KR102471867B1 - Auto-scaling benchmark method and apparatus thereof - Google Patents

Abstract

The present invention relates to a benchmark method and apparatus for a high performance computing device. An auto scaling method according to an embodiment of the present invention is a method performed by a computing device, using a first benchmark tool corresponding to a first parameter among a plurality of parameters included in input information, the computing device Performing a first benchmark for the computing device using a second benchmark tool corresponding to a second parameter different from the first parameter, and performing a second benchmark for the computing device and the first benchmark. The method may include providing benchmark result information based on the mark and the result of the second benchmark.

Description

Auto scaling benchmark method and device thereof {AUTO-SCALING BENCHMARK METHOD AND APPARATUS THEREOF}

The present invention relates to a benchmark method and apparatus for a high performance computing device. More specifically, it relates to a method and apparatus for performing a benchmark while changing various parameters within a specified range.

In order to rapidly derive a result desired by a user based on a large amount of data input into the computing device, the computing speed of the computing device corresponding to the result must necessarily be supported. In particular, in response to the demand for technological advancement in the field of artificial intelligence, the demand for technological advancement for high-performance computing devices for increasing computation speed is also increasing.

The technical field of evaluating the performance of such high-performance computing devices (HPC) is also increasing in importance in response to the demand for technological development of high-performance computing devices.

Conventional technologies for evaluating the performance of high-performance computing devices do not support inputs for various parameters, and thus a user has to specify test ranges of parameters for each individual parameter during performance evaluation. This problem of specifying test ranges of parameters one by one at each performance evaluation is accompanied by a problem that it is not possible to continuously evaluate the computing device. Therefore, a technique for evaluating the performance of a high performance computing device supporting inputs for various parameters is required.

In addition, conventional techniques for evaluating the performance of high-performance computing devices do not support a function of providing profiling data based on hardware performance counters. Accordingly, users desiring profiling data required additional efforts such as extracting and processing hardware performance counter-based data. Accordingly, a technique for providing hardware performance counter-based profiling data is also required.

A technical problem to be solved through some embodiments of the present invention is to provide a device for evaluating the performance of a high-performance computing device and a method performed in the device.

Another technical problem to be solved through some embodiments of the present invention is to provide a benchmark device that supports input for various parameters and a method performed in the device.

Another technical problem to be solved through some embodiments of the present invention is to provide a device for providing hardware performance counter-based profiling data and a method performed by the device.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

In order to solve the above technical problem, an auto scaling benchmark method according to an embodiment of the present invention is a method performed by a computing device, wherein a first parameter corresponding to a first parameter among a plurality of parameters included in input information Performing a first benchmark on the computing device using a benchmark tool, and performing a second benchmark on the computing device using a second benchmark tool corresponding to a second parameter different from the first parameter. The method may include performing a mark and providing benchmark result information based on the results of the first benchmark and the second benchmark.

In an embodiment, the method may further include initializing a plurality of parameters included in the input information.

In one embodiment, the computing device may include a plurality of computing nodes. Here, the input information includes an execution indicator indicating a computing node to be benchmarked, and the step of performing the first benchmark and the second benchmark, if the execution indicator indicates two or more computing nodes, The method may include performing the first benchmark and the second benchmark for the benchmark target computing node based on a Message Passing Interface (MPI) command.

In one embodiment, the performing of the first benchmark includes performing the first benchmark while changing the first parameter within a predetermined range, and performing the second benchmark. may include changing the second parameter within a predetermined range and performing the second benchmark.

In an embodiment, the method may further include performing a benchmark according to a parameter order for each of the plurality of parameters. Here, the performing of the benchmark includes repeatedly performing a predetermined number of times for each of the plurality of parameters, and the providing of the benchmark result information includes benchmarking according to the iteratively performing the benchmark. Based on the average value of the result, providing the benchmark result information may be included.

In an embodiment, the method may further include providing profiling result information generated for the computing device when the input information includes a hardware performance counter event. Here, the providing of the profiling result information may include providing profiling result information generated for the computing device based on a Perf command. In addition, the computing device includes a plurality of computing nodes, and the providing of the profiling result information includes generating profiling result information for each computing node and processing the profiling result information generated for each computing node. , providing processed profiling result information. The providing of the profiling result information may include repeatedly generating the profiling result information for the computing device a predetermined number of times, and may further include providing an average value of the profiling result information. can

A computer readable recording medium according to another embodiment of the present invention is a computer program for benchmarking a computing device, including computer program instructions executable by a processor, wherein the computer program instructions are executed by a processor of the computing device. , performing a first benchmark on the computing device using a first benchmark tool corresponding to a first parameter among a plurality of parameters included in the input information, and performing a second benchmark different from the first parameter. Performing a second benchmark on the computing device using a second benchmark tool corresponding to the parameter, and providing benchmark result information based on the results of the first benchmark and the second benchmark. A computer program performing operations, including the step of doing, may be recorded.

An auto scaling benchmark apparatus according to another embodiment of the present invention includes a processor, a network interface, a memory, and a computer program loaded into the memory and executed by the processor, wherein the computer program includes input An instruction for performing a first benchmark on the computing device using a first benchmark tool corresponding to a first parameter among a plurality of parameters included in the information, corresponding to a second parameter different from the first parameter An instruction for performing a second benchmark for the computing device using a second benchmark tool and an instruction for providing benchmark result information based on the results of the first benchmark and the second benchmark can include Here, when a hardware performance counter event is included in the input information, the method may further include an instruction for providing profiling result information generated with respect to the computing device.

1 illustrates an exemplary environment to which an auto scaling benchmark apparatus according to an embodiment of the present invention can be applied.
FIG. 2 is an exemplary block diagram illustrating the auto scaling benchmark apparatus shown in FIG. 1 in more detail.
3 is an exemplary flowchart illustrating an auto scaling benchmark method according to another embodiment of the present invention.
FIG. 4 is an exemplary flowchart illustrating an operation of installing the benchmark tool described with reference to FIG. 3 .
FIG. 5 is an exemplary flowchart illustrating operations of performing a benchmark on a computing device using the benchmark tool described with reference to FIG. 3 .
6 is an exemplary flowchart illustrating an auto scaling benchmark method according to another embodiment of the present invention.
7 is an exemplary diagram illustrating input information that may be referred to in some embodiments of the present invention.
8 is an exemplary diagram illustrating scripts and execution commands generated based on input information that may be referred to in some embodiments of the present invention.
9 is an exemplary hardware configuration diagram capable of implementing an auto scaling benchmark apparatus according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the technical idea of the present invention is not limited to the following embodiments and can be implemented in various different forms, only the following embodiments complete the technical idea of the present invention, and in the technical field to which the present invention belongs It is provided to fully inform those skilled in the art of the scope of the present invention, and the technical spirit of the present invention is only defined by the scope of the claims.

In adding reference numerals to components of each drawing, it should be noted that the same components have the same numerals as much as possible even if they are displayed on different drawings. In addition, in describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description will be omitted.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used in a meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined. Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase.

In addition, in describing the components of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term. When an element is described as being “connected,” “coupled to,” or “connected” to another element, that element is or may be directly connected to the other element, but there is another element between the elements. It will be understood that elements may be “connected”, “coupled” or “connected”.

As used herein, "comprises" and/or "comprising" means that a stated component, step, operation, and/or element is the presence of one or more other components, steps, operations, and/or elements. or do not rule out additions.

Hereinafter, for a clear understanding of various embodiments of the present invention, some terms described in this specification will be defined.

Benchmark refers to comparing and evaluating the performance of hardware or software to quantify them in order to build an optimized system environment. In addition, the profile refers to dynamic program analysis that measures time complexity, memory, use of a specific instruction, function call cycle, function call frequency, and the like of a program. In this specification, the performance evaluation result value of the computing device evaluated by the benchmark tool will be referred to as benchmark result information. In addition, the profiling result value based on the hardware performance counter will be referred to as profiling result information.

In addition, the auto scaling benchmark means that benchmarking is performed while changing each of various parameters within a specified range.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 illustrates an exemplary environment to which the auto scaling benchmark apparatus 100 according to an embodiment of the present invention can be applied. Figure 1 only shows a preferred embodiment for achieving the object of the present invention, some components may be added or deleted as needed. In addition, it should be noted that components of the exemplary environment shown in FIG. 1 represent functionally differentiated functional elements, and a plurality of components may be implemented in a form integrated with each other in an actual physical environment. For example, the auto scaling benchmark device 100 and the computing node 400 may be implemented in the form of different logics in the same computing device. Hereinafter, each component shown in FIG. 1 will be described in more detail.

The auto scaling benchmark device 100 may perform a benchmark for a computing device. More specifically, the auto scaling benchmark device 100 may perform a benchmark for at least one of a plurality of computing nodes included in the computing device. Also, the auto scaling benchmark device 100 may generate profiling result information based on a hardware performance counter of a computing device. More specifically, the auto scaling benchmark apparatus 100 may generate profiling result information based on a hardware performance counter for at least one of a plurality of computing nodes included in the computing device. In order to exclude redundant description, specific methods for the auto scaling benchmark device 100 to perform a benchmark on a computing device or a profile based on a hardware performance calculator on a computing device will be described later in the specification. to materialize through

Next, the data storage 200 may store input information input to the auto scaling benchmark apparatus 100 . In addition, the data storage 200 may store benchmark result information and profiling result information provided to the user by the auto scaling benchmark apparatus 100 .

Next, the benchmark tool storage 300 may include a request processing unit 310 and a storage space 320 . It should be noted that the benchmark tool files shown in the storage space 320 described above represent examples of benchmark tool files that may be stored in the storage space 320 and do not limit the scope of the present invention.

In response to the benchmark tool request 10 of the auto scaling benchmark device 100, the benchmark tool storage 300 stores the benchmark tool file 20 stored in the storage space 320 as the auto scaling benchmark device 100. ) can be provided. More specifically, if there is a benchmark tool request 10 from the auto scaling benchmark device 100, the request processing unit 310 stores a benchmark tool file corresponding to the benchmark tool request 10 in the storage space 320 ( 20) to see if it exists. When the benchmark tool file 20 is retrieved from the storage space 320, the request processing unit 310 sends the benchmark tool file 20 corresponding to the benchmark tool request 10 to the auto scaling benchmark device 100. send A benchmark tool installation operation performed by the auto-scaling benchmark apparatus 100 using the above-described benchmark tool storage 300 will be detailed later with reference to FIGS. 3 and 4 .

Next, computing node 400 may be included in a computing device. At this time, a plurality of computing nodes 400 may be connected to operate as one system.

Also, the computing node 400 may include a node to be evaluated for performance by the auto scaling benchmark device 100 . It should be noted that those shown in the computing node 400 of FIG. 1 represent examples of computing nodes that may be subject to performance evaluation by the auto scaling benchmark device 100, but do not limit the scope of the present invention. .

Hereinafter, the auto scaling benchmark apparatus 100 according to an embodiment will be described with reference to FIG. 2 for a more detailed description. FIG. 2 is an exemplary block diagram for describing the auto scaling benchmark apparatus 100 shown in FIG. 1 in more detail. Figure 2 only shows a preferred embodiment for achieving the object of the present invention, some components may be added or deleted as necessary. In addition, it should be noted that the exemplary auto scaling benchmark apparatus 100 shown in FIG. 2 represents functionally differentiated functional elements, and a plurality of components may be implemented in a form integrated with each other in an actual physical environment. . For example, they may be implemented in the form of different logics within the same computing device.

Also, the auto scaling benchmark device 100 shown in FIG. 2 may be implemented with one or more computing devices. For example, all functions of the auto scaling benchmark device 100 may be implemented in a single computing device. As another example, the first function of the auto scaling benchmark device 100 may be implemented in a first computing device, and the second function may be implemented in a second computing device. Here, the computing device may be a notebook, a desktop, or a laptop, but is not limited thereto and may include any type of device equipped with a computing function. However, in an environment in which the auto scaling benchmark device 100 needs to provide performance evaluation of a high performance computing device according to various parameters, it may be desirable that the auto scaling benchmark device 100 is also implemented as a high performance server-class computing device. have. An example of the above-described auto scaling benchmark apparatus 100 will be described later with reference to FIG. 9 .

The input parser 110 may convert input information written by a user into a form capable of benchmarking a computing device. For example, input information may be converted into an execution command. Here, the input information includes information about the type and range of the parameter, but refers to information written by the user. For example, referring to FIG. 7 , it can be confirmed that the input information includes the “OMP_NUM_THREADS” parameter and information about a range corresponding to the parameter. Here, although the input information may be written in Json language, it should be noted that the present invention is not limited by the language of the input information.

Next, the benchmark tool installation unit 120 may request a benchmark tool from the benchmark tool storage. In addition, the benchmark tool installation unit 120 may receive the benchmark tool file transmitted from the benchmark tool storage and install the benchmark tool file.

More specifically, if installation of the benchmark tool is required, the benchmark tool installation unit 120 may request the benchmark tool from the benchmark tool storage. If the benchmark tool file exists in the benchmark tool storage, the benchmark tool installation unit 120 may receive the benchmark tool file transmitted from the benchmark tool storage and install the benchmark tool file. If the benchmark tool file does not exist in the benchmark tool storage, the benchmark tool installation unit 120 searches for a benchmark target file from the benchmark target computing node, and if the benchmark target file exists, the benchmark tool installation unit 120 may request a benchmark tool update from a benchmark tool repository. According to the above-described operations of the benchmark tool installation unit 120 and the benchmark tool storage, a benchmark for the computing device may be performed based on the benchmark tool corresponding to the input information. For example, a benchmark for a computing device may be performed based on “hpl_intel_mkl18-1.0” of the “TEST_NAME” item shown in FIG. 7 .

Next, the benchmark tool monitoring unit 130 may provide a list of usable benchmark tools and benchmark information, which is information about items evaluated by each benchmark tool.

Next, the task executor 140 performs a benchmark for the computing device using a benchmark tool. More specifically, the task executor 140 may sequentially initialize the parameters included in the input information. In addition, the task executor 140 changes the first parameter within a predetermined range using a first benchmark tool corresponding to the first parameter, which is a parameter of the highest priority, and performs a first benchmark for the computing device, , It is possible to change the second parameter within a pre-specified range using the second benchmark tool corresponding to the second parameter according to the following order, and perform the second benchmark for the computing device. Here, the task executor 140 may perform a benchmark according to the parameter order for all parameters included in the input information. By performing a benchmark using various parameters according to the above-described operations of the task executor 140, the evaluation of the computing device can be continuously performed.

Next, the profiler 150 may collect profiling result information on the computing device using the Perf command. Also, the profiler 150 may process the collected profiling result information. According to the above-described operations of the profiler 150, profiling result information based on the hardware performance counter of the computing device may be collected.

Next, the result information providing unit 160 may provide benchmark result information to the user based on the result of the benchmark performed by the task executor 140 . In addition, the result information providing unit 160 may provide profiling result information to the user based on the result of the profile performed by the profiler 150 .

Meanwhile, each component shown in FIGS. 1 and 2 may mean software or hardware such as a Field Programmable Gate Array (FPGA) or an Application-Specific Integrated Circuit (ASIC). However, components are not meant to be limited to software or hardware, and may be configured to reside in an addressable storage medium or configured to execute one or more processors. Functions provided in the components may be implemented by more subdivided components, or may be implemented as a single component that performs a specific function by combining a plurality of components.

So far, an auto scaling benchmark device according to an embodiment of the present invention and exemplary environments to which the device can be applied have been described with reference to FIGS. 1 and 2 . According to this embodiment, benchmark result information can be provided to a user by continuously performing benchmarks on various parameters. Also, according to the present embodiment, profiling result information based on hardware performance counters may be provided to the user. Hereinafter, exemplary flowcharts illustrating an auto scaling benchmark method according to some embodiments of the present invention will be described in detail with reference to FIGS. 3 to 6 .

Each step of the methods shown in FIGS. 3 to 6 is only a preferred embodiment for achieving the object of the present invention, and it goes without saying that some steps may be added or deleted as needed.

In addition, each step of the methods shown in FIGS. 3 to 6 may be performed by a computing device. In other words, each step of the methods may be implemented as one or more instructions executed by a processor of a computing device. All steps included in the methods could be performed by one physical computing device, but first steps of the method are performed by a first computing device and second steps of the method are performed by a second computing device. may be Hereinafter, description will be continued on the assumption that each step of the methods is performed by the auto scaling benchmark apparatus 100 illustrated in FIG. 1 . However, for convenience of description, the description of the subject of operation of each step included in the methods may be omitted.

Referring to FIG. 3 , input information is converted in step S100. This step may be performed by the input parser 110 previously described with reference to FIG. 2 .

Next, in step S200, a benchmark tool is installed. This step may be performed by the benchmark tool installation unit 120 described above with reference to FIG. 2 . For a more detailed description of this step, it will be described with reference to FIG. 4 .

Referring to FIG. 4, if the installation of the benchmark tool is required (S210), the benchmark tool is requested in the benchmark tool repository (S220), and if the benchmark tool exists in the benchmark tool repository (S230), the benchmark tool The file is received from the benchmark tool repository and installed (S240). If the benchmark tool file does not exist in the benchmark tool repository (S230), the benchmark target file is searched for in the benchmark target computing node (S250), and if the benchmark target file does not exist as a result of the search (S260), the benchmark Tool update is requested (S270). According to the operations described with reference to FIG. 4 , a benchmark for a computing device may be performed based on a benchmark tool corresponding to input information. It will be described with reference to FIG. 3 again.

Next, in step S300, benchmarking of the computing device is performed using a benchmark tool. This step may be performed by the task executor 140 previously described with reference to FIG. 2 . For a more detailed description of this step, it will be described with reference to FIG. 5 .

Referring to FIG. 5 , parameters included in the input information are initialized in step S310. According to this step, the resource resource of the computing device corresponding to the parameter may be initialized.

Next, in step S320, the first parameter is changed within a predetermined range using a first benchmark tool corresponding to the first parameter, and a first benchmark for the computing device is performed. For example, referring to FIG. 7 , a first benchmark for a computing device may be performed while changing the value of the first parameter “OMP_NUM_THREADS” from an initial value of 72 to a final value of 288.

Next, in step S330, the second parameter is changed within a predetermined range by using a second benchmark tool corresponding to the second parameter, which is a lower order of the first parameter, and a second benchmark is performed for the computing device. For example, referring to FIG. 7 , the second benchmark for the computing device may be performed by changing the value of the second parameter “ProblemSize” from the initial value of 10000 to the final value of 150000.

Next, benchmarking is performed for all parameters included in the input information in step S340 according to the parameter order, in a similar operation to the benchmark execution operation that can be referred to in steps S320 and S330.

In some embodiments related to the benchmark execution operation described with reference to FIG. 5 , the input information includes an execution indicator indicating a benchmark target computing node, and when the execution indicator indicates two or more computing nodes, MPI (Message Passing) A benchmark may be performed on a computing node to be benchmarked based on an Interface) command. According to this embodiment, a benchmark may be performed for two or more computing nodes by using an MPI command. For example, referring to FIG. 7 , since "MPI_HOSTS", an execution indicator, indicates "knl01" and "knl02", benchmarking of target computing nodes may be performed based on the MPI command.

In some other embodiments, the benchmark may be repeatedly performed a predetermined number of times for all parameters included in the input information. By repeatedly performing the benchmark using each parameter, more accurate benchmark result information can be provided to the user. For example, according to FIG. 7 , since “TIME_TO_RUN” is “5”, it can be understood that the benchmark using each parameter is repeatedly performed 5 times. At this time, it should be noted that a predetermined time for resource initialization may be granted whenever the benchmark is repeatedly performed.

It will be described with reference to FIG. 3 again. Referring to FIG. 3 , benchmark result information is provided in step S400 . This step may be performed by the result information provider 160 described above with reference to FIG. 2 .

In some embodiments related to step S400, benchmark result information may be provided based on an average value of benchmark results as the benchmark is repeatedly performed. According to the present embodiment, more accurate benchmark result information can be provided to the user by repeatedly performing the benchmark and providing the benchmark result information based on the average value of the result of each performance.

Referring to FIG. 6 , input information is converted in step S10. A description of this step will be understood by referring to the description of step S100 of FIG. 3 .

Next, in step S20, profiling result information is generated for the computing device based on the Perf command. This step may be performed by the profiler 150 previously described with reference to FIG. 2 . For a more specific example with reference to FIG. 7 , when “PERF_EVENT” is included in the input information, the script is executed in the computing device to be evaluated based on the Perf command, and profiling result information based on the hardware performance counter is generated. It can be. 8 is an exemplary script generated based on input information and an execution command corresponding to the script. Profiling result information may be generated according to the exemplary script shown in FIG. 8 .

In some embodiments related to step S20, profiling result information may be generated for each of a plurality of computing nodes included in the computing device.

In some other embodiments, the generated profiling result information may be processed. Here, profiling result information generated for each node may be combined into one file. Also, garbage data included in the generated profiling result information may be deleted. In addition, the generated profiling result information may be sorted according to a predetermined criterion. It should be noted that processing of profiling result information according to the present embodiment may include all known data post-processing operations.

In some other embodiments, based on the Perf command, profiling result information on the computing device may be repeatedly generated a specified number of times. According to this embodiment, more accurate profiling result information on a computing device can be derived.

Next, profiling result information is provided in step S30. This step may be performed by the result information provider 160 described above with reference to FIG. 2 .

In some embodiments related to step S30, an average value of profiling result information may be provided as profiling result information is repeatedly generated. According to the present embodiment, more accurate profiling result information can be provided to a user by repeatedly generating profiling result information and providing profiling result information based on an average value of each generated profiling result information.

So far, auto scaling benchmark methods according to some embodiments of the present invention have been described with reference to FIGS. 3 to 6 . According to the methods described above, a benchmark for a high-performance computing device including two or more computing nodes may be performed using various parameters. That is, by supporting inputs for various parameters, it is possible to perform continuous benchmarking by various parameters. Also, profiling result information based on a hardware performance counter may be generated.

Hereinafter, an exemplary computing device 1500 capable of implementing an auto scaling benchmark device according to an embodiment of the present invention will be described in more detail with reference to FIG. 9 .

The computing device 1500 includes one or more processors 1510, a bus 1550, a communication interface 1570, a memory 1530 for loading a computer program 1591 executed by the processor 1510, and a computer A storage 1590 for storing the program 1591 may be included. However, only components related to the embodiment of the present invention are shown in FIG. 9 . Accordingly, those skilled in the art to which the present invention pertains can know that other general-purpose components may be further included in addition to the components shown in FIG. 9 .

The processor 1510 controls the overall operation of each component of the computing device 1500 . The processor 1510 includes a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), a Graphic Processing Unit (GPU), or any type of processor well known in the art. It can be. Also, the processor 1510 may perform an operation for at least one application or program for executing a method according to embodiments of the present invention. Computing device 1500 may include one or more processors.

Memory 1530 stores various data, commands and/or information. Memory 1530 may load one or more programs 1591 from storage 1590 to execute a method according to embodiments of the present invention. The memory 1530 may be implemented as a volatile memory such as RAM, but the technical scope of the present invention is not limited thereto.

The bus 1550 provides a communication function between components of the computing device 1500 . The bus 1550 may be implemented as various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 1570 supports wired and wireless Internet communication of the computing device 1500 . Also, the communication interface 1570 may support various communication methods other than Internet communication. To this end, the communication interface 1570 may include a communication module well known in the art.

According to some embodiments, communication interface 1570 may be omitted.

The storage 1590 may non-temporarily store the one or more programs 1591 and various data.

The storage 1590 may be a non-volatile memory such as read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, or the like, a hard disk, a removable disk, or a device well known in the art. It may be configured to include any known type of computer-readable recording medium.

Computer program 1591 may include one or more instructions that when loaded into memory 1530 cause processor 1510 to perform methods/operations in accordance with various embodiments of the invention. That is, the processor 1510 may perform methods/operations according to various embodiments of the present disclosure by executing the one or more instructions.

In the above case, the auto scaling benchmark device according to an embodiment of the present invention may be implemented through the computing device 1500 .

So far, various embodiments of the present invention and effects according to the embodiments have been described with reference to FIGS. 1 to 9 . Effects according to the technical idea of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the specification.

The technical idea of the present invention described with reference to FIGS. 1 to 9 so far can be implemented as computer readable code on a computer readable medium. The computer-readable recording medium may be, for example, a removable recording medium (CD, DVD, Blu-ray disc, USB storage device, removable hard disk) or a fixed recording medium (ROM, RAM, computer-equipped hard disk). can The computer program recorded on the computer-readable recording medium may be transmitted to another computing device through a network such as the Internet, installed in the other computing device, and thus used in the other computing device.

In the above, even though all the components constituting the embodiment of the present invention have been described as being combined or operated as one, the technical spirit of the present invention is not necessarily limited to these embodiments. That is, within the scope of the object of the present invention, all of the components may be selectively combined with one or more to operate.

Although actions are shown in a specific order in the drawings, it should not be understood that the actions must be performed in the specific order shown or in a sequential order, or that all depicted actions must be performed to obtain a desired result. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of the various components in the embodiments described above should not be understood as requiring such separation, and the described program components and systems may generally be integrated together into a single software product or packaged into multiple software products. It should be understood that there is

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art can implement the present invention in other specific forms without changing the technical spirit or essential features. can understand that there is Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. The protection scope of the present invention should be construed according to the claims below, and all technical ideas within the equivalent range should be construed as being included in the scope of the technical ideas defined by the present invention.

Claims (17)

In a method performed by a computing device,
performing a first benchmark on the computing device using a first benchmark tool corresponding to a first parameter among a plurality of parameters included in input information;
performing a second benchmark on the computing device using a second benchmark tool corresponding to a second parameter different from the first parameter; and
Based on the results of the first benchmark and the second benchmark, providing benchmark result information,
Auto scaling benchmark method. According to claim 1,
Further comprising initializing a plurality of parameters included in the input information,
Auto scaling benchmark method. According to claim 1,
The computing device includes a plurality of computing nodes,
Auto scaling benchmark method. According to claim 3,
The input information is
Includes an execution indicator instructing a benchmark target computing node;
Performing the first benchmark and the second benchmark,
If the execution indicator indicates two or more computing nodes, performing the first benchmark and the second benchmark for the benchmark target computing node based on a Message Passing Interface (MPI) command,
Auto scaling benchmark method. According to claim 1,
Performing the first benchmark,
Changing the first parameter within a predetermined range and performing the first benchmark;
Performing the second benchmark,
Changing the second parameter within a pre-specified range and performing the second benchmark,
Auto scaling benchmark method. According to claim 1,
Further comprising performing a benchmark according to the parameter order for each of the plurality of parameters,
Auto scaling benchmark method. According to claim 6,
The step of performing the benchmark,
Comprising the step of repeatedly performing a predetermined number of times for each of the plurality of parameters,
Auto scaling benchmark method. According to claim 7,
The step of providing the benchmark result information,
Providing the benchmark result information based on the average value of the benchmark results according to the iteratively performing,
Auto scaling benchmark method. According to claim 1,
Further comprising providing profiling result information generated for the computing device when the input information includes a hardware performance counter event.
Auto scaling benchmark method. According to claim 9,
Providing the profiling result information,
Providing profiling result information generated for the computing device based on a Perf command,
Auto scaling benchmark method. According to claim 9,
The computing device includes a plurality of computing nodes;
Providing the profiling result information,
generating profiling result information for each computing node; and
Processing profiling result information generated for each computing node and providing processed profiling result information,
Auto scaling benchmark method. According to claim 9,
Providing the profiling result information,
Repetitively generating profiling result information for the computing device a predetermined number of times,
Auto scaling benchmark method. According to claim 12,
Further comprising providing an average value of the profiling result information,
Auto scaling benchmark method. According to claim 1,
Further comprising performing an N-th benchmark on the computing device using an N-th benchmark tool corresponding to the N-th parameter;
The step of providing the benchmark result information,
Based on the result of the Nth benchmark, providing benchmark result information;
The Nth parameter is a parameter different from the first parameter and the second parameter,
Auto scaling benchmark method. A computer program for benchmarking a computing device comprising computer program instructions executable by a processor, the computer program instructions being executed by a processor of the computing device comprising:
performing a first benchmark on the computing device using a first benchmark tool corresponding to a first parameter among a plurality of parameters included in input information;
performing a second benchmark on the computing device using a second benchmark tool corresponding to a second parameter different from the first parameter; and
Based on the results of the first benchmark and the second benchmark, a computer program performing operations comprising providing benchmark result information is recorded,
A computer-readable recording medium. processor;
network interface;
Memory; and
A computer program loaded into the memory and executed by the processor,
The computer program,
instructions for performing a first benchmark on the computing device by using a first benchmark tool corresponding to a first parameter among a plurality of parameters included in the input information;
instructions for performing a second benchmark on the computing device using a second benchmark tool corresponding to a second parameter different from the first parameter; and
Including instructions for providing benchmark result information based on the results of the first benchmark and the second benchmark,
Auto scaling benchmark device. According to claim 16,
Further comprising an instruction for providing profiling result information generated for the computing device when the input information includes a hardware performance counter event.
Auto scaling benchmark device.

Images