 is implemented.