-
公开(公告)号：KR1020130021172A
公开(公告)日：2013-03-05
申请号：KR1020110083541
申请日：2011-08-22
Applicant: 삼성전자주식회사
Abstract: PURPOSE: A terminal and a method for executing an application in the terminal are provided to improve the performance of the whole terminal and dynamic resource allocation by generating a code corresponding to an application as a binary type by confirming the performance and workload of a CPU and a GPU during compiling. CONSTITUTION: When a code for an application to be executed is inputted, a workload of an operation processing unit and a graphic operation processing unit is confirmed(520). The workload of the operation processing unit and the graphic operation processing unit is compared with a threshold value of a set processing amount(530). The code is compiled according to the comparison result(535). The operation processing unit and/or graphic operation processing unit generates a binary for executing the application by using the compiled code(540). The application is executed through the generated binary(550). [Reference numerals] (510) Inputting a code for an application to be executed; (515) Generating an intermediate code; (520) Confirming workloads of a CPU and a GPU; (525) Workload of the CPU is equal to or greater than a first processing amount threshold?; (530) Workload of the CPU is equal to or greater than a second processing amount threshold?; (535) Performing compiling for a binary for the GPU; (540) Generating a binary for the GPU; (545) Feeding back for each generated binary; (550) Executing the corresponding application; (560) Performing compiling for a binary for the CPU; (565) Generating a binary for the CPU; (AA) Start; (BB) End
Abstract translation: 目的：提供一种用于在终端中执行应用的终端和方法，以通过产生与应用程序相对应的代码作为二进制类型，通过确定CPU的性能和工作负载来提高整个终端的性能和动态资源分配， 一个GPU在编译时。 构成：当输入要执行的应用程序的代码时，确认操作处理单元和图形操作处理单元的工作量（520）。 操作处理单元和图形操作处理单元的工作量与设定处理量的阈值（530）进行比较。 代码根据比较结果进行编译（535）。 操作处理单元和/或图形操作处理单元通过使用编译代码生成用于执行应用程序的二进制码（540）。 该应用程序通过生成的二进制执行（550）。 （附图标记）（510）输入要执行的应用的代码; （515）生成中间代码; （520）确认CPU和GPU的工作负载; （525）CPU的工作负载等于或大于第一处理量阈值α; （530）CPU的工作负载等于或大于第二处理量阈值α; （535）执行GPU的二进制编译; （540）为GPU生成二进制码; （545）为每个生成的二进制文件反馈; （550）执行相应的申请; （560）执行CPU的二进制编译; （565）为CPU生成二进制数据; （AA）开始; （BB）结束
-
公开(公告)号：KR1020130087257A
公开(公告)日：2013-08-06
申请号：KR1020120008453
申请日：2012-01-27
Applicant: 삼성전자주식회사
IPC: G06F9/50
CPC classification number: G06F9/5011 , G06F9/5066 , G06T1/20
Abstract: PURPOSE: A method for allocating resources of a graphic processing unit (GPU) and a device thereof are provided to reduce the request processing number of the GPU by improving the efficiency of core application. CONSTITUTION: A memory (230) stores requests, and a control unit (210) stores operations by receiving a task including the operations. The control unit sets the request unit by combining the data of the operations and allocates resources of a GPU (240) according to the execution order of the request. The control unit sets the operation according to the request units and determines whether the number of the requests is over threshold values. [Reference numerals] (220) GPU resource assignor; (230) Memory
Abstract translation: 目的：提供一种用于分配图形处理单元（GPU）及其装置的资源的方法，以通过提高核心应用的效率来减少GPU的请求处理次数。 构成：存储器（230）存储请求，并且控制单元（210）通过接收包括操作的任务来存储操作。 控制单元通过组合操作的数据来设置请求单元，并根据请求的执行顺序分配GPU（240）的资源。 控制单元根据请求单元设置操作，并确定请求数是否超过阈值。 （220）GPU资源分配器; （230）内存
-
Patent Agency Ranking
-
公开(公告)号：KR101885211B1
公开(公告)日：2018-08-29
申请号：KR1020120008453
申请日：2012-01-27
Applicant: 삼성전자주식회사
IPC: G06F9/50
CPC classification number: G06F9/5011 , G06F9/5066 , G06T1/20
Abstract: 본발명은 GPU의자원을효율적으로할당하기위한방법및 장치에대한것으로, 자원할당방법은적어도하나이상의오퍼레이션을포함하는테스크(Task)를수신하고, 상기오퍼레이션각각을리퀘스트단위로구분하여저장하는단계; 둘이상의동일한오퍼레이션을하나의리퀘스트단위로설정하기위하여머징(Merging)하는단계; 및상기리퀘스트의실행순서에따라상기 GPU의자원을할당하는단계를포함하는것을특징으로한다. 본발명에따르면, 오퍼레이션레벨에서 GPU 자원을효율적으로할당할수 있는효과가있다.
-
公开(公告)号：KR1020140011671A
公开(公告)日：2014-01-29
申请号：KR1020120078326
申请日：2012-07-18
Applicant: 삼성전자주식회사
Abstract: The present invention relates to a data processing method and a device using a graphic processing unit (GPU). The data processing method using the GPU according to an embodiment of the present invention include a step of obtaining calculation target data and operator information applied to calculation target data; a step of dividing the calculation target data into a channel element size unit of texel of the GPU; a step of controlling the GPU to return a calculation result by performing calculation on the divided data; a step of transmitting the divided data to the GPU; and a step of obtaining a calculation result value from the GPU. In accordance with the embodiment of the present invention, the present invention can provide the data processing device using the GPU capable of improving calculation performance and the method thereof. [Reference numerals] (112) Data handler; (114) Frame buffer manager; (205) Calulator, data size; (210) Frame buffer size determination; (215) Frame buffer generation command; (220) Frame buffer generation; (225) Calculation target data, the kind of data; (230) Division of data; (235) Determination of the kind of texts; (240) Texel generation command, data transmission; (245) Texel generation, store data; (250) Calculation command generation; (255) Calculation command; (262) Restore data; (265) Calculation; (267) Store a frame buffer; (270) Bringing a frame buffer value; (275) Frame buffer value; (280) Recover a result value; (285) Extinction of the texel; (290) Extinction of the frame buffer; (295) Result value
Abstract translation: 本发明涉及一种数据处理方法和使用图形处理单元（GPU）的装置。 使用根据本发明的实施例的GPU的数据处理方法包括获得应用于计算目标数据的计算目标数据和操作者信息的步骤; 将计算目标数据划分成GPU的纹素像素的通道元素大小单位的步骤; 通过对划分的数据进行计算来控制GPU返回计算结果的步骤; 将划分的数据发送到GPU的步骤; 以及从GPU获得计算结果值的步骤。 根据本发明的实施例，本发明可以提供使用能够提高计算性能的GPU的数据处理装置及其方法。 （附图标记）（112）数据处理程序; （114）帧缓冲管理器; （205）计数器，数据大小; （210）帧缓冲区大小确定; （215）帧缓冲区生成命令; （220）帧缓冲区生成; （225）计算目标数据，数据种类; （230）数据分割; （235）确定文本的种类; （240）Texel生成命令，数据传输; （245）Texel生成，存储数据; （250）计算命令生成; （255）计算命令; （262）恢复数据; （265）计算; （267）存储帧缓冲区; （270）引入帧缓冲区值; （275）帧缓冲器值; （280）恢复结果值; （285）纹身的灭绝; （290）帧缓冲器的消光; （295）结果值
-
公开(公告)号：KR101845328B1
公开(公告)日：2018-04-04
申请号：KR1020110083541
申请日：2011-08-22
Applicant: 삼성전자주식회사
CPC classification number: G06F9/505 , G06F7/504 , G06F9/5044 , G06T1/20 , Y02D10/22
Abstract: 본발명은단말및 그단말에서어플리케이션을수행하는방법에관한것으로, 수행될어플리케이션에대한코드가입력되면, 연산처리부와그래픽연산처리부의워크로드를확인하는과정과, 상기확인된연산처리부의워크로드와그래픽연산처리부의워크로드를미리설정된처리량임계값과비교하는과정과, 상기비교결과에따라상기코드를컴파일하는과정과, 상기컴파일된코드를이용하여상기연산처리부또는상기그래픽연산처리부에서어플리케이션수행을위한바이너리를생성하는과정과, 상기생성된바이너리를통해상기어플리케이션을수행하는과정으로구성된다. 따라서 CPU와 GPU에서처리되는작업량을조절할수 있으며, 전체시스템의성능이향상되며, 수행시간이짧아지기때문에저전력의효과를얻을수 있다.
Abstract translation: 本发明是一种终端和一种用于在终端上执行的应用程序的方法，该代码被输入到该应用程序被执行时，图形操作处理的操作部的工作，并且所识别的操作处理，以确定处理负荷的工作负荷 和图形比较预设吞吐量阈值，编译基于比较结果的代码的过程的操作部分的工作量，并进行了使用经编译的代码中的操作处理的应用或图形运算处理 为生成的二进制生成二进制文件，并通过生成的二进制文件执行应用程序。 因此，可以控制CPU和GPU处理的工作量，提高整个系统的性能，缩短执行时间，从而实现低功耗。
-
公开(公告)号：KR1020130019674A
公开(公告)日：2013-02-27
申请号：KR1020110081775
申请日：2011-08-17
Applicant: 삼성전자주식회사
CPC classification number: G06F9/505 , G06F2209/509
Abstract: PURPOSE: A terminal and a method for performing an application in the terminal are provided to prevent a workload from being leant on a processing unit by allocating an application to a CPU or a GPU(Graphic Processing Unit) according to a weighted valued designated by a user. CONSTITUTION: A storage unit(130) stores a workload control information table. If a code of an application is inputted, a CPU(110) confirms a weighted value of the application stored in the workload control information table to calculate an allocation index and selects a processing device to perform the application. A GPU(Graphic Processing Unit)(120) performs the application through a binary delivered from the CPU. The CPU confirms a priority processing mode included in the workload control information table and a processing threshold value according to the priority processing mode and compares the processing threshold value with the allocation value to select the processing device. [Reference numerals] (117) GPU workload monitoring unit; (130) Storage unit; (135) Workload control information table
Abstract translation: 目的：提供一种用于在终端中执行应用的终端和方法，以通过根据由CPU或GPU（图形处理单元）指定的加权值向CPU或GPU（图形处理单元）分配应用来防止工作负担在处理单元上 用户。 构成：存储单元（130）存储工作负载控制信息表。 如果应用程序的代码被输入，则CPU（110）确认存储在工作负载控制信息表中的应用的加权值，以计算分配索引并选择执行应用的处理装置。 GPU（图形处理单元）（120）通过从CPU传递的二进制程序执行应用程序。 CPU根据优先处理模式确认工作负载控制信息表中包含的优先处理模式和处理阈值，并将处理阈值与分配值进行比较，选择处理装置。 （附图标记）（117）GPU工作负荷监视单元; （130）存储单元; （135）工作量控制信息表
-
-
-
-
-
-
Search Results
7
records
(took 0.019 seconds)
