WO2015015756A1 - 不揮発性メモリ搭載サーバの省電力制御システム、制御装置、制御方法および制御プログラム - Google Patents
不揮発性メモリ搭載サーバの省電力制御システム、制御装置、制御方法および制御プログラム Download PDFInfo
- Publication number
- WO2015015756A1 WO2015015756A1 PCT/JP2014/003848 JP2014003848W WO2015015756A1 WO 2015015756 A1 WO2015015756 A1 WO 2015015756A1 JP 2014003848 W JP2014003848 W JP 2014003848W WO 2015015756 A1 WO2015015756 A1 WO 2015015756A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- memory
- power saving
- saving control
- cpu
- power
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3234—Power saving characterised by the action undertaken
- G06F1/3287—Power saving characterised by the action undertaken by switching off individual functional units in the computer system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3234—Power saving characterised by the action undertaken
- G06F1/329—Power saving characterised by the action undertaken by task scheduling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
- G06F9/4893—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues taking into account power or heat criteria
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5094—Allocation of resources, e.g. of the central processing unit [CPU] where the allocation takes into account power or heat criteria
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Definitions
- the present invention relates to a technology for enhancing the power saving effect in a server equipped with a non-volatile memory.
- Non-Patent Document 1 describes a computing technology that achieves normally-off, which actively shuts off power sources other than the components that should truly operate, even while operating as a system.
- Non-Patent Document 1 an increase in power consumption due to power OFF / ON of a CPU circuit or memory and a power saving amount due to reduction of leakage current during the power OFF period are offset. Therefore, in order to enhance the power saving effect, it is necessary to reduce the frequency of the power OFF / ON as much as possible and to prolong the continuous power OFF time. However, there is a limit in extending the time for continuous power off as long as power control is performed by devising the hardware layer and the OS layer.
- Patent Document 1 discloses a technology for reducing wasteful power consumption in a multi-core processor system although it is not normally-off computing.
- This is a task scheduling apparatus provided with a scheduler that turns off the power supply to the processor when it detects that there is an idle processor.
- the task scheduling apparatus measures the workload of each task before processing the task.
- the usage rates of multiple processors are predicted based on the measured amount of work. Therefore, it is not suitable for normally-off computing which has the complexity of performing these processes and requires instantaneous power-off / on.
- Patent Document 2 Patent Document 3
- Patent Document 4 Patent Document 5
- the present invention has been made in view of the above problems, and an object thereof is to realize normally-off computing with a power saving effect enhanced in a server equipped with a non-volatile memory.
- the power saving control system of the present invention has a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- the operating system comprises: a process characteristic collection unit collecting characteristics of the process; a core allocation determination unit determining an allocation of the CPU core to the process based on the characteristics of the process; and a process based on the allocation
- a process scheduler for performing the process of turning on / off the power of the CPU core or the memory based on the execution of the process.
- a power saving control device includes a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system unit operating on the physical machine, and one or more process units operating by the operating system unit.
- the operating system unit includes a process characteristic collection unit that collects the characteristics of the process unit, and a core allocation determination unit that determines the allocation of the CPU core to the process unit based on the characteristics of the process unit.
- a process scheduler that executes a process unit based on the assignment, and turns off / on the power of the CPU core or the memory based on the execution of the process unit.
- a power saving control method includes a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- the operating system collects characteristics of the process, determines assignment of the CPU core to the process based on characteristics of the process, and processes the process based on the assignment. And turn off / on the power of the CPU core or the memory based on the execution of the process.
- a power saving control program includes a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- a process of collecting characteristics of the process, a process of determining assignment of the CPU core to the process based on characteristics of the process, and assignment A process of executing the process based on the process and a process of powering off / on the CPU core or the memory based on the process execution are executed.
- the power saving control system of the present invention has a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- the operating system is a process characteristic collection unit that collects load characteristics of the process or interrupt characteristics of the process; and an interrupt aggregation unit that aggregates interrupts to the process based on the load characteristics or the interrupt characteristics.
- a process scheduler that executes an interrupt based on the concentration of the interrupts, and turns off / on the power of the CPU core or the memory based on the execution of the interrupt.
- a power saving control device comprises a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system unit operating on the physical machine, and one or more processing units operating on the operating system.
- the operating system unit includes a process characteristic collection unit that collects load characteristics of the process unit or interrupt characteristics of the process unit, and aggregation of interrupts to the process unit based on the load characteristics or the interrupt characteristics And a process scheduler that executes an interrupt based on the concentration of the interrupts, and powers off / on the CPU core or the memory based on the execution of the interrupts.
- a power saving control method includes a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- the operating system collects load characteristics of the process or interrupt characteristics of the process, and aggregates interrupts into the process based on the load characteristics or the interrupt characteristics. And executing an interrupt based on the concentration of the interrupts, and powering off / on the CPU core or the memory based on the execution of the interrupt.
- a power saving control program includes a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- a process of collecting load characteristics of the process or interrupt characteristics of the process in the operating system, and concentration of interrupts into the process based on the load characteristics or the interrupt characteristics Processing of executing an interrupt based on the aggregation of the interrupts, and processing of turning the power of the CPU core or the memory off / on based on the execution of the interrupts.
- the power saving control system of the present invention comprises a physical machine having a CPU and a memory, an operating system running on the physical machine, and one or more processes running on the operating system, the operating system comprising A process scheduler for scheduling the process; and a memory power control unit for turning on / off the power of the memory based on the scheduling.
- the power saving control device includes a physical machine having a CPU and a memory, an operating system unit operating on the physical machine, and one or more processing units operated by the operating system unit.
- the operating system unit has a process scheduler that schedules the process unit, and a memory power control unit that turns on / off the power of the memory based on the scheduling.
- a power saving control method saves a power saving control system having a physical machine having a CPU and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- the operating system schedules the process and powers on / off the memory based on the scheduling.
- a power saving control program saves a power saving control system having a physical machine having a CPU and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system.
- the operating system is caused to execute a process of scheduling the process and a process of powering on / off the memory based on the scheduling.
- FIG. 1 is a block diagram showing a configuration of a power saving control system of a non-volatile memory mounted server according to a first embodiment of this invention.
- the power saving control system of this embodiment includes a physical machine 1 having a CPU 11 having a plurality of CPU cores 18 and a memory 19, an operating system 14 operating on the physical machine 1, and one or more operating on the operating system 14. And a process 150.
- the operating system 14 has a process characteristic collection unit 141 that collects the characteristics of the process 150.
- the core allocation determination unit 142 determines the allocation of the CPU core 18 to the process 150 based on the characteristics of the process 150, and a process scheduler 143 that executes the process 150 based on the allocation. Furthermore, based on the execution of the process 150, the power supply of the CPU core 18 or the memory 19 is turned off / on.
- FIG. 2 is a block diagram showing a configuration of a power saving control system of a non-volatile memory mounted server according to a second embodiment of the present invention.
- the power saving control system of the present embodiment has a physical machine 1.
- the physical machine 1 has a CPU 11 having a plurality of CPU cores 18 (k in FIG. 2, k is a positive integer of 2 or more), an I / O device 12 (Input / Output device), and one or more memories 19 (FIG. 2).
- m is a positive integer
- m is a positive integer).
- the physical machine 1 also has an OS 14 (Operating System) operating on the CPU 11, the I / O device 12, and the memory unit 13.
- the OS 14 includes a process characteristic collection unit 141, a core allocation determination unit 142, and a process scheduler 143. Furthermore, on the OS 14, processes 151 to 15 n (n is a positive integer) as applications run.
- the CPU 11 has a plurality of CPU cores 18, and in addition to the normal CPU processing, it has a function called power gating that can shut off a part of the power supply of the CPU circuit at an arbitrary timing. In the present embodiment, power shutoff / recovery control of the CPU core 18 is possible.
- the I / O device 12 has an interface for exchanging data with various devices such as a network.
- the memory unit 13 has a memory 19 that can operate in response to the operation of each CPU core 18.
- the memory 19 can be a non-volatile memory, and can shut off the power of the memory 19 not in use. In that case, the stored data can be read out after power recovery without erasing the data.
- the memory unit 13 may also be a memory having one or more memory areas.
- the memory area corresponds to the memory 19 described above.
- ReRAM Resistive RAM
- MRAM Magneticoresistive RAM
- STT-MRAM Spin Transfer Torque-MRAM
- PRAM Phase change RAM
- FeRAM Feroelectric RAM
- the OS 14 has a process characteristic collection unit 141, a core allocation determination unit 142, and a process scheduler 143.
- Process execution control for enhancing the power saving effect by powering off the CPU core 18 of the CPU 11 and the memory 19 of the memory unit 13. I do.
- the process characteristic collection unit 141 collects information such as load characteristics of the processes 151 to 15 n, for example, CPU utilization, cache hit rate, frequency of context switch, resource utilization characteristics, for example, network input / output amount.
- the core allocation determination unit 142 determines allocation of which process should be executed by which CPU core 18 based on the load characteristics of the processes 151 to 15n collected by the process characteristic collection unit 141 and the utilization characteristics of resources.
- the core allocation determination unit 142 monitors the high load of each process and the high utilization frequency of resources, collects the low ones and allocates them to the same CPU core 18. Similarly, processes with high load and resource usage frequency are collected and allocated to the same CPU core 18. In this case, in order to avoid deterioration of the processing capacity of the CPU core 18, it is assumed that the CPU core 18 and resources are collected and allocated in a range not exceeding the upper limit. In this way, in the CPU core 18 where the processes with low load are collected, it is possible to increase the pause time in the CPU power gating.
- the process scheduler 143 schedules the execution of the process according to the determination of the core allocation determination unit 142.
- FIG. 3 is a flow chart showing the procedure of the power saving control method of the power saving control system of the present embodiment shown in FIG.
- the process characteristic collection unit 141 collects load characteristics and resource use characteristics of the processes 151 to 15 n operating on the physical machine 1 (step A 1).
- the core allocation determination unit 142 determines allocation of which process should be executed by which CPU core 18 from the process characteristics collected by the process characteristic collection unit 141 (step A2).
- allocation is performed to the CPU core 18 in the range not exceeding the upper limit value of the CPU utilization rate in order from the process with the highest CPU utilization rate.
- the best fit algorithm in the bin packing problem can be used, but it is not limited to a specific algorithm. In this manner, in the CPU core 18 where the processes with low load are collected, it is possible to lengthen the pause time in the power gating of the CPU 11.
- the process scheduler 143 executes and controls the process according to the correspondence between each process determined by the core allocation determination unit 142 and the CPU core 18 (step A3).
- the power OFF / ON of the CPU core 18 and the memory 19 interlocked with the CPU core 18 is automatically controlled by the power gating function of the CPU 11. That is, even if the system is operating, at that moment, the power of components other than the components such as the CPU core 18 and the memory 19 which should operate truly is turned off, and the CPU core 18 which should operate truly Power of components such as the memory 19 can be turned on.
- the memory unit 13 uses the non-volatile memory 19, when the power of the memory 19 is turned on / off, another data stored in the memory 19 is saved or saved. There is no need to recover from the memory to the memory 19. Therefore, smooth and normally off computing with reduced delay is possible.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 3.
- the CPU core to which a process having a low CPU usage rate is allocated has a long idle time, the CPU core and the memory interlocked therewith are continuously kept long by the power OFF. it can. Thereby, the power saving effect can be enhanced.
- FIG. 4 is a block diagram showing a configuration of a power saving control system according to a third embodiment of the present invention.
- the inside of the physical machine 1 in FIG. 2 is the physical machine 2 virtualized by the hypervisor 16, and a plurality of virtual machines (Virtual Machine, VM, etc.) are generated on the hypervisor 16.
- the present embodiment differs from the power saving system of the second embodiment in that 171 to 17 n operate.
- VMs operating on the hypervisor 16 are identified by assigning reference numerals 171 to 17 n (n is a positive integer).
- FIG. 5 is a flow chart showing the procedure of the power saving control method of the power saving control system of the present embodiment shown in FIG. In this power saving control system, the power saving effect is enhanced by devising the scheduling of VMs.
- the VM characteristic collection unit 161 collects load characteristics and resource use characteristics of the VMs 171 to 17n operating on the physical machine 2 (step B1).
- the core allocation determination unit 162 determines allocation of which VM should be executed by which CPU core 18 from the VM characteristics collected by the VM characteristics collection unit 161 (step B2).
- VMs with the highest CPU usage rate are allocated to a smaller number of CPU cores 18 in order not to exceed the upper limit value of the CPU usage rate.
- the best fit algorithm in the bin packing problem can be used, but it is not limited to a specific algorithm. In this manner, in the CPU core 18 where VMs with low load are collected, it is possible to increase the pause time in power gating of the CPU.
- the VM scheduler 163 executes and controls the VM according to the correspondence between each VM and the CPU core determined by the core allocation determination unit 162 (step B3).
- the power OFF / ON of the CPU core 18 and the memory 19 is automatically controlled by the power gating function of the CPU 11. That is, even if the system is operating, at that moment, the power of components other than the components such as the CPU core 18 and the memory 19 which should operate truly is turned off, and the CPU core 18 which should operate truly Power of components such as the memory 19 can be turned on.
- the memory unit 13 uses the non-volatile memory 19, when the power of the memory 19 is turned on / off, another data stored in the memory 19 is saved or saved. There is no need to recover from the memory to the memory 19. Therefore, smooth and normally off computing with reduced delay is possible.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 4. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 5.
- the CPU core to which a VM having a low CPU utilization rate is allocated has a long idle time, the CPU core and the memory interlocked with it are required to have a long pause time due to power off. it can. Thereby, the power saving effect can be enhanced.
- the fourth embodiment of the present invention has the configuration of the physical machine 1 shown in FIG. Further, the process allocation method of the core allocation determination unit 142 in FIG. 2 is different from the second embodiment in that allocation is performed so that the usage frequency is equalized as much as possible, instead of aggregating those with high usage frequency. .
- Step A1 of FIG. 3 is the same processing as that of the second embodiment.
- the core allocation determination unit 142 determines allocation of which process should be executed by which CPU core 18 from the load characteristics and resource utilization characteristics of the processes 151 to 15n collected in step A1. .
- each process is allocated to the CPU core 18 so that the sum of the CPU utilization of each process allocated to each CPU core 18 for each CPU core 18 is as equal as possible based on the CPU utilization of each process . That is, each process is assigned to the CPU core 18 so that the difference in CPU utilization rate among the CPU cores 18 becomes as small as possible.
- Step A3 in FIG. 3 is the same process as in the second embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 3.
- each CPU core since each CPU core has a similar idle time, it is possible to ensure on average from all the CPU cores the pause time due to the power off of the CPU core and the memory linked thereto. . Thereby, the power saving effect can be enhanced.
- the fifth embodiment of the present invention has the configuration of the physical machine 2 shown in FIG.
- the VM allocation method of the core allocation determination unit 162 in FIG. 4 is different from the third embodiment in that allocation is performed so that the usage frequency is equalized as much as possible, instead of aggregating those with high usage frequency. .
- Step B1 of FIG. 5 is the same processing as that of the second embodiment.
- the core allocation determination unit 162 determines allocation of which VM should be executed by which CPU core 18 from the load characteristics of VMs 171 to 17n collected in step B1 and the resource utilization characteristics.
- each VM is allocated to the CPU core 18 so that the sum of the CPU utilization of each VM allocated to the CPU core 18 for each CPU core 18 is as equal as possible based on the CPU utilization of each VM. That is, each VM is allocated to the CPU core 18 so that the difference in CPU utilization rate for each CPU core 18 becomes as small as possible.
- Step B3 in FIG. 5 is the same process as in the third embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 4. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 5.
- each CPU core since each CPU core has a similar idle time, it is possible to ensure on average from all the CPU cores the pause time due to the power off of the CPU core and the memory linked thereto. . Thereby, the power saving effect can be enhanced.
- the sixth embodiment of the present invention has the configuration of the physical machine 1 shown in FIG. Then, the process allocation method of the core allocation determination unit 142 in FIG. 2 is based on the high frequency of use, not the high CPU utilization of each process as an example, but the high frequency of process context switches. Is different from the second embodiment in that
- Step A1 of FIG. 3 is the same processing as that of the second embodiment.
- step A2 of FIG. 3 the core allocation determination unit 142 determines allocation of which process should be executed by which CPU core 18 from the load characteristics and resource utilization characteristics of the processes 151 to 15n collected in step A1. .
- processes are allocated to the CPU core 18 in the range not exceeding the upper limit value of the frequency of context switches in order from the process with the high frequency of context switches.
- Step A3 in FIG. 3 is the same process as in the second embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 3.
- the CPU core to which the process with low context switching frequency is allocated since the CPU core to which the process with low context switching frequency is allocated has a long idle time, the CPU core and the memory interlocked with it are required to have a long pause time due to the power OFF. it can. Thereby, the power saving effect can be enhanced.
- the seventh embodiment of the present invention has the configuration of the physical machine 2 shown in FIG.
- the VM allocation method of the core allocation determination unit 162 in FIG. 4 is based on the height of the usage frequency, for example, the height of the frequency of context switches of VMs, not the height of CPU utilization of each VM. Is different from the third embodiment in that
- Step B1 in FIG. 5 is the same process as in the third embodiment.
- step B2 of FIG. 5 the core allocation determination unit 162 determines allocation of which VM should be executed by which CPU core 18 from the load characteristics of VMs 171 to 17n collected in step B1 and the resource utilization characteristics.
- VMs with the highest frequency of context switches are allocated to the CPU core 18 in order not to exceed the upper limit of the frequency of context switches.
- Step B3 in FIG. 5 is the same process as in the third embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 4. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 5.
- the CPU core to which a VM with a low context switching frequency is allocated has a long idle time, the CPU core and the memory interlocked with it are required to have a long pause time due to power off. it can. Thereby, the power saving effect can be enhanced.
- the eighth embodiment of the present invention has the configuration of the physical machine 1 shown in FIG. Then, the process allocation method of the core allocation determination unit 142 in FIG. 2 is not to consolidate high frequency context switches, but to allocate so that the frequency of process context switches is equalized as much as possible. It differs from the sixth embodiment.
- Step A1 of FIG. 3 is the same process as the sixth embodiment.
- the core allocation determination unit 142 determines allocation of which process should be executed by which CPU core 18 from the load characteristics and resource utilization characteristics of the processes 151 to 15n collected in step A1. .
- each process is assigned to the CPU core so that the sum of the context switch frequency of the process assigned to the CPU core 18 for each CPU core is as equal as possible based on the frequency of the context switch of each process To go. That is, each process is assigned to the CPU core such that the difference in the frequency of context switches for each CPU core 18 becomes as small as possible.
- Step A3 in FIG. 3 is the same process as in the sixth embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 3.
- each CPU core since each CPU core has a similar idle time, it is possible to ensure on average from all the CPU cores the pause time due to the power off of the CPU core and the memory linked thereto. . Thereby, the power saving effect can be enhanced.
- the ninth embodiment of the present invention has the configuration of the physical machine 2 shown in FIG. Then, the VM allocation method of the core allocation determination unit 162 of FIG. 4 is not to consolidate the ones with high frequency of context switches, but to allocate so that the frequency of context switches of VMs is equalized as much as possible. It differs from the seventh embodiment.
- Step B1 in FIG. 5 is the same process as in the seventh embodiment.
- the core allocation determination unit 162 determines allocation of which VM should be executed by which CPU core 18 from the load characteristics of VMs 171 to 17n collected in step B1 and the resource utilization characteristics.
- each VM is assigned to the CPU core 18 so that the sum of the frequency of context switches of the VMs assigned to the CPU core 18 for each CPU core is as equal as possible based on the frequency of the context switch of each VM. I will assign. That is, each VM is allocated to the CPU core 18 so that the difference in the frequency of context switches for each CPU core 18 becomes as small as possible.
- Step B3 in FIG. 5 is the same process as in the seventh embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 4. Further, the power saving control program of the present embodiment is a power saving control program for executing the flowchart of FIG. 5.
- each CPU core since each CPU core has a similar idle time, it is possible to ensure on average from all the CPU cores the pause time due to the power off of the CPU core and the memory linked thereto. . Thereby, the power saving effect can be enhanced.
- FIG. 6 is a block diagram showing a configuration of a power saving control system of a non-volatile memory mounted server according to a tenth embodiment of the present invention.
- the power saving control system of the present embodiment has a physical machine 1, and the physical machine 1 has a CPU 11 having a plurality of CPU cores 18 (k in FIG. 6, k is a positive integer of 2 or more), I / O devices. 12, a memory unit 13 having one or more memories 19 (m in FIG. 6, m is a positive integer).
- the physical machine 1 also has an OS 14 operating on the CPU 11, the I / O device 12, and the memory 13, and the OS 14 has a process characteristic collection unit 141, an interrupt aggregation unit 144, and a process scheduler 143. Furthermore, on the OS 14, processes 151 to 15 n (n is a positive integer) as applications run.
- the CPU 11 has a plurality of CPU cores 18, and in addition to the normal CPU processing, it has a function called power gating that can shut off a part of the power supply of the CPU circuit at an arbitrary timing. In the present embodiment, power shutoff / recovery control of the CPU core 18 is possible.
- the I / O device 12 has an interface for exchanging data with various devices such as a network.
- the memory unit 13 has a memory 19 that can operate in response to the operation of each CPU core 18.
- the memory 19 can be a non-volatile memory, and can shut off the power of the memory 19 not in use. In that case, the stored data can be read out after power recovery without erasing the data.
- the memory unit 13 may also be a memory having one or more memory areas.
- the memory area corresponds to the memory 19 described above.
- Non-volatile memory ReRAM, MRAM, STT-MRAM, PRAM, FeRAM can be used. Further, as long as the memory is nonvolatile, it is not limited to these.
- the OS 14 has a process characteristic collection unit 141, an interrupt concentration unit 144, and a process scheduler 143, and performs process execution control for enhancing the power saving effect by powering off the CPU core 18 of the CPU 11 and the memory 19 of the memory unit 13. Do.
- the process characteristic collection unit 141 includes information such as load characteristics of the processes 151 to 15 n, for example, CPU utilization, cache hit rate, frequency of context switch, and interrupt characteristics, for example, frequency of interrupts due to input and output of devices. collect.
- the interrupt aggregation unit 144 aggregates interrupts to a process with low load characteristics from the load characteristics and interrupt characteristics of the processes 151 to 15 n collected by the process characteristic collection unit 141.
- the interrupt aggregating unit 144 monitors the level of load characteristic of each process and the level of interrupt frequency, and integrates interrupts to those low processes. In this way, it is possible to increase the continuous pause time of the CPU core 18 in the process with low load characteristics and interrupt frequency.
- the process scheduler 143 executes the interrupts collected by the interrupt collection unit 144.
- FIG. 7 is a flow chart showing the procedure of the power saving control method of the power saving control system of the present embodiment shown in FIG.
- the process characteristic collection unit 141 collects the load characteristics of the processes 151 to 15 n operating on the physical machine 1 and the interrupt characteristics, that is, the resource use characteristics (step C1).
- the interrupt aggregation unit 144 refers to the process characteristic collected by the process characteristic collection unit 141, and aggregates interrupts for the process with a small load (step C2).
- interrupts are aggregated if they fall below a preset threshold.
- the aggregation of interrupts is performed by storing randomly generated interrupts in a queue, and accumulating a certain number of cases or when a certain period of time has elapsed. In this way, it is possible to increase the continuous pause time of the CPU core 18 in a process with low load and interrupt frequency.
- step C3 the process scheduler 143 executes the interrupt consolidated by the interrupt aggregating unit 144 (step C3).
- the power OFF / ON of the CPU core 18 and the memory 19 is automatically controlled by the power gating function of the CPU 11. That is, even if the system is operating, at the moment, the power of components other than the CPU core 18 and the memory 19 which should operate truly is turned off, and the CPU core 18 and the memory 19 should operate truly.
- the power of the components of can be turned on.
- the memory unit 13 uses the non-volatile memory 19, when the power of the memory 19 is turned on / off, another data stored in the memory 19 is saved or saved. There is no need to recover from the memory to the memory 19. Therefore, smooth and normally off computing with reduced delay is possible.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG.
- the power saving control program of the present embodiment is a power saving control program that executes the flowchart of FIG. 7.
- interrupts to processes with low CPU utilization and low interrupt frequency are concentrated, and therefore, it is possible to ensure a long pause time due to power-off of the CPU core and the memory linked thereto. Thereby, the power saving effect can be enhanced.
- FIG. 8 is a block diagram showing a configuration of a power saving control system according to an eleventh embodiment of the present invention.
- the inside of the physical machine 1 of FIG. 6 is the physical machine 2 virtualized by the hypervisor 16, and a plurality of virtual machines (VMs) 171 to 171 are provided on the hypervisor 16. It differs from the power saving system of the tenth embodiment in that 17n operates.
- VMs operating on the hypervisor 16 are identified by assigning reference numerals 171 to 17 n (n is a positive integer).
- FIG. 9 is a flow chart showing the procedure of the power saving control method of the power saving control system of the present embodiment shown in FIG.
- the CPU pause time is extended to enhance the power saving effect.
- the VM characteristic collecting unit 161 collects load characteristics of the VMs 171 to 17n operating on the physical machine 2 and interrupt characteristics, that is, resource utilization characteristics (step D1).
- the interrupt collecting unit 164 collects interrupts for VMs with low load characteristics (step D2).
- interrupts are aggregated if they fall below a preset threshold.
- the aggregation of interrupts is performed by storing randomly generated interrupts in a queue, and accumulating a certain number of cases or when a certain period of time has elapsed. In this way, it is possible to increase the continuous pause time of the CPU core 18 in the VM with low load and interrupt frequency.
- the VM scheduler 163 executes the interrupt aggregated by the interrupt aggregation unit 164 (step D3).
- the power OFF / ON of the CPU core 18 and the memory 19 is automatically controlled by the power gating function of the CPU 11. That is, even if the system is operating, at that moment, the power of components other than the CPU core 18 and the memory 19 which should operate truly is turned off, and the CPU core 18 and the memory 18 should operate truly.
- the power of the components of can be turned on.
- the memory unit 13 uses the non-volatile memory 19, when the power of the memory 19 is turned on / off, another data stored in the memory 19 is saved or saved. There is no need to recover from the memory to the memory 19. Therefore, smooth and normally off computing with reduced delay is possible.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 8 described above. Further, the power saving control program of the present embodiment is a power saving control program for executing the above-described flowchart of FIG. 9.
- the twelfth embodiment of the present invention has the configuration of the physical machine 1 shown in FIG. Further, the interrupt concentration of the interrupt concentration unit 144 in FIG. 6 is different from that of the tenth embodiment in that interrupts to processes with low priority are collected.
- Step C1 of FIG. 7 is the same processing as the tenth embodiment. However, process priority is collected as an interrupt characteristic.
- step C2 of FIG. 7 the interrupt aggregation unit 144 aggregates interrupts to processes with low process priority from the load characteristics and interrupt characteristics of the processes 151 to 15n collected in step C1.
- a process with a low process priority may reduce the frequency of execution when the load on the physical machine 1 becomes high, so even if interrupts are preferentially aggregated, the influence is small.
- Step C3 in FIG. 7 is the same process as the tenth embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG.
- the power saving control program of the present embodiment is a power saving control program that executes the flowchart of FIG. 7.
- the thirteenth embodiment of the present invention has the configuration of the physical machine 2 shown in FIG. Further, the interrupt concentration of the interrupt concentration unit 164 in FIG. 8 is different from the eleventh embodiment in that interrupts to VMs with low priorities are collected.
- Step D1 in FIG. 9 is the same process as in the eleventh embodiment.
- VM priorities are collected as interrupt characteristics.
- step D2 of FIG. 9 the interrupt aggregation unit 164 integrates interrupts to VMs with low VM priority from the load characteristics and interrupt characteristics of the VMs 171 to 17n collected in step D1. Since VMs with low VM priority may be executed less frequently if the load on the physical machine 2 is high, even if interrupts are preferentially aggregated, the impact is small.
- Step D3 in FIG. 9 is the same process as in the eleventh embodiment.
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 8 described above. Further, the power saving control program of the present embodiment is a power saving control program for executing the above-described flowchart of FIG. 9.
- FIG. 10 is a block diagram showing a configuration of a power saving control system of a non-volatile memory mounted server according to a fourteenth embodiment of the present invention.
- the power saving control system of the present embodiment includes a physical machine 1, and the physical machine 1 includes a CPU 11, an I / O device 12, and a memory unit 13.
- the physical machine 1 also has an OS 14 operating on the CPU 11, the I / O device 12, and the memory unit 13, and the OS 14 has a process scheduler 143 and a memory power control unit 145.
- processes 151 to 15n (n is a positive integer), which are applications, operate.
- the CPU 11 is a CPU that executes the OS 14 and the processes 151 to 15 n.
- the I / O device 12 has an interface for exchanging data with various devices such as a network.
- the memory unit 13 has one or more memories 19 (m in FIG. 10, m is a positive integer).
- the memory 19 can be a non-volatile memory, and can shut off the power of the memory 19 not in use. In that case, the stored data can be read out after power recovery without erasing the data.
- the memory unit 13 may also be a memory having one or more memory areas.
- the memory area corresponds to the memory 19 described above.
- Non-volatile memory ReRAM, MRAM, STT-MRAM, PRAM, FeRAM or the like can be used. Further, as long as the memory is nonvolatile, it is not limited to these.
- the OS 14 includes a process scheduler 143 and a memory power control unit 145, and performs process execution control for enhancing the power saving effect by the power shutoff of the memory 19 of the memory unit 13.
- the process scheduler 143 schedules the processes 151 to 15n, notifies the memory power control unit 145 of the process 15i (i is 1 to n) to be executed from now, and then executes the scheduled process 15i.
- the memory power control unit 145 turns on the power of the memory 19 of the process 15i notified from the process scheduler 143, and turns off the power of the memory 19 of the process which has been executed up to that point.
- the power saving control system turns on the power of the memory 19 of the process operating on the OS 14 and turns off the power of the memory 19 of the other non-executing processes.
- the power supply cutoff range can be broadened, and the power saving effect can be enhanced.
- the memory unit 13 uses the non-volatile memory 19, when the power of the memory 19 is turned on / off, another data stored in the memory 19 is saved or saved. There is no need to recover from the memory to the memory 19. Therefore, smooth and normally off computing with reduced delay is possible.
- FIG. 11 is a flow chart showing the procedure of the power saving control method of the power saving control system of the present embodiment shown in FIG.
- the process scheduler 143 determines a process 15i to be executed among the processes 151 to 15n operating on the physical machine 1 (step E1).
- the selection of the process 15i to be performed assumes a general scheduler scheduling algorithm and is not limited to a specific algorithm here.
- FIFO First In, First Out
- round robin which is executed in order of each process
- the algorithm of the process scheduler is not limited to these.
- the process scheduler 143 notifies the memory power control unit 145 of the process 15i to be executed (step E2).
- the memory power control unit 145 turns on the power of the memory 19 operated in the process 15i (step E3).
- the memory power control unit 145 turns off the power of the memory 19 operated in the process which has been executed so far (step E4). At this time, if the memory 19 operated in the process executed so far overlaps with the memory 19 operated in the process 15 i executed from now on, the power supply of the memory 19 operated in the process executed so far Will remain ON.
- the process scheduler 143 executes the process 15i to be executed (step E5).
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 1 of FIG. 10 described above. Further, the power saving control program according to the present embodiment is a power saving control program for executing the above-described flowchart of FIG.
- the power of the memory of the process to be executed is turned on by the process execution control and the memory power control described above, and the power of the memory of the other non-running processes is turned off It is possible to widen the cutoff range of the power supply. Thereby, the power saving effect can be enhanced.
- FIG. 12 is a block diagram showing a configuration of a power saving control system according to a fifteenth embodiment of the present invention.
- the inside of the physical machine 1 in FIG. 10 is the physical machine 2 virtualized by the hypervisor 16, and a plurality of virtual machines (VMs) 171 to 171 are provided on the hypervisor 16.
- VMs virtual machines
- This embodiment differs from the power saving system of the fourteenth embodiment in that 17n operates.
- VMs operating on the hypervisor 16 are identified by assigning reference numerals 171 to 17 n (n is a positive integer).
- the power supply cutoff range of the memory unit 13 can be expanded by turning on the power supply of the memory 19 of the running VM and turning off the power supply of the memory 19 of the other non-running VMs. Power saving effect can be enhanced.
- the memory unit 13 uses the non-volatile memory 19, when the power of the memory 19 is turned on / off, another data stored in the memory 19 is saved or saved. There is no need to recover from the memory to the memory 19. Therefore, smooth and normally off computing with reduced delay is possible.
- FIG. 13 is a flowchart showing the procedure of the power saving control method of the power saving control system of the present embodiment shown in FIG.
- the VM scheduler 163 determines the VM 17i (i is 1 to n) to be executed from the VMs 171 to 17n operating on the physical machine 2 (step F1).
- the selection of the VM 17i to be executed assumes a general VM scheduler scheduling algorithm, and is not limited to a specific algorithm here.
- the VM scheduler algorithm can use a FIFO that executes in the order of arrival to the executable queue, and a round robin that executes each VM in a certain order. It is also possible to use a priority preemptive which is preferentially executed in order from a VM having a high priority for a certain period of time.
- the algorithm of VM scheduler is not limited to these.
- the VM scheduler 163 notifies the memory power control unit 165 of the VM 17i to be executed (step F2).
- the memory power control unit 165 turns on the power of the memory 19 operated in the VM 17i (step F3).
- the memory power control unit 165 turns off the power of the memory 19 operated in the VM which has been executed so far (step F4). At this time, if the memory 19 operated in the VM executed so far overlaps with the memory 19 operated in the VM 17i executed from now on, the power supply of the memory 19 operated in the VM executed so far is Leave it ON.
- the VM scheduler 163 executes the VM 17i to be executed (step F5).
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 2 of FIG. 12 described above. Further, the power saving control program of the present embodiment is a power saving control program for executing the above-described flowchart of FIG. 13.
- the power of the memory of the VM to be executed is turned on by the VM execution control and the memory power control described above, and the power of the memory of the VM not in execution is turned off. It is possible to widen the cutoff range of the power supply. Thereby, the power saving effect can be enhanced.
- the sixteenth embodiment of the present invention has the configuration of the physical machine 3 shown in FIG.
- the configuration of the physical machine 3 of the present embodiment differs from that of the fourteenth embodiment in that the OS 14 shown in FIG. 10 includes the OS 15 to which the cache information collecting unit 146 is added.
- the other configuration is the same as that of the fourteenth embodiment.
- FIG. 15 is a flow chart showing the processing procedure of the power saving control system shown in FIG.
- the cache information collecting unit 146 collects information of pages of the memory 19 held in the cache by the CPU 11 (step G1). This information shall be collected synchronously with the cache update.
- the process scheduler 143 determines a process 15i (i is 1 to n) to be executed among the processes 151 to 15n operating on the physical machine 3 (step G2).
- the selection of the process 15i to be performed assumes a general scheduler scheduling algorithm and is not limited to a specific algorithm here.
- the process scheduler 143 notifies the memory power control unit 145 of the process 15i to be executed (step G3).
- the memory power control unit 145 acquires, from the cache information collection unit 146, information on pages in the cache in the memory 19 operated in the process 15i (step G4).
- the memory power control unit 145 powers on the memory 19 having the page in the cache among the memories of the process 15i (step G5).
- the memory power control unit 145 turns off the power of the memory 19 operated in the process which has been executed so far (step G6). At this time, if the memory 19 operated in the process executed so far overlaps with the memory 19 operated in the process 15 i executed from now on, the power supply of the memory 19 operated in the process executed so far Will remain ON.
- the process scheduler 143 executes the process 15i to be executed (step G7).
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 3 of FIG. 14 described above. Also, the power saving control program of the present embodiment is a power saving control program that executes the above-described flowchart of FIG.
- the page in the cache is turned on to store the non-cached page memory and the other memory.
- the seventeenth embodiment of the present invention has the configuration of a physical machine 4 shown in FIG.
- the configuration of the physical machine 4 of this embodiment is different from the fifteenth embodiment in that the hypervisor 16 shown in FIG. 12 has the hypervisor 17 to which the cache information collecting unit 166 is added.
- the other configuration is the same as that of the fifteenth embodiment.
- FIG. 17 is a flowchart showing the processing procedure of the power saving control system shown in FIG.
- the cache information collecting unit 166 collects information of pages of the memory 19 held in the cache by the CPU 11 (step H1). This information shall be collected synchronously with the cache update.
- the VM scheduler 163 determines a VM 17i (i is 1 to n) to be executed among the VMs 171 to 17n operating on the physical machine 4 (step H2).
- the selection of the VM 17i to be executed assumes a general VM scheduler scheduling algorithm, and is not limited to a specific algorithm here.
- the VM scheduler 163 notifies the memory power control unit 165 of the VM 17i to be executed (step H3).
- the memory power control unit 165 acquires, from the cache information collection unit 166, information on pages in the cache in the memory 19 operated by the VM 17i (step H4).
- the memory power control unit 165 powers on the memory 19 having the page in the cache among the memory of the VM 17i (step H5).
- the memory power control unit 165 turns off the power of the memory 19 operated by the VM which has been executed so far (step H6). At this time, if the memory 19 operated in the VM which has been executed up to now overlaps with the memory 19 which is operated in the VM 15i to be executed from now on, the power supply of the memory 19 operated in the VM which has been executed so far is Leave it ON.
- the VM scheduler 163 executes the VM 17i to be executed (step H7).
- the power saving control device of the present embodiment is a power saving control device having the configuration of the physical machine 4 of FIG. 16 described above. Also, the power saving control program of the present embodiment is a power saving control program that executes the above-described flowchart of FIG. 17.
- the power of the page in the cache is turned on by the VM execution control and the memory power control, and the memory of the non-cached page and the other memory Shut down the memory of VMs that are not running. Therefore, the cutoff range of the power supply of the memory can be broadened. Thereby, the power saving effect can be enhanced.
- the eighteenth embodiment of the present invention has the configuration of the physical machine 3 shown in FIG. 14, which is the same as the sixteenth embodiment.
- the cache information collecting unit 146 collects information of pages of the memory 19 held in the cache by the CPU 11 (step G1). Furthermore, the memory power control unit 145 acquires, from the cache information collection unit 146, information on pages in the cache of the memory 19 operated in the process 15i (step G4). Further, the memory power control unit 145 turns on the power of the memory 19 having the page in the cache among the memories of the process 15i (step G5), and the memory 19 operated by the process which has been executed so far. Power off (step G6).
- the cache information collection unit 146 collects the cache read / write ratio of the CPU 11. Furthermore, the memory power control unit 145 acquires, from the cache information collection unit 146, information on the cache read / write ratio of the CPU 11 in the memory 19 operated in the process 15i. Furthermore, the memory power control unit 145 turns off the power of the memory 19 corresponding to the cache when the write ratio of the read / write ratio of the cache of the CPU 11 is lower than the threshold among the memories of the process 15i.
- the nineteenth embodiment of the present invention has the configuration of the physical machine 4 shown in FIG. 16, which is the same as the seventeenth embodiment.
- the cache information collecting unit 166 collects information of pages of the memory 19 held in the cache by the CPU 11 (step H1). Furthermore, the memory power control unit 165 acquires, from the cache information collection unit 166, information on pages in the cache in the memory 19 operated by the VM 17i (step H4). Further, the memory power control unit 165 turns on the power of the memory 19 having the page in the cache among the memories of the VM 17i (step H5), and the memory 19 operated by the VM which has been executed so far. The power is turned off (step H6).
- the cache information collecting unit 166 collects the read / write ratio of the cache of the CPU 11. Furthermore, the memory power control unit 165 acquires, from the cache information collection unit 166, information on the cache read / write ratio of the CPU 11 in the memory 19 operated by the VM 17i. Furthermore, the memory power control unit 165 turns off the power of the memory 19 corresponding to the cache when the write ratio of the read / write ratio of the cache of the CPU 11 is lower than the threshold among the memories of the VM 17i.
- a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system running on the physical machine, and one or more processes running on the operating system, the operating system including the process
- a process characteristic collection unit collecting characteristics of the core, a core allocation determination unit determining an allocation of the CPU core to the process based on the characteristics of the process, and a process scheduler executing the process based on the allocation
- a power saving control system comprising: powering off / on said CPU core or said memory based on execution of said process.
- the power saving control system according to any one of appendices 1 to 9, wherein the operating system is a hypervisor and the process is a virtual machine.
- a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system unit operating on the physical machine, and one or more processing units operated by the operating system unit;
- a process characteristic collection unit collecting characteristics of the process unit, a core allocation determination unit determining an allocation of the CPU core to the process unit based on the characteristics of the process unit, and a process unit based on the allocation
- a power saving control device comprising: a process scheduler for executing, and powering off / on the CPU core or the memory based on execution of the process unit.
- the power saving control device according to any one of appendices 11 to 19, wherein the operating system unit is a hypervisor and the process unit is a virtual machine.
- Power saving control method of power saving control system having physical machine having CPU and memory having multiple CPU cores, operating system running on the physical machine, and one or more processes running on the operating system The operating system collects the characteristics of the process, determines an assignment of the CPU core to the process based on the characteristics of the process, executes a process based on the assignment, and executes the process A power saving control method, wherein the power of the CPU core or the memory is turned on / off based on the power.
- Power saving control program for a power saving control system comprising: a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system running on the physical machine, and one or more processes running on the operating system A process of collecting characteristics of the process, a process of determining assignment of the CPU core to the process based on the process characteristic, and a process of executing the process based on the assignment.
- Appendix 40 The power saving control program according to any one of appendices 31 to 39, wherein the operating system is a hypervisor and the process is a virtual machine.
- Appendix 41 A physical machine having a CPU having a plurality of CPU cores and a memory, an operating system running on the physical machine, and one or more processes running on the operating system, the operating system including the process A process characteristic collection unit that collects the load characteristics of the process or the interrupt characteristic of the process, an interrupt aggregation unit that aggregates the interrupts to the process based on the load characteristic or the interrupt characteristic, and an interrupt based on the aggregation of the interrupts A power saving control system for performing power off / on of the CPU core or the memory based on the execution of the interrupt.
- (Appendix 42) The power saving control system according to claim 41, wherein the memory is a non-volatile memory.
- (Appendix 43) The power saving control system according to appendix 41 or 42, wherein the load characteristic is a CPU usage rate, a cache hit rate, or a frequency of context switch.
- (Appendix 44) The power saving control system according to any one of Appendices 41 to 43, wherein the interrupt characteristic is an interrupt frequency due to input / output to a device, or a priority of the process.
- (Appendix 45) The power saving control system according to any one of appendices 41 to 44, wherein the interrupt aggregation unit aggregates interrupts to a process with low load characteristics or low interrupt characteristics.
- a power saving control system according to one of the claims 41 to 45, wherein the operating system is a hypervisor and the process is a virtual machine.
- a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system unit operating on the physical machine, and one or more processing units operating on the operating system;
- a process characteristic collection unit that collects load characteristics of the process unit or interrupt characteristics of the process unit;
- an interrupt aggregation unit that aggregates interrupts to the process unit based on the load characteristics or the interrupt characteristics;
- a process scheduler that executes an interrupt based on the aggregation of the power saving control device, and performs power supply on / off of the CPU core or the memory based on the execution of the interrupt.
- Appendix 48 The power saving control device according to appendix 47, wherein the memory is a non-volatile memory.
- Appendix 50 The power saving control device according to any one of Appendices 47 to 49, wherein the interrupt characteristic is an interrupt frequency due to input / output to a device, or a priority of the process.
- Appendix 51 The power saving control device according to any one of Appendices 47 to 50, wherein the interrupt aggregation unit aggregates interrupts to a process unit having the low load characteristic or the low interrupt characteristic.
- (Appendix 52) The power saving control device according to any one of Appendices 47 to 51, wherein the operating system unit is a hypervisor and the process unit is a virtual machine.
- (Appendix 53) Power saving control method of power saving control system having physical machine having CPU and memory having multiple CPU cores, operating system running on the physical machine, and one or more processes running on the operating system Wherein the operating system collects load characteristics of the process or interrupt characteristics of the process, aggregates interrupts to the process based on the load characteristics or the interrupt characteristics, and interrupts based on the aggregation of interrupts The power saving control method, wherein the power of the CPU core or the memory is turned on / off based on the execution of the interrupt. (Appendix 54) 53.
- the power saving control method according to appendix 53 wherein the memory is a non-volatile memory.
- Appendix 55 The power saving control method according to appendix 53, wherein the load characteristic is a CPU usage rate, a cache hit rate, or a frequency of context switch.
- Appendix 56 The power saving control method according to any one of Appendices 53 to 55, wherein the interrupt characteristic is an interrupt frequency by input / output to a device, or a priority of the process.
- Power saving control program for a power saving control system comprising: a physical machine having a CPU having a plurality of CPU cores and a memory, an operating system running on the physical machine, and one or more processes running on the operating system A process of collecting load characteristics of the process or interrupt characteristics of the process in the operating system, a process of aggregating interrupts into the process based on the load characteristics or the interrupt characteristics, and aggregation of the interrupts A power saving control program for executing a process of executing an interrupt based on the process and a process of turning the power of the CPU core or the memory off / on based on the execution of the interrupt.
- Appendix 60 The power saving control program according to statement 59, wherein the memory is a non-volatile memory.
- the load characteristic is a CPU usage rate, a cache hit rate, or a frequency of context switch.
- the interrupt characteristic is an interrupt frequency due to input / output to a device, or a priority of the process.
- the operating system has a cache information collecting unit that collects page information of a memory held by the CPU in a cache, and the memory power control unit acquires the page information from the cache information collecting unit.
- the power saving control system according to any one of appendices 65 to 69, wherein the power of the memory is turned on based on the page information.
- a power save control system according to any of clauses 65 to 70, wherein the operating system is a hypervisor and the process is a virtual machine.
- Appendix 72 A physical machine having a CPU and a memory, an operating system unit operating on the physical machine, and one or more process units operated by the operating system unit, the operating system unit including the process unit A power saving control device, comprising: a process scheduler that performs scheduling; and a memory power control unit that turns on / off the power of the memory based on the scheduling.
- Appendix 73 The power saving control device according to statement 72, wherein the memory is a non-volatile memory.
- Appendix 74 The power saving control device according to appendix 72 or 73, wherein the process scheduler executes the process unit based on the scheduling.
- the power saving control device according to any one of appendices 72 to 74, wherein the memory power control unit turns on the power of the memory related to the process unit executed by the process scheduler.
- the power saving control device according to any one of Appendices 72 to 75, wherein the memory power control unit turns off the power of the memory other than the memory related to the process unit executed by the process scheduler.
- the operating system unit includes a cache information collection unit that collects page information of a memory held by the CPU in a cache, and the memory power control unit acquires the page information from the cache information collection unit.
- the power saving control device according to any one of appendices 72 to 76, wherein the power of the memory is turned on based on the page information.
- the power saving control device according to any one of Appendices 72 to 77, wherein the operating system unit is a hypervisor and the process unit is a virtual machine.
- Appendix 79 A power saving control method of a power saving control system, comprising: a physical machine having a CPU and a memory; an operating system operating on the physical machine; and one or more processes operating on the operating system A power saving control method of scheduling the process and powering on / off the memory based on the scheduling.
- the operating system collects page information of a memory held by the CPU in a cache, and control of the power supply of the memory acquires the page information and turns on the power of the memory based on the page information.
- the power saving control method according to any one of appendices 79 to 83.
- the power saving control method according to any one of appendices 79 to 84, wherein the operating system is a hypervisor and the process is a virtual machine.
- a power saving control program of a power saving control system comprising: a physical machine having a CPU and a memory, an operating system operating on the physical machine, and one or more processes operating on the operating system
- a power saving control program for executing the process of scheduling the process and the process of turning on / off the power of the memory based on the scheduling.
- Appendix 87 The power saving control program according to appendix 86, wherein the memory is a non-volatile memory.
- Appendix 88 The power saving control program according to appendix 86 or 87, wherein the process of scheduling the process executes the process based on the scheduling.
- the power saving control program according to any one of appendices 86 to 90, wherein the power of the memory is turned on based on the power saving control program.
- Appendix 92 The power saving control program according to any one of appendices 86 to 91, wherein the operating system is a hypervisor and the process is a virtual machine.
- the present invention can be used as a technology for enhancing the power saving effect in the normally-off computing technology in a server equipped with a nonvolatile memory.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Power Sources (AREA)
Abstract
Description
(第1の実施形態)
図1は、本発明の第1の実施形態の不揮発性メモリ搭載サーバの省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、複数のCPUコア18を有するCPU11とメモリ19とを有する物理マシン1と、物理マシン1で稼動するオペレーティングシステム14と、オペレーティングシステム14で稼動する一つ以上のプロセス150とを有する。オペレーティングシステム14は、プロセス150の特性を収集するプロセス特性収集部141を有する。さらに、プロセス150の特性に基づいてプロセス150に対するCPUコア18の割り当てを決定するコア割当決定部142と、割り当てに基づいてプロセス150の実行をするプロセススケジューラ143とを有する。さらに、プロセス150の実行に基づいて、CPUコア18あるいはメモリ19の電源のOFF/ONを行う。
(第2の実施形態)
図2は、本発明の第2の実施形態の不揮発性メモリ搭載サーバの省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、物理マシン1を有する。物理マシン1は、複数個のCPUコア18(図2ではk個、kは2以上の正整数)を有するCPU11、I/Oデバイス12(Input/Outputデバイス)、一個以上のメモリ19(図2ではm個、mは正整数)を有するメモリ部13を有する。また、物理マシン1は、CPU11、I/Oデバイス12、メモリ部13の上で動作するOS14(Operating System)を有する。OS14は、プロセス特性収集部141、コア割当決定部142、プロセススケジューラ143を有する。さらに、OS14の上では、アプリケーションとしてのプロセス151~15n(nは正整数)が稼働する。
(第3の実施形態)
図4は、本発明の第3の実施形態の省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、図2の物理マシン1の内部が、ハイパーバイザ16によって仮想化された物理マシン2であり、ハイパーバイザ16の上で、複数の仮想マシン(Virtual Machine、VM)171~17nが稼働する点で第2の実施形態の省電力システムと異なる。図4では、ハイパーバイザ16上での稼働するVMに符号171~17n(nは正整数)を付与して識別している。
(第4の実施形態)
次に、本発明の第4の実施形態について説明する。本発明の第4の実施形態は、図2に示す物理マシン1の構成を有する。そして、図2のコア割当決定部142のプロセスの割り当て方式が、利用頻度の高いものを集約するのではなく、利用頻度ができるだけ平準化されるように割り当てる点で、第2の実施形態と異なる。
(第5の実施形態)
次に、本発明の第5の実施形態について説明する。本発明の第5の実施形態は、図4に示す物理マシン2の構成を有する。そして、図4のコア割当決定部162のVMの割り当て方式が、利用頻度の高いものを集約するのではなく、利用頻度ができるだけ平準化されるように割り当てる点で、第3の実施形態と異なる。
(第6の実施形態)
次に、本発明の第6の実施形態について説明する。本発明の第6の実施形態は、図2に示す物理マシン1の構成を有する。そして、図2のコア割当決定部142のプロセスの割り当て方式が、利用頻度の高さ、例として、各プロセスのCPU使用率の高さではなく、プロセスのコンテキストスイッチの頻度の高さを基準にする点で、第2の実施形態と異なる。
(第7の実施形態)
次に、本発明の第7の実施形態について説明する。本発明の第7の実施形態は、図4に示す物理マシン2の構成を有する。そして、図4のコア割当決定部162のVMの割り当て方式が、利用頻度の高さ、例として、各VMのCPU使用率の高さではなく、VMのコンテキストスイッチの頻度の高さを基準にする点で、第3の実施形態と異なる。
(第8の実施形態)
次に、本発明の第8の実施形態について説明する。本発明の第8の実施形態は、図2に示す物理マシン1の構成を有する。そして、図2のコア割当決定部142のプロセスの割り当て方式が、コンテキストスイッチの頻度の高いものを集約するのではなく、プロセスのコンテキストスイッチの頻度ができるだけ平準化されるように割り当てる点で、第6の実施形態と異なる。
(第9の実施形態)
次に、本発明の第9の実施の形態について詳細に説明する。本発明の第9の実施形態は、図4に示す物理マシン2の構成を有する。そして、図4のコア割当決定部162のVMの割り当て方式が、コンテキストスイッチの頻度の高いものを集約するのではなく、VMのコンテキストスイッチの頻度ができるだけ平準化されるように割り当てる点で、第7の実施形態と異なる。
(第10の実施形態)
図6は、本発明の第10の実施形態の不揮発性メモリ搭載サーバの省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、物理マシン1を有し、物理マシン1は、複数のCPUコア18(図6ではk個、kは2以上の正整数)を有するCPU11、I/Oデバイス12、一個以上のメモリ19(図6ではm個、mは正整数)を有するメモリ部13を有する。また、物理マシン1は、CPU11、I/Oデバイス12、メモリ13の上で動作するOS14を有し、OS14は、プロセス特性収集部141、割込み集約部144、プロセススケジューラ143を有する。さらに、OS14の上では、アプリケーションとしてのプロセス151~15n(nは正整数)が稼働する。
(第11の実施形態)
図8は、本発明の第11の実施形態の省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、図6の物理マシン1の内部が、ハイパーバイザ16によって仮想化された物理マシン2であり、ハイパーバイザ16の上で、複数の仮想マシン(VM)171~17nが稼働する点で、第10の実施形態の省電力システムと異なる。図8では、ハイパーバイザ16上での稼働するVMに符号171~17n(nは正整数)を付与して識別している。
(第12の実施形態)
次に、本発明の第12の実施形態について説明する。本発明の第12の実施形態は、図6に示す物理マシン1の構成を有する。そして、図6の割込み集約部144の割込み集約が、優先度の低いプロセスへの割込みを集約する点で、第10の実施形態と異なる。
(第13の実施形態)
次に、本発明の第13の実施形態について説明する。本発明の第13の実施形態は、図8に示す物理マシン2の構成を有する。そして、図8の割込み集約部164の割込み集約が、優先度の低いVMへの割込みを集約する点で、第11の実施形態と異なる。
(第14の実施形態)
図10は、本発明の第14の実施形態の不揮発性メモリ搭載サーバの省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、物理マシン1を有し、物理マシン1は、CPU11、I/Oデバイス12、メモリ部13を有する。また、物理マシン1は、CPU11、I/Oデバイス12、メモリ部13の上で動作するOS14を有し、OS14は、プロセススケジューラ143、メモリ電源制御部145を有する。さらに、OS14の上では、アプリケーションであるプロセス151~15n(nは正整数)が稼働する。
(第15の実施形態)
図12は、本発明の第15の実施形態の省電力制御システムの構成を示すブロック図である。本実施形態の省電力制御システムは、図10の物理マシン1の内部が、ハイパーバイザ16によって仮想化された物理マシン2であり、ハイパーバイザ16の上で、複数の仮想マシン(VM)171~17nが稼働する点で、第14の実施形態の省電力システムと異なる。図12では、ハイパーバイザ16上での稼働するVMに符号171~17n(nは正整数)を付与して識別している。
(第16の実施形態)
次に、本発明の第16の実施形態について説明する。本発明の第16の実施形態は、図14に示す物理マシン3の構成を有する。本実施形態の物理マシン3の構成は、図10に示すOS14に、キャッシュ情報収集部146が追加されたOS15を有する点で、第14の実施形態と異なる。その他の構成は、第14の実施形態の構成と同様である。
(第17の実施形態)
次に、本発明の第17の実施形態について説明する。本発明の第17の実施形態は、図16に示す物理マシン4の構成を有する。本実施形態の物理マシン4の構成は、図12に示すハイパーバイザ16に、キャッシュ情報収集部166が追加されたハイパーバイザ17を有する点で、第15の実施形態と異なる。その他の構成は、第15の実施形態の構成と同様である。
(第18の実施形態)
次に、本発明の第18の実施形態について説明する。本発明の第18の実施形態は、第16の実施形態と同じ、図14に示す物理マシン3の構成を有する。
(第19の実施形態)
次に、本発明の第19の実施形態について説明する。本発明の第19の実施形態は、第17の実施形態と同じ、図16に示す物理マシン4の構成を有する。
(付記1)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有し、前記オペレーティングシステムは、前記プロセスの特性を収集するプロセス特性収集部と、前記プロセスの特性に基づいて前記プロセスに対する前記CPUコアの割り当てを決定するコア割当決定部と、前記割り当てに基づいて前記プロセスの実行をするプロセススケジューラと、を有し、前記プロセスの実行に基づいて、前記CPUコアあるいは前記メモリの電源のOFF/ONを行う、省電力制御システム。
(付記2)
前記メモリは不揮発メモリである、付記1記載の省電力制御システム。
(付記3)
前記プロセスの特性は、前記プロセスの負荷特性、あるいは、リソースの利用特性である、付記1または2記載の省電力制御システム。
(付記4)
前記プロセスの負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記3記載の省電力制御システム。
(付記5)
前記リソースの利用特性は、ネットワーク入出力量である、付記3記載の省電力制御システム。
(付記6)
前記コア割当決定部は、前記CPU使用率の高い順に前記CPUコアに割り当てる、付記4記載の省電力制御システム。
(付記7)
前記コア割当決定部は、前記CPU使用率が平準化するように前記CPUコアに割り当てる、付記4記載の省電力制御システム。
(付記8)
前記コア割当決定部は、前記コンテキストスイッチの頻度の高い順に前記CPUコアに割り当てる、付記4記載の省電力制御システム。
(付記9)
前記コア割当決定部は、前記コンテキストスイッチの頻度が平準化するように前記CPUコアに割り当てる、付記4記載の省電力制御システム。
(付記10)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記1から9の内の1項記載の省電力制御システム。
(付記11)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステム部と、前記オペレーティングシステム部によって稼動する一つ以上のプロセス部と、を有し、前記オペレーティングシステム部は、前記プロセス部の特性を収集するプロセス特性収集部と、前記プロセス部の特性に基づいて前記プロセス部に対する前記CPUコアの割り当てを決定するコア割当決定部と、前記割り当てに基づいてプロセス部の実行をするプロセススケジューラと、を有し、前記プロセス部の実行に基づいて、前記CPUコアあるいは前記メモリの電源のOFF/ONを行う、省電力制御装置。
(付記12)
前記メモリは不揮発メモリである、付記11記載の省電力制御装置。
(付記13)
前記プロセス部の特性は、前記プロセス部の負荷特性、あるいは、リソースの利用特性である、付記11または12記載の省電力制御装置。
(付記14)
前記プロセス部の負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記13記載の省電力制御装置。
(付記15)
前記リソースの利用特性は、ネットワーク入出力量である、付記13記載の省電力制御装置。
(付記16)
前記コア割当決定部は、前記CPU使用率の高い順に前記CPUコアに割り当てる、付記14記載の省電力制御装置。
(付記17)
前記コア割当決定部は、前記CPU使用率が平準化するように前記CPUコアに割り当てる、付記14記載の省電力制御装置。
(付記18)
前記コア割当決定部は、前記コンテキストスイッチの頻度の高い順に前記CPUコアに割り当てる、付記14記載の省電力制御装置。
(付記19)
前記コア割当決定部は、前記コンテキストスイッチの頻度が平準化するように前記CPUコアに割り当てる、付記14記載の省電力制御装置。
(付記20)
前記オペレーティングシステム部はハイパーバイザであり、前記プロセス部は仮想マシンである、付記11から19の内の1項記載の省電力制御装置。
(付記21)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御方法において、前記オペレーティングシステムは、前記プロセスの特性を収集し、前記プロセスの特性に基づいて前記プロセスに対する前記CPUコアの割り当てを決定し、前記割り当てに基づいてプロセスの実行をし、前記プロセスの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONをする、省電力制御方法。
(付記22)
前記メモリは不揮発メモリである、付記21記載の省電力制御方法。
(付記23)
前記プロセスの特性は、前記プロセスの負荷特性、あるいは、リソースの利用特性である、付記21または22記載の省電力制御方法。
(付記24)
前記プロセスの負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記23記載の省電力制御方法。
(付記25)
前記リソースの利用特性は、ネットワーク入出力量である、付記23記載の省電力制御方法。
(付記26)
前記コア割当ての決定は、前記CPU使用率の高い順に前記CPUコアに割り当てる、付記24記載の省電力制御方法。
(付記27)
前記コア割当ての決定は、前記CPU使用率が平準化するように前記CPUコアに割り当てる、付記24記載の省電力制御方法。
(付記28)
前記コア割当ての決定は、前記コンテキストスイッチの頻度の高い順に前記CPUコアに割り当てる、付記24記載の省電力制御方法。
(付記29)
前記コア割当ての決定は、前記コンテキストスイッチの頻度が平準化するように前記CPUコアに割り当てる、付記24記載の省電力制御方法。
(付記30)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記21から29の内の1項記載の省電力制御方法。
(付記31)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御プログラムにおいて、前記オペレーティングシステムに、前記プロセスの特性を収集する処理と、前記プロセスの特性に基づいて前記プロセスに対する前記CPUコアの割り当てを決定する処理と、前記割り当てに基づいてプロセスの実行をする処理と、前記プロセスの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONを行う処理と、を実行させる省電力制御プログラム。
(付記32)
前記メモリは不揮発メモリである、付記31記載の省電力制御プログラム。
(付記33)
前記プロセスの特性は、前記プロセスの負荷特性、あるいは、リソースの利用特性である、付記31または32記載の省電力制御プログラム。
(付記34)
前記プロセスの負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記33記載の省電力制御プログラム。
(付記35)
前記リソースの利用特性は、ネットワーク入出力量である、付記33記載の省電力制御プログラム。
(付記36)
前記コア割当てを決定する処理は、前記CPU使用率の高い順に前記CPUコアに割り当てる、付記34記載の省電力制御プログラム。
(付記37)
前記コア割当てを決定する処理は、前記CPU使用率が平準化するように前記CPUコアに割り当てる、付記34記載の省電力制御プログラム。
(付記38)
前記コア割当てを決定する処理は、前記コンテキストスイッチの頻度の高い順に前記CPUコアに割り当てる、付記34記載の省電力制御プログラム。
(付記39)
前記コア割当てを決定する処理は、前記コンテキストスイッチの頻度が平準化するように前記CPUコアに割り当てる、付記34記載の省電力制御プログラム。
(付記40)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記31から39の内の1項記載の省電力制御プログラム。
(付記41)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有し、前記オペレーティングシステムは、前記プロセスの負荷特性あるいは前記プロセスの割込み特性を収集するプロセス特性収集部と、前記負荷特性あるいは前記割込み特性に基づいて前記プロセスへの割込みの集約をする割込み集約部と、前記割込みの集約に基づいて割込みの実行をするプロセススケジューラと、を有し、前記割込みの実行に基づいて、前記CPUコアあるいは前記メモリの電源のOFF/ONを行う、省電力制御システム。
(付記42)
前記メモリは不揮発メモリである、付記41記載の省電力制御システム。
(付記43)
前記負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記41または42記載の省電力制御システム。
(付記44)
前記割込み特性は、デバイスへの入出力による割込み頻度、あるいは、前記プロセスの優先度である、付記41から43の内の1項記載の省電力制御システム。
(付記45)
前記割込み集約部は、前記負荷特性あるいは前記割込み特性の低いプロセスへの割込みを集約する、付記41から44の内の1項記載の省電力制御システム。
(付記46)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記41から45の内の1項記載の省電力制御システム。
(付記47)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステム部と、前記オペレーティングシステムで稼動する一つ以上のプロセス部と、を有し、前記オペレーティングシステム部は、前記プロセス部の負荷特性あるいは前記プロセス部の割込み特性を収集するプロセス特性収集部と、前記負荷特性あるいは前記割込み特性に基づいて前記プロセス部への割込みの集約をする割込み集約部と、前記割込みの集約に基づいて割込みの実行をするプロセススケジューラと、を有し、前記割込みの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONを行う、省電力制御装置。
(付記48)
前記メモリは不揮発メモリである、付記47記載の省電力制御装置。
(付記49)
前記負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記47または48記載の省電力制御装置。
(付記50)
前記割込み特性は、デバイスへの入出力による割込み頻度、あるいは、前記プロセスの優先度である、付記47から49の内の1項記載の省電力制御装置。
(付記51)
前記割込み集約部は、前記負荷特性あるいは前記割込み特性の低いプロセス部への割込みを集約する、付記47から50の内の1項記載の省電力制御装置。
(付記52)
前記オペレーティングシステム部はハイパーバイザであり、前記プロセス部は仮想マシンである、付記47から51の内の1項記載の省電力制御装置。
(付記53)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御方法において、前記オペレーティングシステムは、前記プロセスの負荷特性あるいは前記プロセスの割込み特性を収集し、前記負荷特性あるいは前記割込み特性に基づいて前記プロセスへの割込みの集約をし、前記割込みの集約に基づいて割込みの実行をし、前記割込みの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONをする、省電力制御方法。
(付記54)
前記メモリは不揮発メモリである、付記53記載の省電力制御方法。
(付記55)
前記負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記53または54記載の省電力制御方法。
(付記56)
前記割込み特性は、デバイスへの入出力による割込み頻度、あるいは、前記プロセスの優先度である、付記53から55の内の1項記載の省電力制御方法。
(付記57)
前記割込み集約は、前記負荷特性あるいは前記割込み特性の低いプロセスへの割込みを集約する、付記53から56の内の1項記載の省電力制御方法。
(付記58)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記53から57の内の1項記載の省電力制御方法。
(付記59)
複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御プログラムにおいて、前記オペレーティングシステムに、前記プロセスの負荷特性あるいは前記プロセスの割込み特性を収集する処理と、前記負荷特性あるいは前記割込み特性に基づいて前記プロセスへの割込みの集約をする処理と、前記割込みの集約に基づいて割込みの実行をする処理と、前記割込みの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONを行う処理と、を実行させる省電力制御プログラム。
(付記60)
前記メモリは不揮発メモリである、付記59記載の省電力制御プログラム。
(付記61)
前記負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、付記59または60記載の省電力制御プログラム。
(付記62)
前記割込み特性は、デバイスへの入出力による割込み頻度、あるいは、前記プロセスの優先度である、付記59から61の内の1項記載の省電力制御プログラム。
(付記63)
前記割込みの集約をする処理は、前記負荷特性あるいは前記割込み特性の低いプロセスへの割込みを集約する、付記59から62の内の1項記載の省電力制御プログラム。
(付記64)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記59から63の内の1項記載の省電力制御プログラム。
(付記65)
CPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有し、前記オペレーティングシステムは、前記プロセスをスケジューリングするプロセススケジューラと、前記スケジューリングに基づいて前記メモリの電源のON/OFFを行うメモリ電源制御部と、を有する、省電力制御システム。
(付記66)
前記メモリは不揮発メモリである、付記65記載の省電力制御システム。
(付記67)
前記プロセススケジューラは、前記スケジューリングに基づいて前記プロセスを実行する、付記65または66記載の省電力制御システム。
(付記68)
前記メモリ電源制御部は、前記プロセススケジューラが実行する前記プロセスに関わる前記メモリの電源をONする、付記65から67の内の1項記載の省電力制御システム。
(付記69)
前記メモリ電源制御部は、前記プロセススケジューラが実行する前記プロセスに関わる前記メモリ以外の前記メモリの電源をOFFする、付記65から68の内の1項記載の省電力制御システム。
(付記70)
前記オペレーティングシステムは、前記CPUがキャッシュに保持しているメモリのページ情報を収集する、キャッシュ情報収集部を有し、前記メモリ電源制御部は、前記キャッシュ情報収集部から前記ページ情報を取得し、前記ページ情報に基づいて前記メモリの電源をONする、付記65から69の内の1項記載の省電力制御システム。
(付記71)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記65から70の内の1項記載の省電力制御システム。
(付記72)
CPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステム部と、前記オペレーティングシステム部により稼動する一つ以上のプロセス部と、を有し、前記オペレーティングシステム部は、前記プロセス部をスケジューリングするプロセススケジューラと、前記スケジューリングに基づいて前記メモリの電源のON/OFFを行うメモリ電源制御部と、を有する、省電力制御装置。
(付記73)
前記メモリは不揮発メモリである、付記72記載の省電力制御装置。
(付記74)
前記プロセススケジューラは、前記スケジューリングに基づいて前記プロセス部を実行する、付記72または73記載の省電力制御装置。
(付記75)
前記メモリ電源制御部は、前記プロセススケジューラが実行する前記プロセス部に関わる前記メモリの電源をONする、付記72から74の内の1項記載の省電力制御装置。
(付記76)
前記メモリ電源制御部は、前記プロセススケジューラが実行する前記プロセス部に関わる前記メモリ以外の前記メモリの電源をOFFする、付記72から75の内の1項記載の省電力制御装置。
(付記77)
前記オペレーティングシステム部は、前記CPUがキャッシュに保持しているメモリのページ情報を収集する、キャッシュ情報収集部を有し、前記メモリ電源制御部は、前記キャッシュ情報収集部から前記ページ情報を取得し、前記ページ情報に基づいて前記メモリの電源をONする、付記72から76の内の1項記載の省電力制御装置。
(付記78)
前記オペレーティングシステム部はハイパーバイザであり、前記プロセス部は仮想マシンである、付記72から77の内の1項記載の省電力制御装置。
(付記79)
CPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御方法において、前記オペレーティングシステムは、前記プロセスをスケジューリングし、前記スケジューリングに基づいて前記メモリの電源のON/OFFをする、省電力制御方法。
(付記80)
前記メモリは不揮発メモリである、付記79記載の省電力制御方法。
(付記81)
前記プロセスのスケジューリングは、前記スケジューリングに基づいて前記プロセスを実行する、付記79または80記載の省電力制御方法。
(付記82)
前記メモリの電源のON/OFFは、前記スケジューリングに基づいて実行される前記プロセスに関わる前記メモリの電源をONする、付記79から81の内の1項記載の省電力制御方法。
(付記83)
前記メモリの電源のON/OFFは、前記スケジューリングに基づいて実行される前記プロセスに関わる前記メモリ以外の前記メモリの電源をOFFする、付記79から82の内の1項記載の省電力制御方法。
(付記84)
前記オペレーティングシステムは、前記CPUがキャッシュに保持しているメモリのページ情報を収集し、前記メモリの電源の制御は、前記ページ情報を取得し、前記ページ情報に基づいて前記メモリの電源をONする、付記79から83の内の1項記載の省電力制御方法。
(付記85)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記79から84の内の1項記載の省電力制御方法。
(付記86)
CPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御プログラムにおいて、前記オペレーティングシステムに、前記プロセスをスケジューリングする処理と、前記スケジューリングに基づいて前記メモリの電源のON/OFFを行う処理と、を実行させる省電力制御プログラム。
(付記87)
前記メモリは不揮発メモリである、付記86記載の省電力制御プログラム。
(付記88)
前記プロセスをスケジューリングする処理は、前記スケジューリングに基づいて前記プロセスを実行する、付記86または87記載の省電力制御プログラム。
(付記89)
前記メモリの電源のON/OFFを行う処理は、前記スケジューリングに基づいて実行される前記プロセスに関わる前記メモリの電源をONする、付記86から88の内の1項記載の省電力制御プログラム。
(付記90)
前記メモリの電源のON/OFFを行う処理は、前記スケジューリングに基づいて実行される前記プロセスに関わる前記メモリ以外の前記メモリの電源をOFFする、付記86から89の内の1項記載の省電力制御プログラム。
(付記91)
前記オペレーティングシステムに、前記CPUがキャッシュに保持しているメモリのページ情報を収集する処理を実行させ、前記メモリの電源のON/OFFを行う処理は、前記ページ情報を取得し、前記ページ情報に基づいて前記メモリの電源をONする、付記86から90の内の1項記載の省電力制御プログラム。
(付記92)
前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、付記86から91の内の1項記載の省電力制御プログラム。
11 CPU
12 I/Oデバイス
13 メモリ部
14、15 OS
141 プロセス特性収集部
142 コア割当決定部
143 プロセススケジューラ
144 割込み集約部
145 メモリ電源制御部
146 キャッシュ情報収集部
150、151~15n プロセス
16、17 ハイパーバイザ
161 VM特性収集部
162 コア割当決定部
163 VMスケジューラ
164 割込み集約部
165 メモリ電源制御部
166 キャッシュ情報収集部
171~17n 仮想マシン(VM)
18 CPUコア
19 メモリ
Claims (10)
- 複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有し、前記オペレーティングシステムは、前記プロセスの特性を収集するプロセス特性収集部と、前記プロセスの特性に基づいて前記プロセスに対する前記CPUコアの割り当てを決定するコア割当決定部と、前記割り当てに基づいて前記プロセスの実行をするプロセススケジューラと、を有し、前記プロセスの実行に基づいて、前記CPUコアあるいは前記メモリの電源のOFF/ONを行う、省電力制御システム。
- 前記メモリは不揮発メモリである、請求項1記載の省電力制御システム。
- 前記プロセスの特性は、前記プロセスの負荷特性、あるいは、リソースの利用特性である、請求項1または2記載の省電力制御システム。
- 前記プロセスの負荷特性は、CPU使用率、あるいは、キャッシュヒット率、あるいは、コンテキストスイッチの頻度である、請求項3記載の省電力制御システム。
- 前記コア割当決定部は、前記CPU使用率の高い順に前記CPUコアに割り当てる、請求項4記載の省電力制御システム。
- 前記コア割当決定部は、前記コンテキストスイッチの頻度の高い順に前記CPUコアに割り当てる、請求項4記載の省電力制御システム。
- 前記オペレーティングシステムはハイパーバイザであり、前記プロセスは仮想マシンである、請求項1から6の内の1項記載の省電力制御システム。
- 複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステム部と、前記オペレーティングシステム部によって稼動する一つ以上のプロセス部と、を有し、前記オペレーティングシステム部は、前記プロセス部の特性を収集するプロセス特性収集部と、前記プロセス部の特性に基づいて前記プロセス部に対する前記CPUコアの割り当てを決定するコア割当決定部と、前記割り当てに基づいてプロセス部の実行をするプロセススケジューラと、を有し、前記プロセス部の実行に基づいて、前記CPUコアあるいは前記メモリの電源のOFF/ONを行う、省電力制御装置。
- 複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御方法において、前記オペレーティングシステムは、前記プロセスの特性を収集し、前記プロセスの特性に基づいて前記プロセスに対する前記CPUコアの割り当てを決定し、前記割り当てに基づいてプロセスの実行をし、前記プロセスの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONをする、省電力制御方法。
- 複数のCPUコアを有するCPUとメモリとを有する物理マシンと、前記物理マシンで稼動するオペレーティングシステムと、前記オペレーティングシステムで稼動する一つ以上のプロセスと、を有する省電力制御システムの省電力制御プログラムにおいて、前記オペレーティングシステムに、前記プロセスの特性を収集する処理と、前記プロセスの特性に基づいて前記プロセスに対する前記CPUコアの割り当てを決定する処理と、前記割り当てに基づいてプロセスの実行をする処理と、前記プロセスの実行に基づいて前記CPUコアあるいは前記メモリの電源のOFF/ONを行う処理と、を実行させる省電力制御プログラムを記録する記録媒体。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US14/904,773 US20160170474A1 (en) | 2013-08-02 | 2014-07-22 | Power-saving control system, control device, control method, and control program for server equipped with non-volatile memory |
JP2015529348A JPWO2015015756A1 (ja) | 2013-08-02 | 2014-07-22 | 不揮発性メモリ搭載サーバの省電力制御システム、制御装置、制御方法および制御プログラム |
Applications Claiming Priority (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2013-161304 | 2013-08-02 | ||
JP2013161302 | 2013-08-02 | ||
JP2013161304 | 2013-08-02 | ||
JP2013161303 | 2013-08-02 | ||
JP2013-161303 | 2013-08-02 | ||
JP2013-161302 | 2013-08-02 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2015015756A1 true WO2015015756A1 (ja) | 2015-02-05 |
Family
ID=52431316
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2014/003848 WO2015015756A1 (ja) | 2013-08-02 | 2014-07-22 | 不揮発性メモリ搭載サーバの省電力制御システム、制御装置、制御方法および制御プログラム |
Country Status (3)
Country | Link |
---|---|
US (1) | US20160170474A1 (ja) |
JP (1) | JPWO2015015756A1 (ja) |
WO (1) | WO2015015756A1 (ja) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10387178B2 (en) * | 2014-10-29 | 2019-08-20 | Red Hat Israel, Ltd. | Idle based latency reduction for coalesced interrupts |
US10976801B2 (en) * | 2018-09-20 | 2021-04-13 | Intel Corporation | System, apparatus and method for power budget distribution for a plurality of virtual machines to execute on a processor |
KR20200097579A (ko) | 2019-02-08 | 2020-08-19 | 삼성전자주식회사 | 프로세스 스케줄링을 위한 전자 장치, 저장 매체 및 방법 |
JP7387308B2 (ja) * | 2019-06-27 | 2023-11-28 | キヤノン株式会社 | 情報処理装置および情報処理装置の制御方法 |
JP2023021662A (ja) * | 2021-08-02 | 2023-02-14 | 富士通株式会社 | プログラムおよび管理方法 |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007122337A (ja) * | 2005-10-27 | 2007-05-17 | Sonac Kk | 演算装置 |
JP2010044460A (ja) * | 2008-08-08 | 2010-02-25 | Renesas Technology Corp | 電源制御装置、計算機システム、電源制御方法、電源制御プログラムおよび記録媒体 |
JP2010277171A (ja) * | 2009-05-26 | 2010-12-09 | Hitachi Ltd | タスク割当装置、および、タスク割当方法 |
JP2012089154A (ja) * | 2005-09-30 | 2012-05-10 | Coware Inc | マルチコアアーキテクチャにおけるスケジューリング |
JP2013513189A (ja) * | 2009-12-16 | 2013-04-18 | クアルコム,インコーポレイテッド | 仮想化されたシステムにおける中央処理装置電力を制御するためのシステムおよび方法 |
US8490103B1 (en) * | 2007-04-30 | 2013-07-16 | Hewlett-Packard Development Company, L.P. | Allocating computer processes to processor cores as a function of process utilizations |
Family Cites Families (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6889277B2 (en) * | 2002-04-18 | 2005-05-03 | Sun Microsystems, Inc. | System and method for dynamically tuning interrupt coalescing parameters |
CN1666180B (zh) * | 2002-07-03 | 2010-05-26 | Nxp股份有限公司 | 多处理器计算机系统及方法 |
JP4213572B2 (ja) * | 2003-11-28 | 2009-01-21 | 株式会社東芝 | 電子機器およびプロセッサ速度制御方法 |
US7650472B2 (en) * | 2005-07-12 | 2010-01-19 | Electronics And Telecommunications Research Institute | Method for reducing memory power consumption |
JP2008257578A (ja) * | 2007-04-06 | 2008-10-23 | Toshiba Corp | 情報処理装置、スケジューラおよび情報処理置のスケジュール制御方法 |
US7882319B2 (en) * | 2007-09-27 | 2011-02-01 | Oracle America Inc. | Method and system for memory management |
JP5182792B2 (ja) * | 2007-10-07 | 2013-04-17 | アルパイン株式会社 | マルチコアプロセッサ制御方法及び装置 |
US8010822B2 (en) * | 2008-03-28 | 2011-08-30 | Microsoft Corporation | Power-aware thread scheduling and dynamic use of processors |
US8296773B2 (en) * | 2008-06-30 | 2012-10-23 | International Business Machines Corporation | Systems and methods for thread assignment and core turn-off for integrated circuit energy efficiency and high-performance |
JP5245722B2 (ja) * | 2008-10-29 | 2013-07-24 | 富士通株式会社 | スケジューラ、プロセッサシステム、プログラム生成装置およびプログラム生成用プログラム |
US8321614B2 (en) * | 2009-04-24 | 2012-11-27 | Empire Technology Development Llc | Dynamic scheduling interrupt controller for multiprocessors |
US8984523B2 (en) * | 2009-05-26 | 2015-03-17 | Telefonaktiebolaget L M Ericsson (Publ) | Method for executing sequential code on the scalable processor at increased frequency while switching off the non-scalable processor core of a multicore chip |
EP2551768A4 (en) * | 2010-03-24 | 2013-10-09 | Fujitsu Ltd | MULTICLE SYSTEM AND STARTING PROCEDURE |
JP5408356B2 (ja) * | 2010-08-10 | 2014-02-05 | 富士通株式会社 | スケジューラプログラム、マルチコアプロセッサシステムおよびスケジューリング方法 |
US9063730B2 (en) * | 2010-12-20 | 2015-06-23 | Intel Corporation | Performing variation-aware profiling and dynamic core allocation for a many-core processor |
US9569279B2 (en) * | 2012-07-31 | 2017-02-14 | Nvidia Corporation | Heterogeneous multiprocessor design for power-efficient and area-efficient computing |
US9329900B2 (en) * | 2012-12-28 | 2016-05-03 | Intel Corporation | Hetergeneous processor apparatus and method |
US9342374B2 (en) * | 2013-06-28 | 2016-05-17 | Dell Products, L.P. | Method of scheduling threads for execution on multiple processors within an information handling system |
JP2016523399A (ja) * | 2013-06-28 | 2016-08-08 | インテル コーポレイション | エネルギー効率的なモバイルプラットフォームのための適応的割り込みコアレッシング |
US9424092B2 (en) * | 2014-09-26 | 2016-08-23 | Microsoft Technology Licensing, Llc | Heterogeneous thread scheduling |
-
2014
- 2014-07-22 JP JP2015529348A patent/JPWO2015015756A1/ja active Pending
- 2014-07-22 WO PCT/JP2014/003848 patent/WO2015015756A1/ja active Application Filing
- 2014-07-22 US US14/904,773 patent/US20160170474A1/en not_active Abandoned
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012089154A (ja) * | 2005-09-30 | 2012-05-10 | Coware Inc | マルチコアアーキテクチャにおけるスケジューリング |
JP2007122337A (ja) * | 2005-10-27 | 2007-05-17 | Sonac Kk | 演算装置 |
US8490103B1 (en) * | 2007-04-30 | 2013-07-16 | Hewlett-Packard Development Company, L.P. | Allocating computer processes to processor cores as a function of process utilizations |
JP2010044460A (ja) * | 2008-08-08 | 2010-02-25 | Renesas Technology Corp | 電源制御装置、計算機システム、電源制御方法、電源制御プログラムおよび記録媒体 |
JP2010277171A (ja) * | 2009-05-26 | 2010-12-09 | Hitachi Ltd | タスク割当装置、および、タスク割当方法 |
JP2013513189A (ja) * | 2009-12-16 | 2013-04-18 | クアルコム,インコーポレイテッド | 仮想化されたシステムにおける中央処理装置電力を制御するためのシステムおよび方法 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2015015756A1 (ja) | 2017-03-02 |
US20160170474A1 (en) | 2016-06-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5885920B2 (ja) | 仮想cpuベースの、周波数、及び電圧制御 | |
JP5075274B2 (ja) | 電力認識スレッドスケジューリングおよびプロセッサーの動的使用 | |
JP5469940B2 (ja) | 計算機システム、仮想計算機モニタ及び仮想計算機モニタのスケジューリング方法 | |
US8892916B2 (en) | Dynamic core pool management | |
EP2430538B1 (en) | Allocating computing system power levels responsive to service level agreements | |
WO2015015756A1 (ja) | 不揮発性メモリ搭載サーバの省電力制御システム、制御装置、制御方法および制御プログラム | |
Sridharan et al. | Holistic run-time parallelism management for time and energy efficiency | |
US9411649B2 (en) | Resource allocation method | |
TWI489265B (zh) | 根據效率評等方案包含在處理器的多頻域間平衡電力的能量效率及節能方法,設備與系統 | |
EP3239844B1 (en) | Processing method and device for memory page in memory | |
WO2010138031A1 (en) | Method and scheduler in an operating system | |
EP2430541A1 (en) | Power management in a multi-processor computer system | |
KR20170054485A (ko) | 운영체제에서의 휴리스틱 프로세서 전력 관리 | |
CN105117285B (zh) | 一种基于移动虚拟化系统的非易失性存储器调度优化方法 | |
Li et al. | Improving preemptive scheduling with application-transparent checkpointing in shared clusters | |
US10768684B2 (en) | Reducing power by vacating subsets of CPUs and memory | |
TW201541347A (zh) | 多核心處理器系統及其排程方法 | |
Sudan et al. | Tiered memory: An iso-power memory architecture to address the memory power wall | |
JP2010039802A (ja) | マルチプロセッサシステム、スケジューリング方法およびそのプログラム | |
Xilong et al. | An energy-efficient virtual machine scheduler based on CPU share-reclaiming policy | |
WO2015146104A1 (ja) | 仮想マシンシステムおよびその制御方法およびその制御プログラム記録媒体 | |
Liao et al. | A novel memory allocation scheme for memory energy reduction in virtualization environment | |
Xiao et al. | Energy-efficiency enhanced virtual machine scheduling policy for mixed workloads in cloud environments | |
HoseinyFarahabady et al. | Data-intensive workload consolidation in serverless (Lambda/FaaS) platforms | |
Thiam et al. | Energy efficient cloud data center using dynamic virtual machine consolidation algorithm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14831842 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2015529348 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14904773 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14831842 Country of ref document: EP Kind code of ref document: A1 |