-
公开(公告)号:US11579934B2
公开(公告)日:2023-02-14
申请号:US17208928
申请日:2021-03-22
Applicant: Apple Inc.
Inventor: Jeremy C. Andrus , John G. Dorsey , James M. Magee , Daniel A. Chimene , Cyril de la Cropte de Chanterac , Bryan R. Hinch , Aditya Venkataraman , Andrei Dorofeev , Nigel R. Gamble , Russell A. Blaine , Constantin Pistol , James S. Ismail
IPC: G06F9/50 , G06F9/48 , G06F1/3234 , G06F1/329 , G06F1/3296 , G06F9/38 , G06F9/26 , G06F9/54 , G06F1/20 , G06F1/324 , G06F1/3206 , G06F9/30
Abstract: Systems and methods are disclosed for scheduling threads on a processor that has at least two different core types, such as an asymmetric multiprocessing system. Each core type can run at a plurality of selectable voltage and frequency scaling (DVFS) states. Threads from a plurality of processes can be grouped into thread groups. Execution metrics are accumulated for threads of a thread group and fed into a plurality of tunable controllers for the thread group. A closed loop performance control (CLPC) system determines a control effort for the thread group and maps the control effort to a recommended core type and DVFS state. A closed loop thermal and power management system can limit the control effort determined by the CLPC for a thread group, and limit the power, core type, and DVFS states for the system. Deferred interrupts can be used to increase performance.
-
公开(公告)号:US09128721B2
公开(公告)日:2015-09-08
申请号:US13913307
申请日:2013-06-07
Applicant: Apple Inc.
Inventor: John G. Dorsey , James S. Ismail , Keith Cox , Gaurav Kapoor
CPC classification number: G09G5/003 , G06F1/20 , G06F1/26 , G06F1/324 , G06F1/3296 , G06T1/20 , G06T1/60 , G06T13/80 , G06T2200/28 , G09G5/18 , G09G2354/00 , G09G2360/08 , G09G2360/127 , Y02D10/126 , Y02D10/172
Abstract: The invention provides a technique for targeted scaling of the voltage and/or frequency of a processor included in a computing device. One embodiment involves scaling the voltage/frequency of the processor based on the number of frames per second being input to a frame buffer in order to reduce or eliminate choppiness in animations shown on a display of the computing device. Another embodiment of the invention involves scaling the voltage/frequency of the processor based on a utilization rate of the GPU in order to reduce or eliminate any bottleneck caused by slow issuance of instructions from the CPU to the GPU. Yet another embodiment of the invention involves scaling the voltage/frequency of the CPU based on specific types of instructions being executed by the CPU. Further embodiments include scaling the voltage and/or frequency of a CPU when the CPU executes workloads that have characteristics of traditional desktop/laptop computer applications.
Abstract translation: 本发明提供了一种用于针对包括在计算设备中的处理器的电压和/或频率进行目标缩放的技术。 一个实施例涉及基于每秒输入帧缓冲器的帧数来缩放处理器的电压/频率,以便减少或消除在计算设备的显示器上显示的动画中的笨拙。 本发明的另一实施例涉及基于GPU的利用率来缩放处理器的电压/频率,以便减少或消除由CPU向GPU缓慢发出指令所引起的任何瓶颈。 本发明的另一个实施例涉及根据由CPU执行的特定类型的指令来调整CPU的电压/频率。 另外的实施例包括在CPU执行具有传统台式/膝上型计算机应用的特征的工作负载时缩放CPU的电压和/或频率。
-
公开(公告)号:US11360820B2
公开(公告)日:2022-06-14
申请号:US15996469
申请日:2018-06-02
Applicant: Apple Inc.
Inventor: John G. Dorsey , Daniel A. Chimene , Andrei Dorofeev , Bryan R. Hinch , Evan M. Hoke , Aditya Venkataraman
IPC: G06F9/50 , G06F9/48 , G06F1/3234 , G06F1/329 , G06F1/3296 , G06F9/38 , G06F9/26 , G06F9/54 , G06F1/20 , G06F1/324 , G06F1/3206 , G06F9/30
Abstract: Systems and methods are disclosed for scheduling threads on an asymmetric multiprocessing system having multiple core types. Each core type can run at a plurality of selectable voltage and frequency scaling (DVFS) states. Threads from a plurality of processes can be grouped into thread groups. Execution metrics are accumulated for threads of a thread group and fed into a plurality of tunable controllers. A closed loop performance control (CLPC) system determines a control effort for the thread group and maps the control effort to a recommended core type and DVFS state. A closed loop thermal and power management system can limit the control effort determined by the CLPC for a thread group, and limit the power, core type, and DVFS states for the system. Metrics for workloads offloaded to co-processors can be tracked and integrated into metrics for the offloading thread group.
-
公开(公告)号:US11119788B2
公开(公告)日:2021-09-14
申请号:US16376785
申请日:2019-04-05
Applicant: Apple Inc.
Inventor: Aditya Venkataraman , Bryan R. Hinch , John G. Dorsey
IPC: G06F1/00 , G06F9/38 , G06F1/3228 , G06F1/3296 , G06F9/48 , G06F9/50
Abstract: Closed loop performance controllers of asymmetric multiprocessor systems may be configured and operated to improve performance and power efficiency of such systems by adjusting control effort parameters that determine the dynamic voltage and frequency state of the processors and coprocessors of the system in response to the workload. One example of such an arrangement includes applying hysteresis to the control effort parameter and/or seeding the control effort parameter so that the processor or coprocessor receives a returning workload in a higher performance state. Another example of such an arrangement includes deadline driven control, in which the control effort parameter for one or more processing agents may be increased in response to deadlines not being met for a workload and/or decreased in response to deadlines being met too far in advance. The performance increase/decrease may be determined by comparison of various performance metrics for each of the processing agents.
-
公开(公告)号:US20210247985A1
公开(公告)日:2021-08-12
申请号:US17244377
申请日:2021-04-29
Applicant: Apple Inc.
Inventor: Aditya Venkataraman , Bryan R. Hinch , John G. Dorsey
IPC: G06F9/38 , G06F9/48 , G06F1/3296 , G06F1/3228 , G06F9/50
Abstract: Closed loop performance controllers of asymmetric multiprocessor systems may be configured and operated to improve performance and power efficiency of such systems by adjusting control effort parameters that determine the dynamic voltage and frequency state of the processors and coprocessors of the system in response to the workload. One example of such an arrangement includes applying hysteresis to the control effort parameter and/or seeding the control effort parameter so that the processor or coprocessor receives a returning workload in a higher performance state. Another example of such an arrangement includes deadline driven control, in which the control effort parameter for one or more processing agents may be increased in response to deadlines not being met for a workload and/or decreased in response to deadlines being met too far in advance. The performance increase/decrease may be determined by comparison of various performance metrics for each of the processing agents.
-
公开(公告)号:US11080095B2
公开(公告)日:2021-08-03
申请号:US15870766
申请日:2018-01-12
Applicant: Apple Inc.
Inventor: Jeremy C. Andrus , John G. Dorsey , James M. Magee , Daniel A. Chimene , Cyril de la Cropte de Chanterac , Bryan R. Hinch , Aditya Venkataraman , Andrei Dorofeev , Nigel R. Gamble , Russell A. Blaine , Constantin Pistol
IPC: G06F9/46 , G06F9/50 , G06F9/48 , G06F1/3234 , G06F1/329 , G06F1/3296 , G06F9/38 , G06F9/26 , G06F9/54 , G06F1/20 , G06F1/324 , G06F1/3206 , G06F9/30
Abstract: Systems and methods are disclosed for scheduling threads on a processor that has at least two different core types, such as an asymmetric multiprocessing system. Each core type can run at a plurality of selectable voltage and frequency scaling (DVFS) states. Threads from a plurality of processes can be grouped into thread groups. Execution metrics are accumulated for threads of a thread group and fed into a plurality of tunable controllers for the thread group. A closed loop performance control (CLPC) system determines a control effort for the thread group and maps the control effort to a recommended core type and DVFS state. A closed loop thermal and power management system can limit the control effort determined by the CLPC for a thread group, and limit the power, core type, and DVFS states for the system. Deferred interrupts can be used to increase performance.
-
7.
公开(公告)号:US20200073714A1
公开(公告)日:2020-03-05
申请号:US16376828
申请日:2019-04-05
Applicant: Apple Inc.
Inventor: Aditya Venkataraman , Bryan R. Hinch , John G. Dorsey
Abstract: Closed loop performance controllers of asymmetric multiprocessor systems may be configured and operated to improve performance and power efficiency of such systems by adjusting control effort parameters that determine the dynamic voltage and frequency state of the processors and coprocessors of the system in response to the workload. One example of such an arrangement includes applying hysteresis to the control effort parameter and/or seeding the control effort parameter so that the processor or coprocessor receives a returning workload in a higher performance state. Another example of such an arrangement includes deadline driven control, in which the control effort parameter for one or more processing agents may be increased in response to deadlines not being met for a workload and/or decreased in response to deadlines being met too far in advance. The performance increase/decrease may be determined by comparison of various performance metrics for each of the processing agents.
-
公开(公告)号:US09672310B1
公开(公告)日:2017-06-06
申请号:US14795013
申请日:2015-07-09
Applicant: Apple Inc.
Inventor: Antonietta Oliva , John G. Dorsey , Keith Cox , Norman J. Rohrer , Sribalan Santhanam , Sung Wook Kang , Mohamed H. Abu-Rama , Ashish R. Jain
CPC classification number: G01R31/2894 , G01R31/2856 , G01R31/2875 , G01R31/2879
Abstract: In an embodiment, the amount of supply voltage guardband to prevent incorrect operation due to aging effects may be modeled using an IC-specific age model generated early in the product life cycle of the IC. For example, high temperature operating life (HTOL) testing may be performed at multiple temperatures and/or voltages to develop the IC-specific age model. The IC-specific age model may be more accurate then the calculations used to develop guardband voltage as discussed previously, which rely on the aging of a single transistor. The IC-specific age model may be used along with monitoring of the aging effects during operation of the IC to predict an amount of increased guardband voltage that is currently desirable to apply to the IC. The predicted amount may vary from about zero when the IC is new to the full amount of guardband voltage when the IC is nearing end of life.
-
公开(公告)号:US20210349726A1
公开(公告)日:2021-11-11
申请号:US17384399
申请日:2021-07-23
Applicant: Apple Inc.
Inventor: Aditya Venkataraman , Bryan R. Hinch , John G. Dorsey
IPC: G06F9/38 , G06F1/3228 , G06F1/3296 , G06F9/48 , G06F9/50
Abstract: Closed loop performance controllers of asymmetric multiprocessor systems may be configured and operated to improve performance and power efficiency of such systems by adjusting control effort parameters that determine the dynamic voltage and frequency state of the processors and coprocessors of the system in response to the workload. One example of such an arrangement includes applying hysteresis to the control effort parameter and/or seeding the control effort parameter so that the processor or coprocessor receives a returning workload in a higher performance state. Another example of such an arrangement includes deadline driven control, in which the control effort parameter for one or more processing agents may be increased in response to deadlines not being met for a workload and/or decreased in response to deadlines being met too far in advance. The performance increase/decrease may be determined by comparison of various performance metrics for each of the processing agents.
-
公开(公告)号:US20210318909A1
公开(公告)日:2021-10-14
申请号:US17208928
申请日:2021-03-22
Applicant: Apple Inc.
Inventor: Jeremy C. Andrus , John G. Dorsey , James M. Magee , Daniel A. Chimene , Cyril de la Cropte de Chanterac , Bryan R. Hinch , Aditya Venkataraman , Andrei Dorofeev , Nigel R. Gamble , Russell A. Blaine , Constantin Pistol , James S. Ismail
IPC: G06F9/50 , G06F9/48 , G06F1/3234 , G06F1/329 , G06F1/3296 , G06F9/38 , G06F9/26 , G06F9/54 , G06F1/20 , G06F1/324
Abstract: Systems and methods are disclosed for scheduling threads on a processor that has at least two different core types, such as an asymmetric multiprocessing system. Each core type can run at a plurality of selectable voltage and frequency scaling (DVFS) states. Threads from a plurality of processes can be grouped into thread groups. Execution metrics are accumulated for threads of a thread group and fed into a plurality of tunable controllers for the thread group. A closed loop performance control (CLPC) system determines a control effort for the thread group and maps the control effort to a recommended core type and DVFS state. A closed loop thermal and power management system can limit the control effort determined by the CLPC for a thread group, and limit the power, core type, and DVFS states for the system. Deferred interrupts can be used to increase performance.
-
-
-
-
-
-
-
-
-