-
1.发明公开그래픽스 프로세싱 장치 및 그래픽스 파이프라인의 텍스쳐링을 위한 LOD(level of detail)를 결정하는 방법 审中-实审用于确定图形处理装置和图形管线的纹理化的细节等级（LOD）的方法
公开(公告)号：KR1020170036419A
公开(公告)日：2017-04-03
申请号：KR1020150135551
申请日：2015-09-24
Applicant: 삼성전자주식회사
CPC classification number: G06T11/001 , G06T1/20 , G06T11/40 , G06T15/04 , G06T2210/36
Abstract: 그래픽스프로세싱장치에서텍스쳐링을위한 LOD를결정하는방법은, 복수의쿼드들중 레퍼런스쿼드를결정하고, 레퍼런스쿼드의텍스쳐좌표들을이용하여레퍼런스쿼드와레퍼런스쿼드를포함하는블록간의유사성을판단하고, 유사성판단결과에기초하여레퍼런스쿼드에대한 LOD를, 주변쿼드들에대한 LOD들로결정한다.
Abstract translation: 如何确定LOD为纹理化的图形处理装置，其中，确定多个四边形的基准四元，并通过使用参考四边形的纹理坐标，并确定包括参考四，相似性判断基准四元块之间的相似性 基于结果，参考四元组的LOD被确定为周围四元组的LOD。
-
公开(公告)号：KR101669989B1
公开(公告)日：2016-10-28
申请号：KR1020100049852
申请日：2010-05-27
Applicant: 삼성전자주식회사
CPC classification number: G06F11/1407 , G06F9/3857 , G06F9/3861 , G06F9/3867
Abstract: 이퀄모델상의제약조건을만족시키는파이프라인프로세서가제공된다. 제공되는파이프라인프로세서는적어도한 개의인스트럭션을다수의스테이지로나누어서처리하는파이프라인처리부및 파이프라인처리부내부의인스트럭션전부또는일부의처리결과를저장하고, 저장된처리결과를각각의인스트럭션의레이턴시를고려하여레지스터파일에기록하는이퀄모델보상부를포함할수 있다.
-
公开(公告)号：KR1020120083800A
公开(公告)日：2012-07-26
申请号：KR1020110005141
申请日：2011-01-18
Applicant: 삼성전자주식회사
CPC classification number: G06F9/30189 , G06F9/30072 , G06F11/3636 , G06F15/7867
Abstract: PURPOSE: A trace data compressing apparatus and a method thereof are provided to select and compress trace data of a FU(Function Unit) performing an actually significant operation except trace data of a FU not performing the actually significant operation, thereby reducing quantity of the generated trace data. CONSTITUTION: A detection unit(301) detects valid trace data which is trace data of a FU performing an actually significant operation among entire trace data of a FU included in a reconfigurable processor. A compression unit(302) compresses the valid trace data. The detection unit confirms the FU, performing an actually significant operation, by using schedule information, provided by a compiler of the reconfigurable processor, and detects the trace data of the confirmed FU as the valid trace data.
Abstract translation: 目的：提供跟踪数据压缩装置及其方法，用于选择和压缩执行实际有效操作的FU（功能单元）的跟踪数据，除了不执行实际有效操作的FU的跟踪数据，从而减少产生的 跟踪数据。 构成：检测单元（301）检测作为可重构处理器中包含的FU的整个轨迹数据中执行实际有效操作的FU的跟踪数据的有效跟踪数据。 压缩单元（302）压缩有效的跟踪数据。 检测单元通过使用由可重配置处理器的编译器提供的调度信息来确认FU执行实际有效的操作，并将确认的FU的跟踪数据检测为有效的跟踪数据。
-
-
公开(公告)号：KR101738640B1
公开(公告)日：2017-05-24
申请号：KR1020110005141
申请日：2011-01-18
Applicant: 삼성전자주식회사
CPC classification number: G06F9/30189 , G06F9/30072 , G06F11/3636
Abstract: 생성되는트레이스데이터의양을줄일수 있는트레이스데이터압축장치및 방법이제공된다. 본발명의일 양상에따른트레이스데이터압축장치는, 재구성가능프로세서에포함된각각의펑션유닛(function unit, FU)의전체트레이스데이터(trace data)에서실질적으로의미있는동작을수행하는펑션유닛의트레이스데이터인유효트레이스데이터를검출하는검출부, 및검출된유효트레이스데이터를압축하는압축부를포함할수 있다.
Abstract translation: 提供了一种能够减少生成的跟踪数据量的跟踪数据压缩设备和方法。 根据本发明的一个方面的跟踪数据压缩装置包括跟踪单元，用于对可重构处理器中包括的每个功能单元（FU）的全部跟踪数据执行实质上有意义的操作， 检测单元，用于检测作为数据的有效跟踪数据;压缩单元，用于压缩检测到的有效跟踪数据。
-
Patent Agency Ranking
公开(公告)号：KR1020140126194A
公开(公告)日：2014-10-30
申请号：KR1020130044434
申请日：2013-04-22
Applicant: 삼성전자주식회사
CPC classification number: G06F17/5081 , G06F11/3684 , G06F17/5009 , G06F9/44 , G06F11/36
Abstract: 프로세서의 검증을 위하여 중복 없는 테스트 케이스를 생성하는 장치에 관한 것으로, 일 실시예에 따르면, 전체 검증 공간(total verification space)에서 복수의 제약 검증 공간(constrained verification space)을 정의하고, 복수의 제약 검증 공간 각각에 대하여 제약 조건 기술(constrined description)을 생성하는 제약조건생성부 및 제약 조건 기술을 이용하여 테스트 케이스를 생성하는 테스트케이스생성부를 포함할 수 있다.
Abstract translation: 本发明涉及一种创建不重叠以验证处理器的测试用例的设备。 根据实施例，设备可以包括：约束条件生成单元，其在总验证空间中定义多个约束验证空间，并且为每个约束验证空间生成约束条件描述; 以及使用约束条件描述生成测试用例的测试用例生成单元。
-
公开(公告)号：KR1020130086489A
公开(公告)日：2013-08-02
申请号：KR1020120007383
申请日：2012-01-25
Applicant: 삼성전자주식회사
CPC classification number: G06F9/38 , G06F11/3636 , G06F11/3648
Abstract: PURPOSE: A hardware debugging device of a program and a method thereof are provided to solve currency problems by generating predication in advance. CONSTITUTION: A predication setting unit (110) sets a trigger repetition count value based on execution block difference information when trigger is generated. A predication initial generation unit (130) generates predication for an execution block based on the trigger repetition count value in advance. The repetition count is the repetition count of calculation executing a program in each function unit of a processor. The predication is initially generated in an execution block difference. [Reference numerals] (110) Predication setting unit; (120) Process reprocessing unit; (130) Predication initial generation unit; (210A) Register file recovery buffer; (220A) Memory recovery buffer
Abstract translation: 目的：提供程序的硬件调试装置及其方法，以便通过事先产生预测来解决货币问题。 构成：当产生触发时，预测设定单元（110）基于执行块差异信息设置触发重复计数值。 预测初始生成单元（130）预先基于触发重复计数值生成对执行块的预测。 重复计数是在处理器的每个功能单元中执行程序的重复计数。 预测最初在执行块的差异中生成。 （附图标记）（110）预测设定单元; （120）加工后处理单元; （130）预测初始生成单位; （210A）寄存器文件恢复缓冲区; （220A）内存恢复缓冲区
-
公开(公告)号：KR1020050088782A
公开(公告)日：2005-09-07
申请号：KR1020040014245
申请日：2004-03-03
Applicant: 삼성전자주식회사
IPC: G01N21/64
CPC classification number: G01N21/648 , G01N21/6454 , G01N21/6404 , G01N21/63 , G01N2021/945 , G01N2223/412 , G02B27/0927
Abstract: An optical detection device for biochemically analyzing samples of microfluidic chips such as DNA chips or protein chips is provided. The optical detection device includes a light source for generating excitation light to excite a sample, a detection plate having an optical waveguide with the sample mounted thereon, a linear symmetry lens for shaping the excitation light into a desired shape having a uniform intensity distribution and transmitting it to the detection plate, and a detector for detecting radiation from the sample.
-
-
-
-
-
-
-
-
Search Results
27
records
(took 0.018 seconds)
