JPH06100949B2 - Power-on control method for computer system - Google Patents

Description

DETAILED DESCRIPTION [Outline] A plurality of devices forming a computer system are divided into partial pieces, and the divided partial pieces are used as power-on units.
Regarding the power-on control method in the power supply control system that connects to each port and turns on the power in units of the ports, a plurality of devices are connected for the purpose of shortening the power-on time of the computer system. Means for storing the power load characteristic of each port, and based on the stored power load characteristic, the time is shifted to Tk and each port Ni (i = 1,2,3
...) is time-sequentially calculated when a power-on instruction is given to the power-on instruction, and the calculated time-series power load characteristics are
When the patented power capacity of the power supply device (CVCF) and the allowable load fluctuation range are not exceeded, the closing instruction time Tki (i = 1,2,
3, ...), the port Ni is registered in the power-on sequence table, and the power-supply control system is configured to power on the port unit according to the registered power-on sequence table.

[Industrial application field]

The present invention divides a plurality of devices constituting a computer system into partial pieces, connects the divided partial pieces to each port as a power-on unit, and turns on the power by each port as a unit. The present invention relates to a power-on control method in a power control system that performs

With the recent increase in the scale of computer systems, the number of input / output devices and the like in the computer system has increased, and as a result, the power-on time of the computer system has been steadily increasing.

Under these circumstances, there has been a demand for a power-on control method that can shorten the power-on time in the computer system.

[Problems to be solved by conventional technology and invention]

FIG. 3 is a diagram for explaining a conventional power-on control method, in which (a) shows a configuration example of a computer system and (b) shows a conventional power load characteristic at power-on. .

Generally, when the computer system is powered on, it is necessary to control so that the maximum capacity (KVA _CMAX ) of the power supply device (CVCF etc.) 5 and the allowable load fluctuation range (KVA _CTRN ) are not _exceeded .

On the other hand, the power consumption load characteristics of each device (for example, an input / output device) 2 are different for each device ( ₀₀ , ₀₁ , ...) 2, but in general, a transient state lasts for a certain period of time and then becomes steady. It becomes a state. Since the power consumption in the transition period exceeds the power consumption in the steady state, the power-on instruction cannot be given to all the input / output devices 2 at once.

Therefore, in the prior art, in order to avoid the overlap in the transitional period, the ports (# 0 to #n) 11 to which a plurality of devices 2 are connected are turned on and off as shown in FIG. It has been used as a control unit.

In other words, after giving the power-on instruction to the input / output device connected to the port (# n-1), wait for the response of the power-on completion from that port (# n-1), and then supply the power to the next port (#n). We have taken the method of giving the input instruction. At this time, a plurality of devices connected to each port (#n) are purely powered on as shown in FIG. Where KVAnm _P : maximum power load capacity of m-th device of port #n (peak value in transient state) KVAnm _N : steady-state power load capacity of m-th device of port #n (maximum value in steady state) ) Tnm _P : Time until the power consumption of the m-th device at port #n reaches its peak Tnm _A : Tnm _P + α, α: Tnm _P error range Tnm _B : Tnm _{P −} α, α: Tnm _P error Range Therefore, at some point in time, the number of devices being powered on is the number of devices connected to the port (#n), and at a certain time, only one device is being powered on. Although guaranteed, the power-on time is the sum of the ports {see Fig. (B)}, so there was a drawback in that it takes a very long time to power on a large-scale system.

In view of the above-mentioned conventional drawbacks, the present invention divides a plurality of devices constituting a computer system into partial pieces, and connects the divided partial pieces to each port as a power-on unit, It is an object of the present invention to provide a power-on control method for a power-on control system in which power is turned on in units of ports, which shortens the power-on time for each of the plurality of devices.

[Means for solving problems]

The above problems are solved by the power-on control system having the following configuration.

A plurality of devices constituting a computer system are divided into partial pieces, and the divided partial pieces are used as a power-on unit.
In a power supply control system for connecting to each port and turning on the power in units of the port, a means for inputting and storing the power load characteristic of each port, and based on the stored power load characteristic, The power load characteristics when the power-on instruction is given to each port Ni (i = 1,2,3 ...) by shifting the time to Tk are calculated in time series, and the calculated time-series power load characteristics are calculated. If the maximum power capacity of the power supply device and the allowable load fluctuation range are not exceeded,
At the input instruction time Tki (i = 1,2,3, ...), the port Ni
Is registered in the power-on sequence table, and the power supply control system is configured to power on in units of ports according to the registered power-on sequence table.

[Action]

That is, according to the present invention, a plurality of devices forming a computer system are divided into partial pieces, and the divided partial pieces are connected as power-on units to respective ports, and the ports are connected to each other. When the power is turned on as a unit, the power load characteristics of each device or each port are input in advance from an input / output device such as a display device and stored in, for example, an external storage device. Based on the specified power load characteristics, shift the timing between each port so that the peaks of power consumption for each device or each port do not overlap, and from the maximum performance of the power supply device (CVCF etc.) The number of ports to which a power-on instruction can be given and the power-on sequence thereof are calculated and registered in the power-on sequence table, and power is turned on for each port in accordance with the registered power-on sequence table. Therefore, it is possible to give power-on instructions to each port one after another without waiting for the completion of power-on of the previous port, and it is possible to greatly reduce the power-on time in a large-scale computer system. There is an effect that can be done.

〔Example〕

 Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram showing an embodiment of the present invention, in which (a)
Shows an example of a system configuration, (b) is a flow chart of a power-on control method of the present invention, and (c) shows an example of a control table necessary for performing power-on control of the present invention. FIG. 2 is an operation time chart when the power is turned on according to the present invention. The power load characteristic table 30 in FIG.
The type definition table 31 for each port of the device connected to each port, the power-on sequence table 32, and the like are necessary means for implementing the present invention. The same reference numerals indicate the same objects throughout the drawings.

The power-on control method of the present invention will be described below with reference to FIGS. 1 and 2.

First, from the input section of the display device 4 of the service processor (SVP) 13 shown in the system configuration of FIG. 1A, the input / output device types A, B, C, ... Connected to the computer system are connected. Is input and stored in the external storage device 3 as the power load characteristic table 30.

Next, the type of the device 2 connected to each port 11 is defined and stored in the external storage device 3 as a type definition table 31.

From the data of the power load characteristic table 30 and the port-specific type definition table 31 thus obtained,
For example, the power consumption when the power-on instruction is given to each port after shifting by 1 second is obtained for each time T _K i (i = 0,1,2 ...), and the power supply device (CVCF etc.) 5 The time series is checked to see if it falls within the permissible range, and if it is recognized that the permissible range is not exceeded, the relevant port #i is repeatedly registered at the input instruction time T _K i, The power input sequence table 32 is generated. {Refer to steps 100 to 107 in Fig. 1 (b)} The process of generating the power input sequence table 32 will be described with a specific example.

If port # 0 is turned on at the start time, both KVAnm _P and KVAnm _N described above are “0” at that time.

At T _K0 one second later, when the port # 1 is turned on, only one type A device of the port # 0 is turned on, and then at the port # 0, for example, every 2 seconds.
The power of the relevant device shall be turned on in sequence.

Therefore, one second later (T _K0 ), the first device (X
OVA) has not reached the peak, so both KVAnm _P and KVAnm _N are '0'.

At T _K1 one second later, assuming that port # 2 is turned on, the equipment (XO) at port # 0 only reaches its peak, so KVAnm _P = 1.3KVA, KVAnm _N = 0.6KVA. Is calculated, and this value is the allowable load capacity KVA of the power supply device 5.
_It is checked whether _CMAX and allowable load fluctuation range KVA _CTRN or less.

The check conditions at this time are as follows, for example.

However, at time T _K3 , the peaks of the device of port # 0 (type A) and the device of port # 1 (type B) overlap and become ΣKVAnm _P = 2.5KVA, so this peak value is the power supply device 5 If the allowable load fluctuation range KVA _CTRN is exceeded, the power of port # 1 cannot be turned on at time T _K0 , so it is necessary to shift it by 1 second, for example.

Thereafter, in the same procedure, the total power consumption for each time (T _K i) is calculated, and the port number that can be input is registered at each time T ki.

In the above example, since the registration conditions are KVA _CMAX and KVA _CTRN or less, the port # 0, is registered at the start time of the power-on sequence table 32, but the time T _K0 is the same as # 1. registration is suppressed, it is shifted to the next T _K1.

In this way, when the power-on sequence table 32 is generated, the service processor (SVP) 13 controls the power control device 12 to turn on the power according to the sequence of the power-on port numbers indicated in the table 32. To do.

As a result, the peak value of power consumption at each time is less than or equal to the power load fluctuation range KVA _CTRN , and the total power consumption at that time is KVA, without waiting for the power-on completion signal of the previous port. It will be below _CMAX .

In the above example, one port is turned on at time T _K i, but the power consumption when a plurality of ports are turned on may be calculated. There is no such thing.

The time chart of FIG. 2 is an example of the power load characteristics when two ports (# n-1, #n) are turned on at the same time, and is shown by a solid line in comparison with the turn-on time of the conventional method shown by the dotted line. It can be seen that the shown dosing time according to the invention is effectively shortened. Of course, the power consumption at each time satisfies the allowable load capacity (maximum capacity) of the power supply device and the allowable load fluctuation range.

As described above, the present invention obtains, for example, the sum of peak values of power consumption per second and the sum of steady values up to that time, based on the power load characteristics of each device and each port. The number of ports that can be powered on at the same time and the order in which they are turned on are calculated so that the value is within the allowable load capacity (maximum capacity) and the allowable load fluctuation range of the power supply device, and based on the calculated order The feature is that the power-on instruction of the next port can be given without waiting for the power-on completion signal of the previous port.

〔The invention's effect〕

As described above in detail, the power-on control method for a computer system according to the present invention divides a plurality of devices constituting the computer system into parts, and divides the divided parts into power sources. In order to connect to each port as a power-on unit and power on in units of that port, the power load characteristics of each device or each port are input in advance from an input / output device such as a display device. , For example, it is stored in an external storage device, and based on the stored power load characteristics, the timing is shifted between ports so that the peaks of power consumption for each device or for each port do not overlap, From the maximum performance of the power supply (CVCF etc.), the number of ports that can simultaneously give power-on instructions and the power-on sequence are calculated and registered in the power-on sequence table. Therefore, since the power is turned on for each port, it is possible to give power-on instructions to each port one after another without waiting for the completion of the power-on of the previous port. This has the effect of significantly reducing the power-on time in the machine system.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of the present invention, FIG. 2 is an operation time chart at power-on according to the present invention, and FIG. 3 is a diagram explaining a conventional power-on control system. In the drawings, 1 is the main unit 11 is the port (# 0~ # n), 12 power supply control unit, 13 a service processor (SVP), 2 is input apparatus, or device ₍₀₀ ~n _2), 3 is External storage device, 4 is a display device, 5 is a power supply device (CVCF, etc.), # 0,1,2, ... is a port number, A, B, C, ... is the type of device 2, _TK0 , _K1 ... The power-on time, is shown respectively.

Claims (1)

[Claims] 1. A plurality of devices (2) constituting a computer system are divided into partial pieces, and the divided partial pieces are connected to respective ports (11) as power-on units.
In a power supply control system (12, 13) for turning on power in units of the port (11), means (3, for inputting and storing power load characteristics of each device (2) or each port (11) 4) and the power load when the time is shifted to Tk and a command to turn on the power is given to each port Ni (i = 1,2,3 ...) (11) based on the stored power load characteristics. The characteristics are calculated in time series, and when the calculated time series power load characteristics do not exceed the maximum power capacity of the power supply device (5) and the allowable load fluctuation range, the closing instruction time Tki (i = 1,2,3, ...) is registered in the power-on sequence table (32) at the above-mentioned port Ni (i = 1,2,3, ...) (11), and the registered power registration sequence table (32) is registered. In accordance with the above, the power supply control device (12) controls to turn on the power supply in units of the ports (11). Input control system.