返回项目页面
对比文件列表
2002-11-29_发明专利_JP2002345147A Motor controller_+++D_E_b_i_l+++.docx
2026-02-23 21:21
2003-01-15_发明专利_JP2003014552A Temperature detector_+++A_D_E_F_H_b_g_i_l+++.docx
2026-02-23 21:21
2003-05-09_发明专利_JP2003134795A Fault sensing system_+++A_D_E_G_H_I_J_b_l+++.docx
2026-02-23 21:21
2005-09-29_发明专利_JP2005269832A Temperature-detecting device and program for temperature detection_+++A_B_D_E_G_I_L_f_h+++.docx
2026-02-23 21:21
2006-12-16_发明专利_TW200643391A 分區之熱監視 ZONED THERMAL MONITORING_+++A_E_F_b_c_d_g_h_i_j_k_l+++.docx
2026-02-23 21:21
2007-02-01_发明专利_JP2007028887A Motor temperature control using estimated motor temperature based on motor power dissipation_+++B_D_E_G_H_a_f_i_l+++.docx
2026-02-23 21:21
2008-06-26_发明申请_US20080155281A1 Method and apparatus to estimate energy consumed by central processing unit core_+++D_j_k_l+++.docx
2026-02-23 21:21
2010-03-18_发明申请_WO2010029780A1 インバータ装置、インバータコントロールシステム、モータ制御システム、インバータ装置の制御方法_+++A_I_d_e_l+++.docx
2026-02-23 21:21
2011-02-17_发明专利_JP2011036095A Power converter and protection method of power converter_+++A_D_E_F_I_b_g_h_l+++.docx
2026-02-23 21:21
2011-05-12_发明专利_JP2011097812A Inverter device_+++A_B_D_E_G_I_L_h+++.docx
2026-02-23 21:21
2011-06-02_发明申请_US20110131004A1 DIE TEMPERATURE ESTIMATOR_+++B_E_J_d_g_l+++.docx
2026-02-23 21:21
2012-02-02_发明专利_JPWO2010029780A1 インバータ装置、インバータコントロールシステム、モータ制御システム、インバータ装置の制御方法_+++A_E_F_b_d_g_h_i_l+++.docx
2026-02-23 21:21
2012-08-30_发明申请_US20120217795A1 INVERTER APPARATUS_+++A_B_D_E_F_I_J_g_h_l+++.docx
2026-02-23 21:21
2013-01-02_发明授权_CN101617314B 具有可配置的输入_输出的裸片设备及其控制方法_+++b_l+++.docx
2026-02-23 21:21
2014-05-07_发明公开_EP2728327A1 Portable electronic device_+++A_E_F_b_d_g_h_j_l+++.docx
2026-02-23 21:21
2014-05-08_发明申请_WO2014067022A1 PORTABLE ELECTRONIC DEVICE_+++A_F_b_d_e_i_j_l+++.docx
2026-02-23 21:21
2015-04-16_发明申请_US20150106642A1 PERFORMANCE STATE BOOST FOR MULTI-CORE INTEGRATED CIRCUIT_+++A_D_L_b_j+++.docx
2026-02-23 21:21
2015-07-01_发明公开_CN104755891A 便携式电子设备_+++A_B_D_E_F_J_g_h_i_l+++.docx
2026-02-23 21:21
2015-07-08_发明公开_KR1020150079652A 휴대용 전자 디바이스_+++A_F_b_d_e_g_h_j_l+++.docx
2026-02-23 21:21
2015-10-15_发明申请_US20150292959A1 PORTABLE ELECTRONIC DEVICE_+++A_B_D_E_F_I_J_c_g_h_k_l+++.docx
2026-02-23 21:21
2015-12-03_发明申请_US20150346798A1 SYSTEM AND METHOD FOR ADJUSTING PERFORMANCE BASED ON THERMAL CONDITIONS WITHIN A PROCESSOR_+++A_D_E_b_g_h_j_l+++.docx
2026-02-23 21:21
2016-03-31_发明申请_US20160092617A1 INTEGRATED CIRCUIT FOR ESTIMATING POWER OF AT LEAST ONE NODE USING TEMPERATURE AND A SYSTEM INCLUDING THE SAME_+++A_b_l+++.docx
2026-02-23 21:21
无说明书_+++A_B_D_E_F_G_J_K_L_c_h_i+++.docx
2026-02-23 21:21
2015-12-03_发明申请_US20150346798A1 SYSTEM AND METHOD FOR ADJUSTING PERFORMANCE BASED ON THERMAL CONDITIONS WITHIN A PROCESSOR_+++A_D_E_b_g_h_j_l+++.docx

# 特征比对分析

## 对比文件名称:

2015-12-03_发明申请_US20150346798A1 SYSTEM AND METHOD FOR ADJUSTING PERFORMANCE BASED ON THERMAL CONDITIONS WITHIN A PROCESSOR

## 目标专利名称:

时间温度传感器位置偏移误差校正CN109642829B

## 模型名称:

GPT-4

## 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
**A:其包括:从传感器接收区的温度测量值** <br>**《直接公开》**[0048] "Additionally, the processing block 210 may send sensor measured temperature values to the power management unit 240. These measured temperature values are shown as sensor input 232 in FIG. 2. In various embodiments, the die 202 may utilize analog or digital thermal sensors placed throughout the die 202. The thermal sensors provide information as to when the die heats up in a particular area due to increased compute activity." <br> [0061] "As described earlier, the processing block 210 may send sensor measured temperature values to the power management unit 240."对比文件明确公开了处理块(processing block)将传感器测量的温度值发送给功率管理单元(power management unit)。这里的“传感器”是遍布裸片的模拟或数字热传感器,用于感测特定区域(例如,因计算活动增加而发热的区域)的温度。这对应于目标专利中“从传感器接收区的温度测量值”,其作用是获取处理器特定区域的温度信息,用于后续的热管理决策,与目标专利中的作用(作为温度校正的输入)相同。因此,该特征被对比文件直接公开。
**B:测量所述区的活动,所测量的活动包括流动于包含所述裸片的装置中的电流或包含所述裸片的所述装置的性能** <br>**《隐含公开》**[0047] "In various embodiments, existing or additional logic may be added to the design for the purpose of sampling and/or maintaining counts (e.g., using flops) corresponding to these signals. These selected signals may be selected for sampling based upon a determination that they have a relatively significant impact on power consumption. For example, various clock enable signals, bus driver enables, mismatches in content-addressable memories (CAM), and CAM word-line (WL) drivers may be chosen for sampling. In one embodiment, a count may be maintained for such signals during operation (e.g., when a clock enable is asserted). Based on these counts, an estimate of power consumption corresponding to the counts may be determined." <br> [0046] "In various embodiments, one or more of the processing units 220a-220d may sample a number of pre-selected signals. The selection of which signals to sample during a particular clock cycle may correspond to how well the selection correlates to the amount of switching node capacitance within a given one of the processing units 220a-220d."对比文件详细描述了通过采样预选信号(如时钟使能、总线驱动器使能)并维护其计数来估计功耗。这些信号的活动(例如,断言状态)直接反映了包含裸片的装置(即处理器或处理单元)内部电路的开关活动。开关活动必然伴随着电流的流动(对电容充放电)和装置性能的变化(例如,指令执行速率)。因此,通过测量这些信号的活动来估计功耗,隐含了测量“流动于包含裸片的装置中的电流”或“包含裸片的装置的性能”。虽然对比文件未明确使用“电流”或“性能”这两个词,但本领域技术人员可以毫无疑义地理解,所采样的信号活动是电流和性能的直接表现。因此,该特征被对比文件隐含公开。
**C:所述性能至少部分基于确定高速缓存未命中的数量或管线暂停的数量** <br>**《未公开》**无对应内容。对比文件中虽然提及了通过采样特定信号(如CAM不匹配、CAM字线驱动器)来估计功耗,这些信号可能与高速缓存或管线活动相关,但并未明确提及或必然隐含“确定高速缓存未命中的数量”或“管线暂停的数量”作为性能测量的具体基础。目标专利中这两个是具体的性能计数器指标。对比文件公开的是一种更通用的、基于信号采样的功耗估计方法,并未教导或暗示使用这两个特定的性能事件计数器。因此,该特征未被对比文件公开。
**D:将所测量的活动转化为所述区的功率** <br>**《直接公开》**[0047] "Based on these counts, an estimate of power consumption corresponding to the counts may be determined." <br> [0050] "Control logic within the power reporting unit 242 receives activity levels sampled or measured during a given time interval. The power reporting unit 242 calculates a power value from the usage measurements 230 received from activity monitors in the processing units 220a-220d."对比文件明确公开了基于采样信号的活动计数(counts)或活动水平(activity levels)来确定功耗估计值(estimate of power consumption)或计算功率值(calculates a power value)。这正是“将所测量的活动转化为所述区的功率”。其作用是为后续的热模型计算提供输入,与目标专利中的作用(为导出温度偏移提供输入)实质相同。因此,该特征被对比文件直接公开。
**E:通过参考所存储的热量模型参数来解释所述区的所述功率或所述活动来导出温度偏移** <br>**《直接公开》**[0055] "The temperature reporting unit 246 receives the scaled calculated power value. The temperature reporting unit 246 translates the received value to a calculated die temperature value. The temperature reporting unit 246 may utilize a transient resistor-capacitance (RC) thermal model for the translation." <br> [0074] "In block 314, a transient RC thermal model is used to translate the scaled calculated power value, Pcalc-scaled, to a calculated temperature value, Tcalc."对比文件明确公开了温度报告单元(或相应方法步骤)使用瞬态电阻-电容(RC)热模型,将缩放后的计算功率值转换为计算出的裸片温度值。这里的“瞬态RC热模型”就是一组预定义的、存储的热量模型参数。通过参考这个模型来解释功率并“导出”一个计算温度值(Tcalc)。在目标专利的上下文中,“温度偏移”可以理解为校正量,而对比文件导出的“计算温度值”本身就可以视为一个基于模型的温度估计,可用于与传感器读数比较以产生偏移,或者其计算过程本身就隐含了偏移的确定。因此,该特征被对比文件直接公开。
**F:以及至少部分地基于所述温度偏移调节所述温度测量值。** <br>**《未公开》**[0058] "The operation parameter selector 248 receives the calculated temperature value or the boost value, and accordingly selects a set of operation parameters for the processing units 220a-220d based on the received value." <br> [0075] "In block 318, the operating parameters of one or more processing units are adjusted based on the temperature headroom and possibly on the allocation of power credits."对比文件公开的是基于计算出的温度值或温度余量(headroom)来调整处理器的操作参数(如P状态、电压/频率),即调节处理器性能以控制发热。其调节的对象是“操作参数”或“性能”。而目标专利权利要求中的技术特征F明确要求“调节所述温度测量值”,即对传感器读取的原始温度值本身进行校正(例如,加上一个偏移量以获得更准确的值)。对比文件没有公开对温度传感器测量值本身进行调节或校正。因此,该特征未被对比文件公开。
**G:其中所述所存储的热量模型参数包含与温度差匹配的所估计的功率。** <br>**《隐含公开》**[0055] "The temperature reporting unit 246 may utilize a transient resistor-capacitance (RC) thermal model for the translation." <br> [0057] "In some embodiments, the operation parameter selector 248 may use the received calculated temperature value to determine the boost value, which may be the difference between the maximum temperature limit and the calculated temperature."对比文件使用的瞬态RC热模型,其本质是建立功率输入与温度响应(包括温度差)之间的关系。模型参数(如热阻Rth、热容Cth)决定了在给定功率输入下,温度相对于某个参考点(如环境温度)的上升(即温度差)。虽然对比文件未明确陈述“模型参数包含与温度差匹配的所估计的功率”,但本领域技术人员公知,任何热模型的核心就是描述功率与温度(或温升)的映射关系。因此,该模型必然隐含了“所估计的功率”与“温度差”之间的匹配关系。此外,对比文件计算“提升值”(boost value)作为最大温度限值与计算温度值的差值,也体现了温度差的概念。因此,该特征被对比文件隐含公开。
**H:其中所述所存储的热量模型参数包含与绝对温度匹配的所估计的功率。** <br>**《隐含公开》**[0055] "The temperature reporting unit 246 receives the scaled calculated power value. The temperature reporting unit 246 translates the received value to a calculated die temperature value. The temperature reporting unit 246 may utilize a transient resistor-capacitance (RC) thermal model for the translation." <br> [0056] "The RC thermal model typically utilizes a worst-case ambient temperature."对比文件公开了使用RC热模型将功率值转换为计算出的裸片温度值(calculated die temperature value),并且该模型通常使用最坏情况环境温度。计算出的裸片温度值是一个绝对温度值(例如,98摄氏度)。因此,该热模型必然建立了“所估计的功率”与最终计算出的“绝对温度”之间的匹配关系,其中环境温度作为模型的边界条件或输入参数之一。虽然对比文件未使用“与绝对温度匹配”这一措辞,但根据其描述,本领域技术人员可以毫无疑义地推断出,所存储的热模型参数(Rth, Cth等)与环境温度条件相结合,能够将功率输入映射为绝对温度输出。因此,该特征被对比文件隐含公开。
**I:其中所述所测量的活动包括电流流动穿过电源开关。** <br>**《未公开》**无对应内容。对比文件提及采样“时钟使能信号、总线驱动器使能”等,这些信号控制电路单元的开关,但并未明确提及或必然指向“电源开关”(power switch)。电源开关通常指用于电源门控(power gating)的特定开关器件,其电流流动是专门用于控制电源域的开启和关闭。对比文件采样的是一般性影响功耗的信号,不特指电源开关的电流。因此,该特征未被对比文件公开。
**J:其中所述所测量的活动包括至少部分地基于所述裸片的所述装置的所述性能。** <br>**《隐含公开》**[0047] "These selected signals may be selected for sampling based upon a determination that they have a relatively significant impact on power consumption. For example, various clock enable signals, bus driver enables, mismatches in content-addressable memories (CAM), and CAM word-line (WL) drivers may be chosen for sampling."对比文件选择采样的信号(如时钟使能、CAM不匹配)与处理器的性能直接相关。时钟使能信号的活动频率反映了处理器核心或功能单元的活动水平,进而反映其性能水平(如指令吞吐量)。CAM不匹配等事件也与高速缓存访问性能有关。因此,通过测量这些信号的活动来估计功耗,其基础至少部分是基于裸片装置(即处理器)的性能。本领域技术人员能够理解,功耗与性能密切相关,测量影响功耗的信号活动间接反映了性能。因此,该特征被对比文件隐含公开。
**K:其中所述性能的测量值是至少部分地基于性能计数器确定的。** <br>**《未公开》**无对应内容。对比文件描述了通过采样特定硬件信号并维护其计数来估计功耗。虽然这涉及“计数”(counts),但这些计数是针对预选的、影响功耗的硬件信号(如时钟使能),而非专门用于衡量处理器性能的“性能计数器”(performance counters)。性能计数器通常指架构上定义的、用于统计特定性能事件(如指令退休数、缓存未命中数)的专用寄存器。对比文件未提及或暗示使用此类性能计数器。因此,该特征未被对比文件公开。
**L:其中所述活动的测量以及所述区所消耗的功率转换是通过数字功率监测器/计量器执行的。** <br>**《隐含公开》**[0047] "In various embodiments, existing or additional logic may be added to the design for the purpose of sampling and/or maintaining counts (e.g., using flops) corresponding to these signals. ... Based on these counts, an estimate of power consumption corresponding to the counts may be determined." <br> [0050] "Control logic within the power reporting unit 242 receives activity levels sampled or measured during a given time interval. The power reporting unit 242 calculates a power value from the usage measurements 230 received from activity monitors in the processing units 220a-220d."对比文件公开了在处理器中增加逻辑来采样信号并维护计数,以及功率报告单元基于这些活动测量值计算功率值。这套用于实时采样活动并计算功率的硬件逻辑(可能包含计数器、采样逻辑、计算单元)本质上构成了一个“数字功率监测器/计量器”。虽然对比文件未使用该特定术语,但其描述的功能(数字化采样活动并转换为功率估计)与本领域技术人员所理解的数字功率监测器/计量器完全相同。因此,该特征被对比文件隐含公开。

<<<A>>><<<D>>><<<E>>><<<b>>><<<g>>><<<h>>><<<j>>><<<l>>>

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