返回项目页面
对比文件列表
1998-02-15_发明专利_ATE162641T1_+++M_N_a_b_c_d_e_g_l_p+++.docx
2026-02-20 00:16
2002-07-04_发明申请_US20020087903A1 Mechanism for managing power generated in a computer system_+++AEF_b+++.docx
2026-02-20 00:16
2007-08-28_发明授权_US07263457B2 System and method for operating components of an integrated circuit at independent frequencies and_or voltages_+++A_L_M_N_b_e_f_j_k+++.docx
2026-02-20 00:16
2008-01-03_发明申请_US20080005591A1 Method, system, and apparatus for dynamic thermal management_+++A_E_F_H_I_J_K_P_b_c_d_g_l_m_n_o+++.docx
2026-02-20 00:16
2009-04-09_发明申请_US20090094438A1 OVER-PROVISIONED MULTICORE PROCESSOR_+++A_L_M_O_P_b_e_f_j_k+++.docx
2026-02-20 00:16
2010-02-18_发明专利_JP2010039802A Multiprocessor system, scheduling method and program therefor_+++A_E_K_L_M_N_b_c_d_g_j_p+++.docx
2026-02-20 00:16
2011-06-09_发明申请_US20110138395A1 THERMAL MANAGEMENT IN MULTI-CORE PROCESSOR_+++A_E_F_b_c_d_e_g_h_i_j_k_l_p+++.docx
2026-02-20 00:16
2011-07-21_发明专利_JP2011141672A Information processor and method for controlling the same_+++A_B_E_F_I_c_d_g_h_j_k_l_p+++.docx
2026-02-20 00:16
2012-10-04_发明专利_DE112010004717T5 WÄRMEMANAGEMENT IN MEHRKERNPROZESSOR_+++A_F_I_P_b_c_d_e_g_h_j_k+++.docx
2026-02-20 00:16
2012-10-25_发明申请_US20120271481A1 METHOD AND SYSTEM FOR THERMAL LOAD MANAGEMENT IN A PORTABLE COMPUTING DEVICE_+++A_B_E_F_M_N_O_P_c_d_g_h_j_k_l+++.docx
2026-02-20 00:16
2012-10-26_发明申请_WO2012145212A2 METHOD AND SYSTEM FOR THERMAL LOAD MANAGEMENT IN A PORTABLE COMPUTING DEVICE_+++A_H_J_K_L_M_N_P_b_c_d_e_f_g_i_o+++.docx
2026-02-20 00:16
2012-10-26_发明申请_WO2012145212A3 THERMAL LOAD MANAGEMENT IN A PORTABLE COMPUTING DEVICE_+++A_B_J_K_L_M_N_O_P_c_d_e_f_g_h+++.docx
2026-02-20 00:16
2013-04-18_发明专利_JP2013513169A Thermal management in a multi-core processor_+++A_E_I_P_b_c_d_f_j_k+++.docx
2026-02-20 00:16
2013-06-20_发明申请_US20130159742A1 Method, System, and Apparatus for Dynamic Thermal Management_+++F_K_P_a_b_e_h_j_l_m_n+++.docx
2026-02-20 00:16
2014-01-01_发明公开_CN103492974A 用于通过监视和控制便携式计算装置中的电流流动来减少热负载的方法和系统_+++A_L_M_N_b_e_f_j_k_o_p+++.docx
2026-02-20 00:16
2014-01-07_发明公开_KR1020140002072A 휴대용 컴퓨팅 디바이스에서의 열 로드 관리_+++A_E_H_K_L_M_N_O_P_b_c_d_f_g_i_j+++.docx
2026-02-20 00:16
2014-02-26_发明公开_EP2699977A2 THERMAL LOAD MANAGEMENT IN A PORTABLE COMPUTING DEVICE_+++A_J_K_L_M_N_O_P_b_c_d_e_f_g_h+++.docx
2026-02-20 00:16
2014-03-27_发明申请_WO2011072001A3 THERMAL MANAGEMENT IN MULTI-CORE PROCESSOR_+++A_B_C_D_F_H_J_K_L_M_N_O_P_e_g+++.docx
2026-02-20 00:16
2014-04-17_发明申请_US20140108834A1 METHOD, SYSTEM, AND APPARATUS FOR DYNAMIC THERMAL MANAGEMENT_+++A_E_F_b_g_h_j_k_l+++.docx
2026-02-20 00:16
2014-06-19_发明申请_US20140173311A1 Methods and Systems for Operating Multi-Core Processors_+++A_L_M_N_O_P_b_e_j_k+++.docx
2026-02-20 00:16
2014-07-10_发明专利_JP2014516443A Management of the heat load in the portable computing device_+++A_E_J_K_L_M_N_O_P_b_c_d_f_g_h_i+++.docx
2026-02-20 00:16
2014-07-10_发明专利_JP2014516446A Method and system for reducing the thermal load by monitoring and controlling the current in the portable computing device_+++L_M_N_a_b_e_j_k_p+++.docx
2026-02-20 00:16
2014-07-23_发明专利_JP5559891B2 Thermal management in a multi-core processor_+++A_E_F_P_b_c_d_g_h_i_j_k_l+++.docx
2026-02-20 00:16
2012-10-25_发明申请_US20120271481A1 METHOD AND SYSTEM FOR THERMAL LOAD MANAGEMENT IN A PORTABLE COMPUTING DEVICE_+++A_B_E_F_M_N_O_P_c_d_g_h_j_k_l+++.docx

对比文件名称:2012-10-25_发明申请_US20120271481A1 METHOD AND SYSTEM FOR THERMAL LOAD MANAGEMENT IN A PORTABLE COMPUTING DEVICE

目标专利名称:166多核处理器的热缓解CN107111518B

模型名称:DeepSeek-R1

### 特征比对表格

技术特征描述及公开性判断结果对比文件原文引用公开性论述
**技术特征A**:所述UE具有包括多个核的处理器,所述多个核包括第一核和剩余核,所述方法包括:确定所述多个核中的所述第一核的温度,所述第一核处理负载<br>**《直接公开》**[0066] “The applications CPU 110 may comprise a multi-core processor that includes a zeroth core 222, a first core 224, and an Nth core 230.”<br>[0063] “The thermal sensors 157 may produce a voltage drop that is converted to digital signals… The thermal sensors 157, in addition to being controlled and monitored by an ADC controller 103, may also be controlled and monitored by one or more thermal policy manager module(s) 101.”<br>[0111] “Based on temperature readings from sensors 157, drivers executed on one or more of the cores themselves may be leveraged to cause a process load reallocation from a ‘hot’ core to a ‘cool,’ or otherwise less utilized, core.”对比文件直接公开了便携式计算设备(PCD)具有包含多个核(如core 222, 224, 230)的处理器(CPU 110)。温度传感器(157)被放置在各核附近以监测温度,热策略管理器模块(101)负责监视这些温度读数。当一个核处理负载并变得“热”时,会触发负载重新分配。这构成了具有多核处理器的UE,并确定处理负载的核的温度。因此,技术特征A被对比文件直接公开。
**技术特征B**:响应于确定所述第一核的温度大于预缓解温度阈值且不大于缓解温度阈值而确定所述剩余核中的第二核的温度,所述缓解温度阈值大于所述预缓解温度阈值<br>**《直接公开》**[0100-0102] “The second policy state 310 may comprise a ‘quality of service’ or ‘QoS’ state… This exemplary second state 310 may be reached or entered into by the thermal policy manager 101 when a change of temperature has been detected… The temperature range for this second, QoS thermal state 310 may comprise a range between about 50° C. to about 80° C.”<br>[0102-0103] “The third thermal state 315 may comprise a ‘severe’ state… The temperature range for this third, severe thermal state 310 may comprise a range between about 80° C. to about 100° C.”<br>[0120] “In block 720, the thermal policy manager 101 may reallocate or issue commands to reallocate the current workloads among the various cores, in order to reduce workload or to shift the workload.”对比文件公开了多个热状态(如QoS状态310、严重状态315),每个状态对应不同的温度范围(例如50-80°C, 80-100°C)。这实质上定义了多个温度阈值。当温度进入QoS状态(可对应于“大于预缓解阈值且不大于缓解阈值”)时,热策略管理器会采取缓解措施,包括审查各核的工作负载(隐含了确定其他核的状态)以进行负载重新分配。虽然阈值名称和具体数值不同,但本领域技术人员能毫无疑义地得出:当检测到核的温度处于两个阈值之间(即一个中间热状态)时,系统会响应并确定其他核的温度以进行后续操作。因此,技术特征B被直接公开。
**技术特征C**:响应于确定所述第二核的温度大于负载共享温度阈值而将所述第一核的所述负载的至少一部分、但非所有所述负载转移到所述第二核,所述负载共享温度阈值小于所述预缓解温度阈值<br>**《隐含公开》**[0113-0114] “Notably, in some embodiments… process loads may be reallocated within a given core. For example, process loads… may normally be scheduled for processing at a sub-core level… An overloaded process queue… may generate excessive thermal energy… In such a scenario, the thermal energy load may be mitigated by reallocating within the given core… all or part of the process load from the high density sub-processor block to the lower power density main process block.”<br>[0120] “The proportion of processing load reallocation, the particular portion of process load which is reallocated and the processing location to which load is reallocated, may be accomplished according to the current thermal state determined by the thermal policy manager 101.”对比文件公开了可以根据热状态重新分配处理负载,并且可以重新分配“all or part of the process load”(全部或部分负载)。虽然未明确记载一个名为“负载共享温度阈值”的第三阈值,但其公开了在QoS状态(对应预缓解状态)下采取负载重新分配等措施。本领域技术人员阅读对比文件后,为了更精细地管理热缓解策略,完全有可能且合理地设置一个比QoS状态触发温度(预缓解阈值)更低的阈值(即负载共享温度阈值),用于决策是转移全部负载还是部分负载。当目标核(第二核)温度高于此更低阈值时,可能意味着其也有一定热负荷,因此只转移部分负载而非全部,这是合乎逻辑的推断。因此,技术特征C被对比文件隐含公开。
**技术特征D**:以及响应于确定所述第二核的温度小于所述负载共享温度阈值而将所述第一核的所有所述负载转移到所述第二核。<br>**《隐含公开》**[0113-0114] “…reallocating within the given core… all or part of the process load…”<br>[0148] “In this scenario, if either or both of the zeroth core 222 and first core 224 reach the second thermal state 310… then the thermal policy manager 101 may apply… a process reallocation thermal load mitigation technique be applied in which all of the workload of the two active cores 222, 224 be shifted to the two inactive cores 226, 228.”对比文件明确公开了可以重新分配“全部”负载(all of the workload)。结合对技术特征C的分析,本领域技术人员可以合理推断:当用于接收负载的核(第二核)温度足够低(例如低于所设置的负载共享温度阈值)时,将热核的全部负载转移到该冷核是更直接有效的降温方式。对比文件图13A的示例(110C到110D)也展示了将所有负载从某些核转移到另一些核的场景。因此,基于对比文件公开的负载全部转移的可能性,以及设置温度阈值来触发不同转移策略的普遍技术手段,技术特征D被隐含公开。
**技术特征E**:进一步包括确定所述剩余核中的每一个核的温度,其中基于所述剩余核中的每一个核的温度的所述确定,所述第一核的所述负载的所述至少一部分被转移到所述第二核。<br>**《直接公开》**[0120] “In block 720, the thermal policy manager 101 may reallocate or issue commands to reallocate the current workloads among the various cores, in order to reduce workload or to shift the workload. The proportion of processing load reallocation… may be accomplished according to the current thermal state determined by the thermal policy manager 101.”<br>[0111] “Based on temperature readings from sensors 157, drivers executed on one or more of the cores themselves may be leveraged to cause a process load reallocation from a ‘hot’ core to a ‘cool,’ or otherwise less utilized, core.”对比文件公开了热策略管理器基于温度读数来触发负载重新分配,从“热”核到“冷”或未充分利用的核。要进行这样的决策,系统必然需要确定或知晓各个核的温度状态,以识别哪个是“热”核,哪个是“冷”核。因此,“确定剩余核中的每一个核的温度”是负载重新分配决策过程的必要前提和隐含步骤。本领域技术人员能够毫无疑义地从对比文件得出这一技术特征。因此,技术特征E被直接公开。
**技术特征F**:进一步包括确定所述剩余核中的哪一个核具有最低温度,其中响应于确定所述第二核具有所述剩余核的所述最低温度,所述负载的所述至少一部分被转移到所述第二核。<br>**《直接公开》**[0111] “Based on temperature readings from sensors 157, drivers executed on one or more of the cores themselves may be leveraged to cause a process load reallocation from a ‘hot’ core to a ‘cool,’ or otherwise less utilized, core.”<br>[0046] “…steer the processing load associated with the thermal load to underutilized, lower temperature or otherwise available processing capacity.”对比文件明确教导了将处理负载从热核重新分配到“较低温度(lower temperature)”或未充分利用的核。为了找到“较低温度”的核,系统需要比较各核的温度,并可以自然地选择其中温度最低的核作为转移目标。将负载转移到温度最低的核是实现有效热缓解的最直接和合理的选择。因此,本领域技术人员能够毫无疑义地从对比文件公开的“转移到较冷核”这一教导中,得出“确定哪个核温度最低并转移到该核”的具体实施方案。技术特征F被直接公开。
**技术特征G**:进一步包括:确定所述剩余核中的每一个核具有大于所述负载共享温度阈值的温度<br>**《隐含公开》**[0114-0115] “For the third or severe terminal state 315… the thermal policy manager 101 may apply or request that the monitor module 114 and/or O/S module 207 apply more aggressive thermal mitigation techniques… the thermal policy manager 101 may also shift workloads among different hardware devices in a spatial manner, to bring active devices off-line and to bring inactive devices on-line.”<br>[0147] “If the thermal policy manager 101 enters any one of the thermal states 310, 315, 320 described above… a process reallocation thermal load mitigation technique… may be implemented. According to this process reallocation… the thermal policy manager 101, the monitor module 114, and/or the O/S module 207 may shift the process workload of one core to one or more other cores in a multi-core processor 110.”对比文件公开了在更严重的热状态(如严重状态315)下,可以在多个核之间转移和共享负载(如图13A中110A到110B的示例)。当系统决定在多个核(而不仅是单个最冷核)之间共享负载时,这通常意味着可用的、适合接收负载的核可能都处于相对较高的温度水平(否则会优先选择单个最冷核)。本领域技术人员可以合理推断,在这种情况下,系统会确认这些候选核的温度状态。虽然对比文件没有明确记载“每一个核的温度都大于某个负载共享阈值”,但设置一个阈值来判断是否启用多核共享模式(而非单核全转),是热管理领域的常规设计选择。因此,技术特征G被隐含公开。
**技术特征H**:以及将所述负载的剩余部分转移到所述剩余核中的一组核以在所述第二核与所述一组核之间共享所述负载。<br>**《隐含公开》**[0147] “…the thermal policy manager 101, the monitor module 114, and/or the O/S module 207 may shift the process workload of one core to one or more other cores in a multi-core processor 110.”<br>[0148] 图13A中多核处理器110B的示例展示了将zeroth core和second core的部分负载重新分配到first core和third core,实现了在多个核(first core和third core)之间共享原属于一个核(zeroth core)的负载。对比文件图13A的示例(110A到110B)直接公开了将一个核(zeroth core 222)的负载的一部分转移到另一个核(first core 224),同时将另一个核(second core 226)的负载的一部分也转移到其他核(third core 228)。这实质上展示了在多个核之间重新分配负载。虽然该示例涉及多个源核,但其清楚表明了“将负载转移到一组核”的概念。本领域技术人员可以很容易地将此概念应用于单热核场景,即将热核的负载拆分并转移到“一组”较冷的核上。因此,技术特征H被隐含公开。
**技术特征I**:进一步包括确定所述剩余核中的每一个核的相应温度与所述第一核的温度的相应温差,其中基于所述剩余核的所述相应温差,所述第一核的所述负载在所述剩余核之间共享。<br>**《未公开》**对比文件中未发现记载根据核之间的具体温差(temperature difference)比例来分配负载的内容。对比文件公开了基于温度读数将负载从热核转移到冷核,也公开了可以在多个核之间分配负载。然而,对比文件没有公开或教导根据“第一核与每个剩余核之间的温差”来计算或决定每个剩余核应接收的负载比例。负载分配可以基于各核的绝对温度、利用率或热状态,但“基于温差”是一种特定的分配算法。对比文件未提供任何关于使用温差作为分配依据的启示或描述。因此,本领域技术人员不能从对比文件中毫无疑义地得出或合理推断出该技术特征。技术特征I未被公开。
**技术特征J**:其中进一步响应于确定所述第一核的温度不大于所述缓解温度阈值,所述负载的所述至少一部分被转移,所述方法进一步包括:确定所述多个核中的所述第一核的第二温度<br>**《隐含公开》**[0105] “The fourth thermal state 320 may comprise a ‘critical’ state in which the thermal policy manager 101 applies or triggers the shutting down of non-essential hardware and/or software. The temperature range for this fourth thermal state may include those of about 100° C. and above.”<br>[0116-0117] “Referring now to the fourth and critical state 320 of FIG. 9, the thermal policy manager 101 may start shutting down or requesting the monitor 114 and/or O/S module 207 to start shutting down all nonessential hardware and/or software modules.”对比文件公开了当温度达到临界状态(如约100°C以上)时,会采取更严厉的措施如关闭非必要硬件/软件。这对应于目标专利中“第一核温度大于缓解温度阈值”时采取的降低功耗等措施(见特征L)。而特征J描述的是在“不大于缓解温度阈值”时进行负载转移的上下文下,还包括后续再次确定温度(第二温度)的步骤。这是持续温度监控的自然延伸。本领域技术人员能够理解,热管理是一个持续过程,在采取任何缓解措施(如负载转移)后,系统会继续监测核的温度(即“第二温度”)以评估措施效果或判断是否需要进一步行动。因此,技术特征J被隐含公开。
**技术特征K**:确定所述第一核的第二温度大于所述缓解温度阈值<br>**《隐含公开》**[0105] “The fourth thermal state 320 may comprise a ‘critical’ state… The temperature range for this fourth thermal state may include those of about 100° C. and above.”<br>[0123] “For the fourth or critical terminal state 320, in block 720, this thermal state 320 may be similar to conventional techniques that are designed to eliminate functionality and operation of a PCD 100 in order to avoid critical temperatures.”对比文件明确定义了“临界状态”(critical state)及其温度范围(例如约100°C及以上)。这实质上公开了一个高温度阈值(缓解温度阈值)。在持续的热管理过程中,如果监测到核的温度(第二温度)达到或超过此临界阈值,系统就会确定该事件。虽然对比文件未使用“第二温度”这一表述,但持续监测并判断温度是否进入更严重状态(如从QoS状态进入严重或临界状态)是其热策略管理器的核心功能。因此,本领域技术人员能够合理推断出“确定第一核的第二温度大于缓解温度阈值”这一步骤。技术特征K被隐含公开。
**技术特征L**:以及响应于确定所述第一核的第二温度大于所述缓解温度阈值而降低所述第一核的功耗。<br>**《隐含公开》**[0105] “The fourth thermal state 320… the thermal policy manager 101 applies or triggers the shutting down of non-essential hardware and/or software.”<br>[0116-0117] “…the thermal policy manager 101 may start shutting down or requesting the monitor 114 and/or O/S module 207 to start shutting down all nonessential hardware and/or software modules.”<br>[0110] “Load scaling may comprise adjusting or ‘scaling’ the maximum clock frequency allowed in DVFS algorithm… This thermal load mitigation technique may also involve adjusting the voltage…”对比文件公开了在临界热状态(对应于温度大于缓解阈值)下,会采取关闭非必要硬件/软件等严厉措施。关闭硬件或降低其性能(如通过DVFS调整频率和电压)是直接降低功耗的手段。虽然对比文件在临界状态的描述中更强调“关闭”,但其在之前描述的热缓解技术(如负载动态缩放)中明确包含了通过调整频率和电压来管理功耗和热量的方法。本领域技术人员可以合理推断,当核的温度达到临界高阈值时,除了负载转移,直接降低该核本身的功耗(例如通过降频、降压甚至关闭)是必然且首选的缓解措施之一。因此,技术特征L被隐含公开。
**技术特征M**:其中所述降低所述第一核的功耗包括执行以下至少一者:降低所述第一核的频率<br>**《直接公开》**[0110] “Load scaling may comprise adjusting or ‘scaling’ the maximum clock frequency allowed in DVFS algorithm… This thermal load mitigation technique may also involve adjusting the voltage to match the standard DVFS table used for a particular and unique PCD 100.”<br>[0095] “In the context of this document, the demand for processors that provide high performance and low power consumption has led to the use of dynamic voltage and frequency scaling (“DVFS”) in processor designs. DVFS enables trade-offs between power consumption and performance. Processors 110 and 126 (FIG. 6) may be designed to take advantage of DVFS by allowing the clock frequency of each processor to be adjusted with a corresponding adjustment in voltage.”对比文件明确且详细地公开了通过动态电压频率缩放(DVFS)来调整处理器的时钟频率,以此作为管理功耗和热量的热缓解技术。降低处理器核的频率是DVFS的核心操作之一。因此,技术特征M被对比文件直接公开。
**技术特征N**:降低所述第一核的供电电压<br>**《直接公开》**[0110] “Load scaling may comprise adjusting or ‘scaling’ the maximum clock frequency allowed in DVFS algorithm… This thermal load mitigation technique may also involve adjusting the voltage to match the standard DVFS table…”<br>[0095] “…allowing the clock frequency of each processor to be adjusted with a corresponding adjustment in voltage.”对比文件在描述DVFS和负载缩放技术时,明确提及了调整电压(adjusting the voltage)。降低供电电压是与降低频率协同进行的、用于降低功耗的直接手段。因此,技术特征N被对比文件直接公开。
**技术特征O**:使所述第一核功率塌陷<br>**《直接公开》**[0116-0117] “Referring now to the fourth and critical state 320 of FIG. 9, the thermal policy manager 101 may start shutting down or requesting the monitor 114 and/or O/S module 207 to start shutting down all nonessential hardware and/or software modules.”<br>[0114] “Further, the thermal policy manager 101 may request that active process loads are completely reallocated… effectively taking the high thermal energy generating sub-processor blocks offline.”对比文件公开了在临界热状态下可以关闭(shut down)硬件模块,以及通过完全重新分配负载使处理块离线(offline)。使一个核“功率塌陷”(power collapse)是实现关闭或使其离线的具体且常用的低功耗技术手段。本领域技术人员能够毫无疑义地将对比文件中的“关闭”或“离线”理解为包括“功率塌陷”这种实现方式。因此,技术特征O被直接公开。
**技术特征P**:以及将所述第一核的所有所述负载转移到所述剩余核中的至少一个核。<br>**《直接公开》**[0148] “…the thermal policy manager 101 may apply or request that a process reallocation thermal load mitigation technique be applied in which all of the workload of the two active cores 222, 224 be shifted to the two inactive cores 226, 228.”<br>[0113-0114] “…reallocating within the given core… all or part of the process load…”对比文件明确公开了将“全部”工作负载(all of the workload)从一个或多个核转移到其他核。图13A的示例(110C到110D)也清晰展示了这种全部转移的场景。因此,技术特征P被对比文件直接公开。

<<<A>>><<<B>>><<<c>>><<<d>>><<<E>>><<<F>>><<<g>>><<<h>>><<<j>>><<<k>>><<<l>>><<<M>>><<<N>>><<<O>>><<<P>>>

权利要求与技术特征
Powered by Django

网站备案号:渝ICP备2023012882号


重庆市非显而易见网络科技有限责任公司 A Anti NPE NPE