公开(公告)号：US07725657B2

公开(公告)日：2010-05-25

申请号：US11726238

申请日：2007-03-21

Applicant: William C. Hasenplaugh ,  Li Zhao ,  Ravishankar Iyer ,  Ramesh Illikkal ,  Srihari Makineni ,  Donald Newell ,  Aamer Jaleel ,  Simon C. Steely, Jr.

Inventor： William C. Hasenplaugh ,  Li Zhao ,  Ravishankar Iyer ,  Ramesh Illikkal ,  Srihari Makineni ,  Donald Newell ,  Aamer Jaleel ,  Simon C. Steely, Jr.

IPC: G06F12/06

CPC classification number: G06F12/084 ,  G06F12/0842

Abstract: In one embodiment, the present invention includes a method for associating a first priority indicator with data stored in a first entry of a shared cache memory by a core to indicate a priority level of a first thread, and associating a second priority indicator with data stored in a second entry of the shared cache memory by a graphics engine to indicate a priority level of a second thread. Other embodiments are described and claimed.

Abstract translation: 在一个实施例中，本发明包括一种用于将第一优先级指示符与存储在共享高速缓冲存储器的第一条目中的数据相关联以指示第一线程的优先级，并且将第二优先级指示符与存储的数据相关联的方法 在图形引擎的共享高速缓冲存储器的第二条目中指示第二线程的优先级。 描述和要求保护其他实施例。

公开(公告)号：US20090172315A1

公开(公告)日：2009-07-02

申请号：US11965131

申请日：2007-12-27

Applicant: Ravishankar Iyer ,  Ramesh Milekal ,  Donald Newell ,  Li Zhao

Inventor： Ravishankar Iyer ,  Ramesh Milekal ,  Donald Newell ,  Li Zhao

IPC: G06F12/00

CPC classification number: G06F12/126 ,  G06F9/5016 ,  G06F12/0888

Abstract: A method and apparatus for is herein described providing priority aware and consumption guided dynamic probabilistic allocation for a cache memory. Utilization of a sample size of a cache memory is measured for each priority level of a computer system. Allocation probabilities for each priority level are updated based on the measured consumption/utilization, i.e. allocation is reduced for priority levels consuming too much of the cache and allocation is increased for priority levels consuming too little of the cache. In response to an allocation request, it is assigned a priority level. An allocation probability associated with the priority level is compared with a randomly generated number. If the number is less than the allocation probability, then a fill to the cache is performed normally. In contrast, a spatially or temporally limited fill is performed if the random number is greater than the allocation probability.

Abstract translation: 本文描述了提供高速缓冲存储器的优先级感知和消耗引导的动态概率分配的方法和装置。 针对计算机系统的每个优先级测量利用高速缓冲存储器的样本大小。 基于所测量的消耗/利用率来更新每个优先级的分配概率，即对于消耗太多高速缓存的优先级降低了分配，并且对于消耗太多缓存的优先级来说，分配增加。 响应于分配请求，它被分配优先级。 将与优先级相关联的分配概率与随机生成的数字进行比较。 如果该数量小于分配概率，则正常地执行对高速缓存的填充。 相比之下，如果随机数大于分配概率，则执行空间或时间有限的填充。

公开(公告)号：US20090165004A1

公开(公告)日：2009-06-25

申请号：US12004756

申请日：2007-12-21

Applicant: Jaideep Moses ,  Don K. Newell ,  Ramesh Illikkal ,  Ravishankar Iyer ,  Srihari Makineni ,  Li Zhao ,  Scott Hahn ,  Tong N. Li ,  Padmashree Apparao

Inventor： Jaideep Moses ,  Don K. Newell ,  Ramesh Illikkal ,  Ravishankar Iyer ,  Srihari Makineni ,  Li Zhao ,  Scott Hahn ,  Tong N. Li ,  Padmashree Apparao

IPC: G06F9/46

CPC classification number: G06F9/505 ,  G06F2209/507 ,  G06F2209/508

Abstract: In one embodiment, a method provides capturing resource monitoring information for a plurality of applications; accessing the resource monitoring information; and scheduling at least one of the plurality of applications on a selected processing core of a plurality of processing cores based, at least in part, on the resource monitoring information.

Abstract translation: 在一个实施例中，一种方法提供捕获用于多个应用的​​资源监视信息; 访问资源监控信息; 以及至少部分地基于所述资源监视信息，在所述多个处理核心的所选择的处理核心上调度所述多个应用中的至少一个。

公开(公告)号：US07552288B2

公开(公告)日：2009-06-23

申请号：US11503777

申请日：2006-08-14

Applicant: Ravishankar Iyer ,  Li Zhao ,  Srihari Makineni ,  Donald Newell

Inventor： Ravishankar Iyer ,  Li Zhao ,  Srihari Makineni ,  Donald Newell

IPC: G06F12/08

CPC classification number: G06F12/0811 ,  G06F12/0831

Abstract: In one embodiment, the present invention includes a method for maintaining data in a first level cache non-inclusively with data in a second level cache coupled to the first level cache. At the same time, at least a portion of directory information associated with the data in the first level cache may be maintained inclusively with a directory portion of the second level cache. Other embodiments are described and claimed.

Abstract translation: 在一个实施例中，本发明包括一种用于在与第一级高速缓存耦合的第二级高速缓存中的数据非包含地维护第一级高速缓存中的数据的方法。 同时，可以与第二级高速缓存的目录部分一起保持与第一级高速缓存中的数据相关联的目录信息的至少一部分。 描述和要求保护其他实施例。

公开(公告)号：US20080244221A1

公开(公告)日：2008-10-02

申请号：US11694322

申请日：2007-03-30

Applicant: Donald K. Newell ,  Jaideep Moses ,  Ravishankar Iyer ,  Rameshkumar G. Illikkal ,  Srihari Makineni

Inventor： Donald K. Newell ,  Jaideep Moses ,  Ravishankar Iyer ,  Rameshkumar G. Illikkal ,  Srihari Makineni

IPC: G06F15/00

CPC classification number: G06F15/16

Abstract: Embodiments of apparatuses, methods, and systems for exposing system topology to an execution environment are disclosed. In one embodiment, an apparatus includes execution cores and resources on a single integrated circuit, and topology logic. The topology logic is to populate a data structure with information regarding a relationship between the execution cores and the resources.

Abstract translation: 公开了将系统拓扑暴露给执行环境的装置，方法和系统的实施例。 在一个实施例中，装置包括在单个集成电路上的执行核心和资源以及拓扑逻辑。 拓扑逻辑是使用关于执行核心和资源之间的关系的信息来填充数据结构。

公开(公告)号：US20200228880A1

公开(公告)日：2020-07-16

申请号：US16833582

申请日：2020-03-28

Applicant: Ravishankar Iyer ,  Nilesh Kumar Jain ,  Rameshkumar Illikkal ,  Carl S. Marshall ,  Selvakumar Panneer ,  Rajesh Poornachandran

Inventor： Ravishankar Iyer ,  Nilesh Kumar Jain ,  Rameshkumar Illikkal ,  Carl S. Marshall ,  Selvakumar Panneer ,  Rajesh Poornachandran

IPC: H04N21/81 ,  H04N21/234 ,  H04N21/235 ,  H04N21/647

Abstract: Various embodiments for dynamically generating an advertisement in a video stream are disclosed. In one embodiment, video stream content associated with a video stream for a user device is received. Video analytics data is obtained for the video stream content, which indicates a scene recognized in the video stream content. An advertisement to be generated and inserted into the video stream content is then selected based on the scene recognized in the video stream content, and an advertisement template for generating the selected advertisement is obtained. Video advertisement content corresponding to the advertisement is then generated based on the advertisement template and the video analytics data. The video advertisement content is then inserted into the video stream content, and the modified video stream content is transmitted to the user device.

公开(公告)号：US09626586B2

公开(公告)日：2017-04-18

申请号：US14320092

申请日：2014-06-30

Applicant: Yong Zhang ,  Ravishankar Iyer ,  Rameshkumar G. Illikkal ,  Donald K. Newell ,  Jianping Zhou

Inventor： Yong Zhang ,  Ravishankar Iyer ,  Rameshkumar G. Illikkal ,  Donald K. Newell ,  Jianping Zhou

IPC: G06K9/00 ,  G06K9/48 ,  G06K9/46

CPC classification number: G06K9/481 ,  G06K9/00973 ,  G06K9/4671 ,  G06K2009/485

Abstract: Methods and systems of recognizing images may include an apparatus having a hardware module with logic to, for a plurality of vectors in an image, determine a first intermediate computation based on even pixels of an image vector, and determine a second intermediate computation based on odd pixels of an image vector. The logic can also combine the first and second intermediate computations into a Hessian matrix computation.

公开(公告)号：US20160306415A1

公开(公告)日：2016-10-20

申请号：US15192134

申请日：2016-06-24

Applicant: Andrew J. Herdrich ,  Rameshkumar G. Illikkal ,  Ravishankar Iyer ,  Sadagopan Srinivasan ,  Jaideep Moses ,  Srihari Makineni

Inventor： Andrew J. Herdrich ,  Rameshkumar G. Illikkal ,  Ravishankar Iyer ,  Sadagopan Srinivasan ,  Jaideep Moses ,  Srihari Makineni

IPC: G06F1/32 ,  G06F12/08

CPC classification number: G06F1/3293 ,  G06F1/3206 ,  G06F1/3287 ,  G06F9/4418 ,  G06F9/5094 ,  G06F12/084 ,  G06F13/24 ,  G06F2212/1028 ,  G06F2212/60 ,  G06F2212/62 ,  H04W52/028 ,  H04W88/02 ,  Y02B70/30 ,  Y02B70/32 ,  Y02D10/122 ,  Y02D10/22 ,  Y02D70/00

Abstract: In one embodiment, the present invention includes a method for receiving an interrupt from an accelerator, sending a resume signal directly to a small core responsive to the interrupt and providing a subset of an execution state of the large core to the first small core, and determining whether the small core can handle a request associated with the interrupt, and performing an operation corresponding to the request in the small core if the determination is in the affirmative, and otherwise providing the large core execution state and the resume signal to the large core. Other embodiments are described and claimed.

公开(公告)号：US20160182345A1

公开(公告)日：2016-06-23

申请号：US14581595

申请日：2014-12-23

Applicant: Andrew J. Herdrich ,  Patrick Connor ,  Dinesh Kumar ,  Alexander W. Min ,  Ravishankar Iyer ,  Daniel J. Dahle ,  Kapil Sood ,  Jeffrey B. Shaw

Inventor： Andrew J. Herdrich ,  Patrick Connor ,  Dinesh Kumar ,  Alexander W. Min ,  Ravishankar Iyer ,  Daniel J. Dahle ,  Kapil Sood ,  Jeffrey B. Shaw

IPC: H04L12/26

CPC classification number: H04L43/50 ,  G06F12/0811 ,  G06F12/084 ,  G06F2212/152 ,  G06F2212/601 ,  H04L41/0896 ,  H04L41/5009 ,  H04L41/5025 ,  H04L41/5035 ,  H04L41/5096 ,  H04L43/16

Abstract: In embodiments, apparatuses, methods and storage media (transitory and non-transitory) are described that are associated with end-to-end datacenter performance control. In various embodiments, an apparatus for computing may receive a datacenter performance target, determine an end-to-end datacenter performance level based at least in part on quality of service data collected from a plurality of nodes, and send a mitigation command based at least in part on a result of a comparison of the end-to-end datacenter performance level determined to the datacenter performance target. In various embodiments, the apparatus for computing may include one or more processors, a memory, a datacenter performance monitor to receive a datacenter performance target corresponding to a service level agreement, and a mitigation module to send a mitigation command based at least in part on a result of a comparison of an end-to-end datacenter performance level to a datacenter performance target.

Abstract translation: 在实施例中，描述了与端对端数据中心性能控制相关联的装置，方法和存储介质（暂时性和非暂时性）。 在各种实施例中，用于计算的装置可以接收数据中心性能目标，至少部分地基于从多个节点收集的服务质量数据来确定端对端数据中心性能级别，并且至少基于从多个节点收集的服务数据发送缓解命令 部分原因是比较确定数据中心性能目标的端对端数据中心性能级别。 在各种实施例中，用于计算的装置可以包括一个或多个处理器，存储器，用于接收与服务水平协议相对应的数据中心性能目标的数据中心性能监视器，以及缓解模块，用于至少部分地基于 这是将端到端数据中心性能级别与数据中心性能目标进行比较的结果。

公开(公告)号：US08799902B2

公开(公告)日：2014-08-05

申请号：US11786019

申请日：2007-04-09

Applicant: Ramesh Kumar Illikkal ,  Ravishankar Iyer ,  Jaideep Moses ,  Don Newell ,  Tryggve Fossum

Inventor： Ramesh Kumar Illikkal ,  Ravishankar Iyer ,  Jaideep Moses ,  Don Newell ,  Tryggve Fossum

IPC: G06F9/46 ,  G06F1/00 ,  G06F1/26 ,  G06F1/32 ,  G06F15/173

CPC classification number: G06F9/5077 ,  G06F9/5094 ,  G06F2209/504 ,  Y02D10/22 ,  Y02D10/36

Abstract: A method and apparatus for throttling power and/or performance of processing elements based on a priority of software entities is herein described. Priority aware power management logic receives priority levels of software entities and modifies operating points of processing elements associated with the software entities accordingly. Therefore, in a power savings mode, processing elements executing low priority applications/tasks are reduced to a lower operating point, i.e. lower voltage, lower frequency, throttled instruction issue, throttled memory accesses, and/or less access to shared resources. In addition, utilization logic potentially trackes utilization of a resource per priority level, which allows the power manager to determine operating points based on the effect of each priority level on each other from the perspective of the resources themselves. Moreover, a software entity itself may assign operating points, which the power manager enforces.

Abstract translation: 这里描述了一种基于软件实体的优先级来节制处理元件的功率和/或性能的方法和装置。 优先级感知功率管理逻辑接收软件实体的优先级，并相应地修改与软件实体相关联的处理元件的工作点。 因此，在省电模式中，执行低优先级应用/任务的处理元件被降低到较低的工作点，即较低的电压，较低的频率，节流的指令问题，节流的存储器访问和/或较少的对共享资源的访问。 此外，利用逻辑潜在地追踪每个优先级别的资源的利用率，这允许电力管理者从资源本身的角度基于每个优先级的影响来确定工作点。 此外，软件实体本身可以分配功率管理器执行的操作点。