KR980011219A - Disk drives for portable computers with adaptive demand-based power management - Google Patents

Abstract

Data recording disk drives used in battery-powered portable computers have several modes of power saving operation. The power save modes are entered after the time calculated after the previous data read or write command has elapsed. The time to enter sleep mode is calculated based on the real-time workload of the computer user, and thus continues to change during operation of the disk drive. The disk drive detects the user's current workload by calculating the frequency of disk drive accesses and based on the history of this workload, the disk drive determines which of the plurality of power save modes is appropriate and when to enter the power save mode .

A read or write access of each disk drive is detected and used to calculate the current access frequency. The current access frequency is compared to a previously calculated and continuously updated critical frequency. The threshold frequency represents a uniform or sporadic access pattern and is calculated from a formula that includes adjustable gain factors.

And enters the appropriate power saving modes when the current access frequency is less than the threshold frequency. The intermediate power saving modes may be omitted based on the detected access pattern. Disk drives can also dynamically adapt to situations where the workload changes, thereby saving more energy without compromising performance. This is accomplished by adjusting the gain factors that determine the actual performance of the system, which changes the threshold frequency. The disk drive also determines when to exit sleep mode without having to wait indefinitely for user access. Sleep mode can also be customized for different disk drive products. In this case, the power saving modes need not follow the standard for setting the number of fixed time or power saving modes. New user-selected parameters, such as ON / OFF and performance and energy goals, can be used in place of the fixed-mode entry time. These parameters adjust the critical frequency by adjusting the gain factors. This changes the formula used to determine when to enter sleep mode.

Description

Disk drives for portable computers with adaptive demand-based power management

FIG. 1 is a block diagram illustrating a disk drive and computer system showing various elements consuming energy in a power source and disk drive; FIG.

FIG. 2 is a flow diagram illustrating a process for counting accesses to one or more power-consuming disk drive components within a predetermined time window. FIG.

Figure 3 is a block diagram illustrating a microprocessor in conjunction with a ring buffer in which access densities are accumulated.

Figure 4 shows the disk drive entering sleep mode. A process for calculating an access frequency and a process for comparing the frequency with a threshold frequency to determine an access frequency.

5 shows a flow illustrating the process of calculating a threshold frequency using a previous access frequency stored in a ring buffer;

6 is a flow chart illustrating a process for calculating a threshold frequency using a previous access frequency stored in a ring buffer;

FIG. 7 is a graph showing time steps for entering and exiting the power saving mode; FIG.

FIG. 8 is a block diagram of a third embodiment of the present invention, with a counter for counting time windows between entry and exit to a power saving mode.

FIG. 9 is a graph showing the cumulative energy / response penalty as a function of time to account for adjustment of the gain factors when the trip level is exceeded.

10 is a flow diagram illustrating a process for calculating an energy / response penalty and an opportunity loss penalty at the end of a power save mode;

FIG. 11 is a flow diagram illustrating the process of adjusting the gain factors and calculating the cumulative opportunity loss penalty when the trip level is exceeded.

DESCRIPTION OF THE REFERENCE NUMERALS

1: Spindle driver 2: VCM driver

3: pre-amp / write driver 4: data write channel

5: Spindle control electronic circuit part 6: Servo control electronic circuit part

7: disk controller electronic circuit part 9: microprocessor

10: buffer memory 11: interface module

12: power control module 13: computer interface controller

19: control electronic circuit part 20: power source

30: Rotary voice coil motor 31: Actuator

32: spindle motor 33: data head

40: disk drive 41: computer

70, 71: bus

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention generally relates to disk drives used in a portable computer that supplies power to a battery, such as a laptop computer or a notebook computer, and more particularly, To disk drives.

Portable computers can only operate for a few hours before they run out of battery power, the main power consumption being the hard disk. Important power management techniques for hard disk drives used in portable computers include several reduced-power operating modes or power-save operating modes, which can be used after a read or write operation of a disk drive After the lapse of the set fixed time period, each mode is entered. For example, at the end of a fixed time period after the user lastly records data on the hard disk, the read / write head moves to the parking position and the spindle motor of the disk drive stops. Then, when the user again accesses the disk drive, the spindle motor rotates and the heads move across the disk to the appropriate data track to read or write data. The biggest disadvantage of such a power saving mode is that it takes time to terminate the power saving mode. During this time, the user has to wait forever. This has a significant impact on the morale of the computer. Generally, the length of a fixed time period is set by a computer user using software.

This conventional power saving technique has a problem that the user does not have the data necessary for selecting a good value as a fixed time period. That is, the user has only limited knowledge of the access type and does not have information about the energy and performance parameters of the disk drive. The user is isolated from the information of the internal access type of the disk drive by the hardware and software in the system. It is not easy to compromise adequately between energy and performance, since fixed time to enter sleep mode is not authorized for the user workload. The user must change the fixed time in anticipation of the workload, and if the fixed time is chosen too short or too long, the opposite effect on performance and energy consumption can be expected.

The entry time of short mode saves energy when the access type is a form of consecutive inactive states followed by a long inactive state. However, excess energy is used when the inactivity state period is close to the entry time of the mode. Also, after entering the mode after a short period of inactivity, it may affect performance and generally result in more frequent access delays due to mode recovery time.

If the period of the longer inactive state is less common than the period of the shorter inactive state, the entry time of the long mode reduces the impact on the computer performance relatively and reduces the tendency to use excessive energy.

The optimal time for a fixed time during a particular user workload may not be constant. Moreover, the workload can be changed without the user realizing it, because it is caused by the application software running by the user.

In fact, users want to choose between energy consumption and computer performance, not between fixed mode entry times. The user chooses a fixed mode entry time only to speculate in achieving energy or performance goals. Obviously, the disk drive should receive energy and performance goals from the user. With these goals, the disk drive can take any suitable way to meet these goals. This also allows the disk drive to use more power saving modes. This is because the user does not need to know the specific power saving modes operating in the drive unless the user chooses a fixed time for each mode entry.

As described above, in the disk drive, it is possible to detect changes in the workload, actively adapt to changes in the workload, and determine energy and performance goals instead of the fixed time to determine when to enter and exit the power- There is a need for a method and system for achieving power management.

Accordingly, an object of the present invention is to provide a disk drive that performs power management for predicting subsequent user requests from a past access history of a disk drive to determine the entry and exit time of the power save mode. Such a disk drive does not know how much performance and energy consumption costs the user is involved in entering and exiting the power save mode, so that the current user-selected pre-determined or mode-entry time or fixed mode- . The disk drive has information about the energy break-even time and recovery time associated with the power save mode. Energy Brakes - The time it takes to keep a disk drive in a certain power-saving mode so that extra energy consumed while returning from power-saving mode is balanced with reduced energy consumption while in power-save mode. The return time is the time it takes the disk drive to return from the sleep mode to the active state. The disk drive keeps track of the trajectory of the access type, i.e., the history of various requests to read / write data or to move the actuator. This causes the disk drive to detect the current user's workload and determine which of the plurality of power saving modes is appropriate and when to enter the power save mode.

In a preferred embodiment, when access to each disk drive is detected, it is used to calculate the current access frequency. The current access frequency is compared to that previously calculated and continues to update the threshold frequency. The threshold frequency represents a uniform or sporadic access form and is calculated from a formula that includes adjustable gain factors. An appropriate power saving mode while the disk drive is running enters when the current access frequency is less than the threshold frequency. The intermediate power saving mode based on the detected access type can be omitted. Disk drives can also actively adapt to changing workload conditions, thereby saving a lot of energy without compromising performance. This is achieved by adjusting the gain factors associated with the actual performance of the system. The disk drive also determines when to exit sleep mode without having to wait for user access.

In the present invention, a number of power saving modes in which a user does not need to memorize a trajectory of a situation change are disclosed. The power saving modes can also be suitably applied to different disk drives without having to follow certain standards for setting the number of fixed time or power saving modes. New user-selected parameters such as ON / OFF and performance and energy goals can be used in place of the fixed mode-entry time. These parameters adjust the critical frequency by adjusting the gain factors. This changes the formula used to determine when to enter sleep mode.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

FIG. 1 is a block diagram illustrating various components used in power management for a disk drive of the present invention. In general, the disk drive 40 is a hard disk drive contained within a housing of a laptop computer, such as the computer 41. The disk drive 40 includes one or more disks, such as a conventional disk 34, and the disks are attached to the spindle motor 32 and rotated. The data head 33 is connected to an actuator 31 which is generally operated by a rotary voice coil motor (VCM) 30. The spindle motor 32 is driven by the spindle driver 1 and the spindle control electronics 5. The servo control electronics section 6 is used to locate the head 33 on each data track of the disk 33 and is connected to the VCM driver 2 which supplies current to the VCM 30. [ 1 also shows a crash stop 37 for the actuator 31 and a load / unload ramp (L / UL) 38. The load / If the disk drive 40 is of the CSS type (the Contact Start / Stop type) in which the head 33 remains in the disk 34 when the spindle motor is not rotating, the actuator is driven by the current supplied to the VCM 30 The head 33 is driven to the crash stop 37 and then the head 33 stops on a landing zone where there is no data close to the inside diameter (ID) of the disk. When the disk drive 40 is of the L / UL type, the actuator 31 is driven to suspend the head 33 when the spindle motor is temporarily stopped, and the suspension 33 ). The ID landing area in the CSS disk drive and the ramp in the L / UL disk drive are often referred to as head-parking locations.

The data head 33 reads / writes user data on the disk 34 and is generally a Thin Film Inductive (TFI) read / write head or a magnetoresistive (MR) Quot; read head "). The data head 33 is connected to the preamplifier / write driver 3 and the data write channel 4 and to the disk controller electronic circuit 7. The data recording channel 4 can be any form such as a peak detector or a partial-response maximum likelihood (PRML), and includes functions such as data detection, encoding and decoding. The disk controller 7 connected to the microprocessor 9 deals with the process of reading and writing data and controls the communication with the computer 41 and manages the buffer memory 10 and controls the servo control electronics 6, And issues a command to the control electronic circuit section 5.

The disk drive 40 also includes a microprocessor 9, an associated memory 8, a buffer memory 10, an interface module 11 and a power control module 12. The microprocessor memory 8 is used to store code and data for the microprocessor 9. [ The buffer memory is used to store data transferred from the computer 41 to the disk drive 40 and is generally a cache memory. The interface module 11 controls the transmission of information from the computer interface controller 13 onto the interface. The integrated drive electronics (IDE) interface and the small computer system interface (SCSI) are the most common types of interfaces.

All elements of the disk drive 40 shown in FIG. 1 require operating power. The power control module 12 controls the power management of the disk drive 40 and receives power from the power source 20 through the bus 71. [ In Figure 1, the power connections of all the components are not shown, and each of these elements receives power directly through the bus 71 or through other components. The power control module 12 may be physically contained as a logic circuit in the controller 7 or stored in the memory 8 in the form of microcode executed by the microprocessor 9. [ The power line from the power source 20 to each energy consuming element may be directly controlled by the output of the module 12 or the microprocessor 9 or may be controlled by turning the elements on, turn-on commands and commands to change the power state of each element.

Although the connection between various elements of the disk drive 40 is schematically shown in FIG. 1, the functions can also be obtained by arranging the elements differently. For example, the microprocessor memory 8 may be included in the microprocessor 9 or may be coupled to the buffer memory 10. Furthermore, it is common to have more than one microprocessor in a disk drive, one of which is used primarily for interface and drive management functions and the other for servo functions. In this case, the servo control electronic circuit section 6 may include a separate servo microprocessor.

The computer 41 includes an interface controller 13, a processor 14, a memory 15, a display 16, a keyboard 17 or other input device, a peripheral device 18, a power source 20 and a control electronics 19). The interface controller 13 controls communication with the disk drive 40. The power source 20 is a power source for the computer 41 and the disk drive 40. The power source used in the portable device may be any other type of power source, such as an ac power supply, but is typically a rechargeable battery. The power supply 20 supplies power to the various elements in the computer 41 through the bus 70 and supplies power to the disk drive 40 through the other bus 71. The power source 20 also communicates with the control electronics 19. For example, a battery used in a portable computer includes an internal control electronic circuit for monitoring a battery condition, such as a discharge state.

Usually, the disk drive 40 does not consume energy at a constant rate. The current is supplied to the VCM 30 rather than the track following such that the head 33 is kept on a single data track by an intermittent current pulse sent to the VCM 30, More energy is consumed when seeking such as moving across the data tracks of the disk 34. [ Further, during the write operation, additional energy is consumed due to the flow of write current to the coil of the TFI head. The active power mode by such disk drive accesses is divided into the following two cases. A seek / read / write power mode and an idle power mode. The seek / read / write power mode is generally estimated because it is based on assumptions of user workload. The idle power mode is a normal track-tracking operation in which no data is written / read and no seek will occur. The term active state is used to indicate the state of the disk drive when in either the seek / read / write or idle power modes.

Two general power saving modes are referred to as " idle 2 "and" standby ". In the idle 2 mode, the actuator 31 is parked (i.e., moved to the crash stop 30 or unloaded onto the RAM 37), and the servo control electronics 6, the preamplifier 3, The containing write / read electronics is turned off. Thus, the idle 2 mode substantially reduces or eliminates the power supplied to the VCM driver 2, the servo control electronics section 6, the preamplifier 3, and the channel 4. In addition, in the idle 2 power save mode, since the servo operation and the write / read operation are not active, it is also possible to reduce the power supplied to the disk controller 7 and the microprocessor 9.

On the other hand, in the standby mode, the actuator 31 is moved to the parking position and the spindle motor 32 is turned off. The standby power saving mode has all of the power saving effect of the idle 2 mode while reducing the power supplied to the spindle control electronic circuit portion 5 and the spindle driver 1. [ In some embodiments, the buffer 10 may be turned off in either or both of the idle 2 mode and the standby mode. Additional power saving modes are also possible. For example, the sleep mode includes power saving features in the standby mode, and also allows power to be supplied only to parts of the interface controller 11 and to those needed to respond to the sleep recovery command from the computer interface controller 13 And nearly all of the remaining electronic circuitry is turned off.

Table 1 shows power values for a typical 2.5 inch disk drive in the two active modes and the two power saving modes. It is clear from this table that the power saving mode, ie idle 2 and standby mode, significantly reduces energy consumption.

[Table 1]

Table 1 also shows the recovery time (T _Rec ) for two power save modes, which is the time it takes the disk drive to go from sleep mode to active. The average recovery power (P _Rec ) is also shown, where the energy break-even time (T _BE ) can be calculated from this information. This is the time that the drive needs to stay in a particular power-saving mode so that the extra energy consumed during recovery from a particular power-saving mode is balanced with the reduced energy consumption while in this mode. Assuming that the normal active mode is idle and the recovery time (T _Rec ) is applied only in the power save mode, T _BE can be estimated by the following equation.

T _BE = P _Rec (T _Rec / P _Idle )

Table 2 shows the energy break-even time (T _BE ) for the example of Table 1.

[Table 2]

In the present invention, the actual user workload is used to determine which power mode is most appropriate and when to enter that mode. It is therefore desirable that the entry into the power save mode is in accordance with the demands of the user's workload rather than being predetermined by the fixed time of the user-selection.

On the other hand, in order to optimize energy consumption, it is very important to know the type of access to disk drives and what type of operation each element has. In the power save mode, access is defined as a task requiring a return from the power save mode. Assuming, for example, that the power save mode is idle 2, the disk 34 is rotating and the servo control electronics 6 and the VCM driver 2 are turned off, and the buffer 10 containing the disk cache In the active state, the disk drive access to move the actuator 31 to read / write data is counted as one access. The cache hit is not an access to this power save mode but it is not a definite other interface command such as status queries from the computer 41 because it does not require reading or writing to the disk 34. [ When designing simple, it is not necessary to consider the effect of cache hits when counting accesses. In this case, the read or write command received through the interface is counted as one access even if the data remains entirely in the cache. This assumption may reduce energy savings, but it can be cost effective because it reduces the complexity of designing power management technologies.

When making a decision to enter a particular power saving mode, for example standby (reducing the rotational speed of the disk 34), the recent access pattern must be considered. The access pattern contains information about the software process that is driving the access. Access patterns can be characterized by frequency, i.e. the rate at which disk drive access occurs, and the distribution of frequency can be determined from the access history. It is possible to determine when the observed access frequency does not belong to the frequency distribution of the recent access history. This determination is made statistically by estimating the probability that the low access frequency is not part of the recent access pattern, thus indicating that the processing of the access pattern and the software associated therewith is discontinued. It is related. In the present invention, these coefficients are also dynamically adjusted based on the performance of the disk drive to provide adaptive power management.

It is also possible to detect the periodic accesses from the access frequency and to enter and exit the power saving mode by predicting the start and end of the periodic accesses. As an example of periodic access, there is a case of automatic storage in which a document storage interval is specified in a word processing program. That is, the software user can use the automatic storage function of the word processing software to automatically record the file to the disc every 5 minutes. If you can exit sleep mode immediately before entering periodic access, you will improve performance because you do not know the response delay. Entering the power saving mode immediately after the termination of the periodic access can shorten the delay time and enter the power saving mode, so that the energy saving is increased.

In the preferred embodiment, the disk access pattern is assumed to belong to two categories: uniform access and sporadic access. For example, the average and standard deviation of the access frequency can be calculated, and if the standard deviation is a fraction of the average, the access frequency can be considered well defined. Otherwise, the access pattern is considered sporadic, ie not well defined by the mean and standard deviation.

In the case of uniform access, if the observed access frequency is less than or equal to a value obtained by subtracting several times the standard deviation from the average, the access pattern is regarded as completed. This is equivalent to selecting the probability that the observed access frequency belongs to the observed access pattern. Alternatively, it is also possible to estimate the termination of the access pattern using a fraction of the minimum observed access frequency. The basic principle is to use the latest access frequency to characterize the access pattern and determine the critical frequency from it. Thus, when the access frequency exceeds this critical frequency, the likelihood that the access pattern will stop is assumed to be very high.

The access frequency is measured by selecting a time window, counting the number of accesses occurring within the window, and converting this number to frequency. Different time windows may be selected for each power saving mode. The number of accesses occurring within a time window is called the access density.

FIG. 2 is a flow chart detailing the process of measuring the access density. The timer is first checked at step 403 to determine if the window time has expired. If the window has not yet been terminated, it is checked in step 401 whether an access has occurred, and if so, the density counter is increased in step 402. FIG. 2 shows the behavior when all the time windows are a plurality of the shortest windows. In this case, the densities for all power saving modes are increased at step 402. The reset of the density counter is not indicated, but the density value for the current window is initialized after being transferred to another storage device, such as a ring buffer, before access to the next window begins. The density count is initialized only at the end of the longer window in the power saving mode with time windows in the minimum time window of a predetermined multiple. FIG. 2 illustrates the operation of a polling loop design. An equivalent interrupt-based design may be easily derived from the second road.

The process shown in FIG. 2 is part of the functions of the power module 12 of FIG. 1, and may be implemented on hardware and / or on software. FIG. 3 shows a suitable hardware configuration for implementing the process of FIG. 2. Timer 201 and counter 202 and ring buffer 203 are shown as part of controller 7. The ring buffer 203 is addressed by the microprocessor 9. The same is true for steps 4 to 6, but a set of program instructions for carrying out the steps of the process shown in FIG. 2 are stored in microcode in the memory 8 addressed by the microprocessor 9 . The access signal 220 is input to the density counter 202 in the disk controller 7. The signal 202 may be sent to the controller 7 at the same time when it responds to a read request from the interface 11 and the interface controller 13 or when the control 7 transfers the read command to the channel 4. [ Lt; / RTI > The access signal 220 is sent to the counter 202 every time the disk drive is accessed and the counter 202 counts the accesses 220. Timer 201 continues to operate to output signal 221 at the end of each time window. The density value at the counter is output as signal 222 and passed to a density storage buffer 203, such as a ring buffer. The end-of-window signal 221 from the timer 201 causes the ring buffer 203 to write the density value 222 to the next memory location and store the write pointer value . The signal 221 then initializes the counter 202. The density value is read out from the ring buffer 203 by the microprocessor 9. The microprocessor 9 transmits clear 223 and read 225 signals to read the buffer 203 and receives output signals 224 that represent the density value. The microprocessor 9 uses the density values to issue commands to change their power state in each component of the disk drive while executing the microcode stored in the memory 8. [ Here, changing the power state means entering or exiting an appropriate power saving mode.

In general, the time window for the power saving mode is selected to enable a range of access frequencies that the disk drive can appropriately respond to the power saving mode in order to obtain a good response. Here, the frequencies of interest are those in which the period has a value similar to the latency and the energy break-even time of the power saving mode. Accordingly, a time window having a value close to the energy break-even time becomes appropriate, The window value can be further optimized by examining the operation of the disk drive or by simulation. Other coefficients, such as performance goals, should also be considered when selecting the time window. In the example of the disk drive of Table 1, it is preferable to select a time window of 400 ms for idle 2 and 1.6 s for standby.

The access density value obtained in the above description is the number of accesses that occur in a particular time window. The density value is then converted to a frequency value by a scale, a scaling method. The dynamic range in frequency is expanded by implementing it so that the density at "0" is not equal to the frequency at "0 ". When the density is "0 ", the frequency is calculated from the number of consecutive" 0 "densities. One of these conversion equations is as follows.

Frequency = density * scale; Density> 0 (2)

Frequency = scale / ("0" density number +1); Density <= 0 (3)

Where the "zero-density" is the number of consecutive time windows with a density of zero. Scale means a scale factor for converting density to frequency. From the above formulas (2) and (3), a good approximate access frequency can be obtained when the access frequency has a value less than 1 / time-window. For ease of processing, it is common to use a frequency unit having an integer value defined by a scaling factor. The density having a value of "1 " corresponds to the frequency of the scale. The magnification factor is selected for ease of calculation and provides the desired dynamic range of the frequency. Normally, a value of 256 is used as a scale, which is suitable for 16-bit processing and also simplifies scaling when the density is not a value of "0 " .

Table 3 below is an example of the density-frequency conversion using the above formulas (2) and (3) and the 400 ms time window, i.e., idle 2 mode. The first line represents the start time for each time window, the second line represents an example of the access density value for each window, and the third line represents the access frequency derived from the access density in the second line. For example, accesses 19 occur in a time window between 0.4 s and 0.8 s, and the scaling factor has a value of 256, which corresponds to frequency 4864. Also, since there is no access in the time window between 1.2s and 2.0s, two "0" density values are obtained. These three density values are converted to a single frequency by the formula 256/3 = 85.

[Table 3]

It is desirable to minimize the energy management calculations if you want to reduce any delays during the disk drive's access period. This approach to density has the advantage that it does not require much computation when an access is made to the disk drive. It is rather desirable that this density measurement or density-frequency conversion is omitted when the disk drive is accessed too frequently. This increases the response time, and the access density for the interval in this case can be set to a fixed value applied in the case of bulk access. In this case, the access density or frequency value for the interval is updated when the operation of the disk drive is slightly reduced. For example, a density of 256 corresponds to a frequency value of 65,536 with a magnification factor of 256, which is a busy interval.

The value of the access frequency obtained in the density-to-frequency transform described above is an estimate since the exact timing of the accesses from this density is not obtained. However, if the window time is appropriately selected, the estimate is sufficiently accurate within the frequency range of interest. Measuring or estimating the access frequency may be accomplished by other techniques such as Fourier transform techniques instead of the above method.

The microprocessor 9 decides to enter the power save mode when the current access frequency becomes less than the threshold frequency determined in the access history. FIG. 4 is a flow chart showing details of entering the power saving mode. In this flowchart, it is assumed that the various power saving modes are assigned values in order from the value 1 to the max mode (maxmode) in the order of increasing the power saving effect. The term tf [mode] represents the critical frequency of a particular power saving mode. The term lf [mode] represents a low-frequency flag for a particular mode. This flag helps control the calculation of the critical frequency. In step 301, it means that all power saving modes have been checked, and the previous mode (last mode) is set to 1. For all modes, the critical frequency is cleared (tf [mode] = 0) and the low-frequency flag is also cleared (lf [mode] = 0).

The process then waits at step 302 until the minimum window time has elapsed. This process has already been shown and described in FIG. Assume that the longer window time is an integer multiple of the minimum window time when more than one window time, that is, a unique time window is used for each power saving mode, comes back to the fourth road again. If window time has elapsed once and access density has been measured, step 303 is performed. In step 303, the power saving mode to be inspected is set to the maximum energy saving mode. In step 304, the process checks whether the current time is window time for the mode to be checked. If not, perform step 308.

If in step 308 it is the current all-pass mode, then return to step 302 and the density is measured for the next time window. At this point, it is necessary to determine which of the active power saving modes is to be checked. If the current mode is not the previous mode in step 308, the next lower mode (mode-1) is selected in the order in which the power saving effect is reduced in step 309. Then, step 304 is performed again. Note that in the check of step 304, the state of the previous mode = 1 is always "true" because it is the same as the state of step 302 is terminated. If the check in step 304 is "TRUE ", then a new density value is made available for the mode and step 305 is reached. In step 305, the access density measured in step 302 is converted to abacus freq using conversion equations (2) and (3). In step 306, the access frequency freq calculated in step 305 is compared with a threshold frequency, tf [mode], for the current mode. If the access frequency is larger than the threshold frequency, the procedure goes to step 307. In step 307, a finite frequency tf [mode] for the current mode is obtained. The threshold frequency may or may not vary depending on the condition. The details of adjusting the critical frequency will be described later with reference to FIG. 5. Then, at step 308, the process continues for the next mode as described above.

On the other hand, if the access frequency is less than or equal to the critical frequency in step 306, the process proceeds to step 310, and the current power saving mode is input to the mode. This means that the microprocessor 9 sends a signal to the appropriate element of the disk drive to reduce power. If the current mode is the Max mode in step 311, there is no power saving mode to be performed, so the process proceeds to step 313, and the flow of this flowchart is stopped. However, if there is still a power save mode to be checked, the process goes to step 312 to update the previous mode to [mode + 1]. This is to reflect that all power saving modes that are less than or equal to the input mode are no longer considered. Thereafter, the process returns to step 302 to perform the subsequent step.

In order to save as much energy as possible, the order of checking the power saving mode in step 304 starts from the mode in which the most energy is saved first, and then proceeds in the order of the next energy saving mode. Thus, in step 310, it is possible to input the most favorable power saving mode to the disk drive. For example, by applying the present invention, the disk drive in the idle mode can enter the standby mode without first entering the idle 2 mode. If a particular power saving mode other than the maximum mode is entered, the disk drive will operate in that particular mode until it is suitable to enter a better mode or until it exits that particular mode.

5 is a flow chart showing the details of a preferred embodiment for calculating and adjusting the critical frequency for the power save mode. The most recent access frequency is stored in a ring buffer (not shown) addressed by the microprocessor 9. The buffers may be different for each power saving mode. A ring buffer is a set of registers, or memory locations, in which frequency measurements are accumulated. The ring buffer holds several frequencies corresponding to several registers or memory locations for a set. When several frequencies are loaded, the addition of additional frequency values causes the oldest value to be lost, which provides the effect of giving a sliding view over recent access history.

In step 501, a check is made to determine whether there are sufficient values in the ring buffer to calculate the threshold frequency for the current power save mode. The minimum number of frequency values is typically two. The higher the number, the higher the statistical accuracy, but the values must be chosen before calculating the critical frequency. If the number of values in the ring buffer is not sufficient, go to step 503. In step 503, it is checked whether the current frequency freq is larger than the scale / 2. Here, the scale is defined in the above-mentioned formulas (2) and (3). If the result of the check is true, the process proceeds to step 504 where the current frequency value is input to the ring buffer. If 'false', the number of consecutive time windows with a density of "0" continues to increase, so the previous frequency value in the ring buffer must be changed. This is done in step 509. Note that this is not an input to the ring buffer, so the number of values in the ring buffer does not change. Step 510 is entered from steps 504 and 509 and returns the process to step 501. [

If it is determined at step 501 that the number of values in the ring buffer is sufficient, step 502 is reached. In step 502, the current access frequency freq is compared to the active frequency threshold freq-act. This value increases the speed of the process and provides a safety threshold. Scale / 2 is the first selectable value for freq-act. If the check at step 502 is 'true', then the disk drive access frequency is greater than the active threshold, so no further computation of the critical frequency is required. Thereafter, the process proceeds to step 505, and if [mode] is cleared, the process proceeds to step 503, and then the process proceeds as described above.

On the other hand, if the check at step 502 is 'false', the process needs to proceed further since the access frequency is less than the active threshold. In step 506, the low-frequency flag lf [mode] is checked. If the flag is set, the process returns to step 503 and the current low-frequency threshold is maintained. Suppose that the end of the access pattern is checked when the low-frequency flag is set. As a result, the current frequency is assumed not to be part of this pattern and can not be used to change the critical frequency. However, since the low-frequency flag is cleared as the active threshold freq-act is increased, this frequency is still in the ring buffer. The assumption that the access pattern is stopped in this case is incorrect. So the frequencies that occur during the inspection are now considered part of the access pattern.

On the other hand, if the low-frequency flag is cleared in step 506, the process proceeds to step 507 because a new frequency has to be calculated. This calculation is performed in step 507, and then in step 508 the low-frequency mode flag is set and then passed to step 503 again.

The process shown in FIG. 5 uses the most recent access history to calculate the critical frequency for each power save mode. The calculation is performed only when the access frequency is less than the active threshold. This is advantageous in that it only performs the calculation of the access frequency when the disk drive is inactive so that there is little influence on the performance.

In step 507, the threshold frequency is calculated based on the access pattern in the associated ring buffer for the previous power save mode. The mean value and the standard deviation of the values in the ring buffer are calculated. These calculations include complex number calculations. Instead, the mean value frequency meanf in the uniform access pattern is calculated from the maximum and minimum frequencies in the ring buffer, maxf and minf by the following equation.

Average frequency = (maximum frequency + minimum frequency) / 2 (4)

This gives a good estimate for the uniform access pattern since the distribution is assumed to be well-specified. The standard deviation sdevf is calculated from the frequency range as follows.

Standard deviation = (maximum frequency - minimum frequency) / 4 (5)

However, since it is not necessary to calculate the mean and standard deviation accurately, they are more preferably calculated using the simplified formulas (4) and (5).

In sporadic access patterns, only a fraction of the minimum frequency is sufficient. Formulas (6) to (8) shown below are the formulas used in step 507.

t1 [mode] = (maximum frequency + minimum frequency) / 2-g1 [mode] * (maximum frequency * minimum frequency)

t2 [mode] = minimum frequency / g2 [mode] (7)

tf [mode] = max (t1, t2) (8)

The t1 [mode] is a value corresponding to the uniform access pattern, while the t2 [mode] is a value corresponding to the sporadic access pattern. The larger of these two values is used as the critical frequency for the mode

There are two gain factors, g1 [mode] and g2 [mode], which are for uniform access patterns and sporadic access patterns. These gain factors have different values depending on each mode. For the disk drives of Table 1, the values of g1 = 1 and g2 = 4 are suitable for both idle 2 and standby modes. As can be seen from formulas (6) to (8), increasing the value of the gain factor has the effect of reducing the critical frequency, and decreasing the value of the gain factor has the effect of increasing the critical frequency. Therefore, the entering action of the power save mode is appropriately harmonized by selecting and adjusting these gain factors. These gain factors can be selected and adjusted by the user through appropriate system or application software.

Since separate ring buffers are used for each power save mode, the size of the buffers must be selected independently. Increasing the size of the buffer increases the time required for the access history to be written, while decreasing the size reduces the time. When choosing the buffer size, the length of the history must be taken into account in order to achieve an improved response to recent events and to reduce the amount of memory required. For the disk drives of Table 1, the buffer size was appropriately chosen to be 16 for the two power saving modes. A more sophisticated way of maintaining access history is possible. For example, frequencies in the ring buffer have weighting factors related to the length of time in the buffer. The bar graph of the access frequency can be used in place of the ring buffer as a mechanism for deleting previous data, such as renormalizing. However, the ring buffer has the advantage of simplicity of design.

The ring buffer can be constructed using density instead of frequency as shown in Figure 3. In this case, the density is converted to frequency using the threshold frequency calculation method. 6 shows a process for calculating a threshold frequency using a density buffer. This flow chart corresponds very closely to the flow chart of FIG. 5, with some exceptions. This flow chart corresponds very closely to the flow chart of FIG. 5, with some exceptions. That is, step 601 corresponds to step 501. And step 603 of converting density density to frequency freq using equations (2) and (3) has been added. Steps 604, 605 and 606 correspond to steps 502, 505 and 506, respectively. Step 607 has been added to convert each density value in the ring buffer to a frequency value using the following formulas (2) and (3). Note that this conversion step results in a smaller number of frequency values than the number of density values in the ring buffer because the continuous "0" density is converted to a single frequency value. Steps 608 and 609 correspond to steps 507 and 508, respectively. Finally, step 602 inputs the current density value to the ring buffer for the selected power saving mode.

Both of the configurations for the ring buffer described above are suitable for calculating the critical frequency. The frequency buffer has the advantage that the history length increases with the lapse of time when there are low frequencies. On the other hand, density buffers have the advantage of having a fixed history length. The density buffer also reduces the computational overhead during the interval of the activity disk access because the density-to-frequency conversion is delayed until the active frequency threshold is reached.

As described above, once the disk drive enters the power save mode, it enters either another power save mode or the disk drive returns to the active state. The former occurs when the estimated access frequency continues to decrease and reaches a critical frequency in another mode. The latter occurs when a periodic access pattern is detected that results in disk drive access or drive-initiated entering. When the disk drive returns to the active state, the ring buffers are desperate to clear because a new access pattern is assumed to be measured. This further reduces the influence of the old pattern in the ring buffer. If it is possible in other configurations to change the weighting factor for the data from the previous pattern, it may be desirable to keep ring buffer data persistent.

Other coefficients may be used to adjust the gain factor, which in turn affects the determination of the critical frequency. For example, the attempt to enter a power saving mode in a standby state where the head can not immediately read / write data can be weighted by the current cache hitrate. The high cache hit rate is used, for example, to adjust the threshold frequency to have a high value by reducing the gain factor. Even though the accesses occur for a short period of time after reaching the critical frequency, these accesses hit the cache. As a result, the possibility of using extra energy is still reduced. Similarly, a lower cache hit rate results in a lower threshold frequency because the probability of successive accesses requiring read or write from the disk is higher. In this case, a further power saving mode for controlling the power required for the same cache as the buffer 10 of FIG. 1 is desirable. By doing so, the cache buffer remains active in standby mode. The gain factors for this mode are, of course, affected by the cache hit ratio. In the case of a high cache hit rate, it is desirable to reduce the threshold frequency while allowing more cache hit for performance. It is also possible to treat read and write operations separately or to characterize cache accesses using locality where access occurs. All of these are used to adjust the gain factor and eventually the critical frequency.

As described above, there are several parameters that can be adjusted in the above-described design. These include gain factors g1 and g2, time window size, ring buffer size, density-frequency conversion scale factor scale and active threshold frequency. In the preferred embodiment, however, the most suitable parameters for adjusting the entering action of the power saving mode are the gain coefficients in formulas (6) and (7). The adjustable parameters have a fixed value at the time of manufacture of the disk drive, but they are adjustable by the user through adjustable or dynamic power management, taking into account performance penalties.

As described above, through the interface controller 13 in FIG. 1, the computer 41 can issue a command to directly set the gain coefficients. However, it is rather advantageous that the computer 41 is not directly involved in the internal detailing process for energy management. So we use a separate performance factor pf. This is adjusted by receiving commands from the computer 41 and is not affected by the characteristic energy management design applied to the disk drive.

In general, energy management with a power saving mode in a disk drive requires a proper tradeoff between access performance and energy savings. This is due to the direct effect of the recovery latencies of the power saving mode described above. Thus, a single weighted coefficient of performance representing the importance of compromise between energy and performance is highly desirable. The single weighted coefficient of performance at one extremity maximizes energy savings without considering performance, and the coefficient at the other extremum maximizes performance without considering energy savings. This method is completely different from conventional disk drive power management techniques that adjust the fixed time to enter the power save mode by the energy management command. These commands are not directly related to performance or energy savings, but also cause different energy savings and performance effects, even for other drives with the same fixed time setting. However, the scale of the energy versus performance factor has a fixed value so that all drives behave similarly to one another.

In the preferred embodiment, the coefficient of performance command is defined as Table 4 below.

[Table 4]

Where a value of " 0 "corresponds to maximum energy savings, and a value of" 255 " This user-selected performance factor command may also be emulated using standard fixed-time commands available on common disk drive interfaces such as SCSI or IDE. In the case of IDE, the standby mode time value value ranges from 0 to 255 without power saving mode. Actually time = 5s * value, so it has a range from 5 seconds to about 20 minutes. In this case, in converting the command into the performance coefficient, "0" indicates the maximum performance, so that the linear scale from Table 4 has a range of 1 to 255 instead of 0 to 254.

In the preferred embodiment, the input from the performance factor command is changed to the operation of the energy management system by adjusting the gain factors g1 and g2. As a result, the value of the gain factor is transformed into a range between the two limits expressing extreme energy and performance. For example, in the case of the uniform access of the formula (6), g1 corresponds to a multiplier when estimating the standard deviation sdev. This is 0.5 <g1 <5 as a practical range for g1. Similarly, since g2 is a fraction of the minimum frequency minf, 1 <g2 <10 is presented as a practical range as well. Statistically, the lower bound of g1 corresponds to a probability of about 20% for regularly distributed data where the critical frequency is actually part of the distribution. If the lower limit is too low, the probability of inadequate entering the power saving mode increases, resulting in an increase in energy use. The upper limit of g1 corresponds to a probability of about 10 ^{-6, which is} slightly larger than a sufficient number. Even though it is not statistically well defined, g2 also resembles the change of the above limit value. It is desirable to consider empirical data on performance and energy behavior to make the selection of the limit values more effective. Furthermore, since many setting values in practice have distinct values that are clearly distinguishable from each other, it is not necessary to obtain a precise resolution as shown in Table 4.

The following equations (9) and (10) show the conversion coefficient of the performance coefficient pf for the command in Table 4 in the disk drive having the power value of Table 1. [

g1 = (12 + pf / 4) / 16 (9)

g2 = (24 + pf / 2) / 16 (10)

The extreme values used here are roughly 0.75 <g1 <4.75 and 1.5 <g2 <9.5. Equations (9) and (10) are designed to have integer values for ease of implementation. As a final step for calculating Formulas (6) and (7), it is preferable to execute the scale factor of 16.

Other parameters, such as time window size, can also be adjusted using the performance factor. Generally, the higher the value, the better the performance of the power consumption. Furthermore, a subset of possible power saving modes is also selected. For example, there are a few power-saving modes that affect performance in comparison to others, but it is better not to use them if performance is considered important. The performance factor pf described above has the advantage of allowing the disk drive designer to determine the parameters that need to be adjusted to achieve the performance goal. At this time, the end user or the system assembler need not know the details of the specific implementation.

In general, entry into sleep modes is based on disk drive access, satisfactory power management actions can be obtained using equations (6) and (7), where parameters are fixed until new values are selected Value. Selecting a set of parameters such as gain factors determines the suitability of the selection according to the actual access pattern. If there are changes in the access pattern that affect the operation of the power-saving mode entry, new parameters must be selected. However, the parameters are dynamically adjusted as the access pattern changes. This is accomplished by using an adaptive system that dynamically adjusts gain factors to determine how to approach performance goals.

In order to achieve dynamic adaptability, it is necessary to have an ability to judge an actual action in comparison with a required action. The best performance is obtained when the deflection magnitude and direction from the performance target is measured. These measurements are treated as penalties and are then used to adjust the gain factors. For power management, it is convenient to define two penalties: Energy / Response Penalties (ERP) and Missed Opportunity Penalties (MOP). The former occurs when the disk drive uses excess energy or affects performance and implies that the critical frequency tf is chosen to be too high. The latter implies that when the disk drive is judged to be appropriate, it occurs at the time of entering the power save mode, and the critical bend tf is chosen to be too low. Since the two types of penalties have opposite effects in determining the critical frequency, they are used to balance the system action.

The penalties are used to adjust the parameters in the formula of the critical frequency, so they can be calculated in any unit as convenient. Calculating the penalty in units of time has the advantage of simple calculation. On the other hand, calculating the penalty in frequency units has the advantage of having several input values available from the demand-driven calculation.

7 is a graph showing the time sequence for entering and exiting the power saving mode. Where the horizontal axis represents time and the vertical axis represents power. Each short tick mark on the horizontal axis represents the time window for the power save mode. The graph uses three power levels: seek / read / write power P0 and idle power P1 and mode power p2. For ease of identification, all accesses to the disk drive are indicated at short intervals at power level P0. Time T0 is the beginning of the time window in which access occurs. At a time T1, which is a little time, the mode enters the intermittent mode and continues until time T2. Then, when the next disk access occurs, the drive returns to the active state. The next disk access is generated in the time window starting at time T4.

The durations of power saving modes in performance penalties are very important. This is illustrated in FIG. 8 by adding a counter 204 in the hardware configuration of FIG. 3 shown and described above. The counter 204 counts the number of time windows 221 between the mode entry 226 and the mode exit shown as access 220. [ Output 227 represents the duration of the mode in time window units. This value can be used directly for a time unit, or it can be converted to a frequency unit through formulas (2) and (3).

Both the energy impact and the response time impact on the energy / response penalty erp are evaluated. The energy penalty ep is a measure of how much extra energy is used and when it improperly enters power saving mode. The response time penalty rp is a measure of how much the actual data throughput is affected when the power-on mode is improperly entered. In the example of FIG. 7, two mode entries are shown, one from T1 to T2 and the other from T5 to T6. The energy / response penalty occurs in the mode entered at T5, not for the mode entered at T1. This is indicated at the ERP level in FIG.

There is an energy break-even time T _BE for the power saving mode that allows the energy saved during the duration of the mode in the energy penalty to be balanced by the return energy. For reference, referring to the first mode entry of FIG. 7, T0 represents the previous time window with access density &gt; 0, T1 represents the time to enter the mode, and T2 represents the time at which the mode ends. The value of T2-T1 is the mode duration counted in counter 9 of the implemented hardware. If (T2-T1) &lt; T _BE, then the energy penalty eq is calculated as a function of (T2-T1) and T _BE . That is, as a preferred embodiment,

Is _{(T2-T1) <T BE} ) that is "true", eq = 16-16 * (( T2-T1) is / T _BE), and the "false" = 0 eq. (11)

In equation (11), the energy penalty ep is calculated to range linearly from a value of 16. At this time, for the value of "0" when the response time exceeds the limit, the duration of the power saving mode is essentially "0". This is because access immediately follows the mode entry. Another energy penalty formula that provides different weights to penalties is also possible. However, equation (11) is very simple and gives a reasonable estimate of the energy impact. In fact, the values of the parameters in equation (11) do not need to be accurate, and the equations are performed as integer calculations. There are 16 penalty levels for integer calculations that provide enough resolution.

The response time impact of the power save mode is calculated based on the additional wait time resulting from the return. This is an estimate of the effect on the throughput, which is a measure of the performance of the disk drive. As a result, the upper limit of the throughput effect, tub, can be derived from the gain factor. The throughput effect from the power save mode is measured as r1 / (T2-T0). As a result, the response time limit trl is determined according to the upper limit tube. That is, trl = rl / tub. If (T2-T0) &lt; trl, a response penalty occurs at any time. (T2-T0) is calculated from the threshold frequency and the mode duration (T2-T1).

tf = scale / (T1-T0 + 1) (12)

(T2-T0) = T2-T1 + scale / tf (13)

The response penalty rp is calculated as a function of (T2-T0) and trl.

In a preferred embodiment,

Rp = 16-16 * ((T2-T0) / trl) if (T2-T0) <trl) is true and rp = 0 if it is false. (14)

In formula (14), the response time penalty rp is calculated to have a linear range from a value of 16. At this time, for the value of "0" when the response time limit exceeds the limit, the duration of the power saving mode is essentially "0". Another energy penalty formula that provides different weights to penalties is also possible. However, equation (14) is very simple and gives a good estimate of the energy impact. In practice, the values of the parameters in equation (14) need not be accurate, and the equations are performed with integer calculations. There are 16 penalty levels for integer calculations that provide enough resolution.

The energy penalty ep and the response time penalty rp are combined to produce the energy / response penalty erp. In the preferred embodiment, the larger of the two penalties is used as the energy / response penalty erp. This will simplify the calculation because only penalties with larger time ranges are calculated.

mt1 = max (T _BE , trl) (15)

However, it is desirable to calculate energy / response penalties using both energy and response penalties, for example, by using a weighted average.

While energy and response penalties can be calculated as frequency units. In this case, the frequency is calculated from the time values T _BE and trl according to the formulas (2) and (3). Here, the energy break-even frequency fbe and the response limiting frequency frl are obtained. The duration of the mode is converted to a frequency value using equations (2) and (3). The penalty formula is as follows.

ep = 16-16 * feb / fmd if (fmd> feb) is "true" and ep = 0 if it is "false". (16)

And,

Rf = 16-16 * fr1 * tf * fmd / (feb (1 / fm-1 / tf &gt; 1 / fr1)

fmd), and if it is "false", rp = 0. (17)

In a preferred embodiment, if only the larger of the two penalties is used again, only penalties with lower frequency limits are calculated.

mf1 = max (fbe, fr1) (18)

The energy and response penalty is calculated when the mode is terminated to return to the active state. Assuming a deeper mode of entry, the penalty does not occur since the return penalty is associated with the selected mode.

On the other hand, we will look at another type of penalty, the opportunity loss penalty mop. Using energy-saving mode in this penalty is ideal compared to the action in which no energy or response penalty occurs at all. An opportunity loss penalty occurs when the power save mode is not used (Type 1), or when the threshold frequency entering mode is too low (Type 2). On occasion, the two types are considered only when it is possible to use an interval without incurring energy or response penalties. In FIG. 7, the time interval M3 of T4 to T4 is an example of a type 1 opportunity, while the time interval of T5 in time intervals MOP2 and T4 in T0 is an example of a type 2 opportunity. Like the energy / response penalty, the opportunity loss penalty can be calculated as the frequency or time that you think is most convenient.

First, the opportunity loss penalty is calculated in units of time. For the power save mode, the loss penalty occurs when the time interval is lost. For Type 1 opportunities, the relevant time is T3, which is a previous time window with an access density greater than "0", and T4, which is the next disk access time when not entering sleep mode within the time interval. The value of (T4-T3) represents the length of the time interval and is calculated from the access frequency using the following formula:

T4-T3 = scale / freq-1 (19)

Here, frequency freq is the measured access frequency. Of course, you should only consider time intervals greater than the energy break-in time and response limit range.

The formula for the Type 1 opportunity loss penalty is:

mop1 = (T4 - T3) / mt1 (20)

When implemented as an integer calculation, if the opportunity (T4-T3) becomes a penalty, it becomes a penalty of "0" by the above formula. The size of the penalty is determined by how big the opportunity is, rather than how big the energy break-in time and response limit are. So the bigger the chance, the bigger the penalty and the smaller the chance, the smaller the penalty.

The value of (T1-T0) for the Type 2 opportunity entering the former 1 is calculated from the critical frequency using the following formula.

T1-T0 = scale / tf-1 (21)

Type 2 penalties are used only when power saving modes are available without energy or response penalties and when an inadequate amount of energy is exhausted before the threshold frequency enters the mode. This is obtained by the following formula (22).

mop2 = (T1-T0) / mt1 (22)

If the opportunity results in an energy or response penalty, as in Type 1, no penalty will occur. In the case of type 1, this is an absolute limit. In the Type 2 case, this is an approximation since the energy penalty or response penalty depends on both the Type 2 opportunity and the mode duration. However, formula (22) has the advantage that it is very easy to implement.

The opportunity loss penalty is also calculated in frequency units. A penalty is generated when the measurement frequency freq is less than the frequency penalty lower limit value mf1 in the type 1 opportunity. In this case, the following formula is a very good penalty formula.

mop1 = mt1 / freq (23)

The formula for Type 2 opportunities is:

mop2 = mt1 / freq (24)

An advantage of calculating penalties using frequency units is that the current frequency freq and the threshold frequency tf can be easily obtained. However, time units are preferred because they can be easily implemented as counters in hardware implementations, and time units are used to maintain consistency with energy and response penalties.

Of course other penalty formulas are also possible. It is assumed in the above description that if an energy or response penalty occurs, the opportunity loss penalty does not occur. In some cases, it is rather desirable that an opportunity loss penalty occurs in this case. The penalty then has an additional weighting factor depending on the magnitude of the energy or response penalty.

When in the multiple power save modes, the penalty values are preferably determined for each mode. Because the modes have different time windows, which are different parts of the access frequency spectrum. For one mode the entry into the other modes is used to determine the penalty. For example, the penalties may be checked for all modes at all time intervals, or only for the most optimal mode. Furthermore, it is convenient to maintain multiple penalties for each power save mode. For example, formulas (6) through (8) show cases of uniform and sporadic access patterns. The set of penalties is maintained with an independent dynamic adaptability for these two cases. This is realized by recognizing which type of pattern is activated during a predetermined time interval and calculating a penalty for the pattern.

The time histories of the penalties are used to adjust parameters such as gain factors g1 and g2 and the like in the calculation equations (6) and (7) of the critical frequency. At the same time as each penalty is generated in the preferred embodiment, they are added to the corresponding cumulative penalty value, cumulative energy / response penalties cerp and Cumulative Opportunity Loss penalties cmop. When one of these values reaches a predetermined trip level, the selected parameters of the threshold frequency calculation are changed accordingly, and the accumulated value is reduced. If the penalty does not occur and enters the power saving mode, the corresponding cumulative penalty value is reduced to a significant amount. This has the effect of decreasing the penalty over time, so that the previous penalties have a relatively small impact relative to the new penalties. This process is illustrated in the graph in FIG. The horizontal axis is the mode interval, and each tick corresponds to mode entry or opportunity. The vertical axis represents the magnitude of the cumulative penalty cerp. The cumulative value increases continuously after the penalty has occurred at the first interval 0. Since the penalty did not occur at interval 1, the cumulative penalty value decreases. Subsequently, the cumulative penalty value for the new penalty increases, and the cumulative penalty value decreases at a fixed rate when there is no penalty. The new penalty size with cumulative penalty at interval 7 is sufficient to exceed the penalty trip level as shown in the graph. At this point, the gain factors in equations (6) and (7) are changed and the cumulative penalty value is reduced to zero. The cumulative penalty is cleared when it is assumed that the action of adjusting the parameter has changed. This is the moment when a new measurement is required. These results are visible in time intervals after interval 8 with few penalties, and penalties are less than in the preceding time interval.

In a preferred embodiment, the energy / response penalty selected as the greater of ep and rp is added to the cumulative penalty cerp when this penalty occurs.

cerp = cerp + max (ep, rp) (25)

Thereafter, the cumulative penalty is compared to a predetermined trip level erpt, and the gain factors g1 and g2 in the threshold frequency calculation are changed. In the preferred embodiment, the penalties are individually detected for uniform and sporadic accesses, respectively, and the appropriate gain factors g1 and g2 are increased. A good value for erpt is 16. As described above, there are practical limit values for the gain factors, which are used to limit the upper range of the gain.

If there is no penalty, the cumulative penalty cerp is decreased by a predetermined amount cerd to the lower limit value of 0 at the end of each mode.

cerp = cerp-cerd (26)

Where cerd acts as the decay rate for the cumulative penalty. Normally, 2 is appropriate as the value of cerd. Larger values cause penalties at higher speeds, while smaller values cause penalties at later speeds.

The opportunity loss cumulative penalty cmop is calculated similar to the energy / response cumulative penalty. It is possible to treat Type 1 and Type 2 penalties as individual cumulative penalties, or apply a cumulative penalty of the same value to them for simplicity. When an opportunity loss penalty occurs, the appropriate value of mop1 or mop2 is added to the cumulative penalty cmop.

In the preferred embodiment, the appropriate gain factors g1 and g2 are reduced when the cumulative penalty exceeds the trip level mopt. A good value for mopt is 16. As described in detail above, there are practical limit values for the gain factors, which are used to limit the lower limit range for the gain. If there is no penalty, the cumulative penalty cmop decreases by a predetermined amount cmod from the end of each mode to the lower limit value of 0. As in the above embodiment, 2 is appropriate for the value of cmod.

The energy / response penalty erp and the type 2 opportunity loss penalty mop2 are calculated when the mode ends. A detailed process for this is shown in FIG. Step 350 enters step 313 of FIG. 4 when entering the deepest mode. Step 352 is entered from step 401 of FIG. 2 when an access occurs and the drive is currently in power save mode. When entering from step 350, it waits for access at step 351. In step 352, the durati on is obtained. In step 353, the energy penalty ep and the response penalty rp are calculated as described above. In step 354, a check is made to see if there is a penalty for the two penalties. If there is no penalty, go to step 356. And if a penalty is present, go to step 355. At step 355 the penalty is added to the cumulative penalty cerp. At step 358, the cumulative penalty is checked against the penalty trip level erpt. If the trip level has not been exceeded, go to step 367 to return the control to step 300 of FIG.

If the trip level has been exceeded, the gain factor is changed to an appropriate value in step 359. In step 360, the cumulative penalty is set to 0, and the process proceeds to step 367. [ Since there is no penalty in step 356, the cumulative penalty value is decreased. In step 357, the penalty for loss of type 2 opportunity is calculated using equation (24). In step 361, it is checked whether a penalty has occurred. If the penalty has not occurred, the cumulative penalty value is decreased in step 362, and the process proceeds to step 367. [ If a penalty has occurred, step 363 adds the penalty generated to the cumulative penalty. At step 364 it is checked whether the cumulative penalty has exceeded the trip level. If not, go to step 367. If the cumulative penalty has exceeded the trip level, the gain is changed to an appropriate value in step 365, and the cumulative penalty is set to zero in step 366, and the process proceeds to step 367.

The Type 1 opportunity loss penalty mop1 is calculated when the access frequency is less than the frequency penalty limit value. Referring to FIG. 4, a comparison and calculation of the values takes place as part of step 307. This process is shown in more detail in FIG. Step 325 is entered from step 307 of the fourth step. In step 326, the penalty value mop1 is calculated as the formula (23). In step 327, it is checked whether or not there is a penalty. If there is no penalty, the process returns from step 328 to step 307. If a penalty has occurred, step 329 adds a cumulative penalty cmop thereto. At step 330 the cumulative penalty is compared to the opportunity loss trip level mopt. If the trip level has not been exceeded, go to step 328. And if the trip level has been exceeded, go to step 311 and the gain factor is changed to the appropriate value. At step 332, the cumulative penalty is cleared and the process proceeds to step 328.

It is also possible to adjust the gain factors without explicitly using an opportunity loss penalty. In this case, the abacus number of energy / response penalties is used instead of the opportunity loss penalty. This applies in the case where it is assumed that optimal action occurs when a predetermined number of energy / response penalties are generated. Too low frequency of occurrence indicates that the system did not save enough energy. Thus, if the occurrence frequency rate drops below a certain level, the gain factors can be changed as if an opportunity loss penalty occurred.

It is also possible to set another method of feeding back the penalty information to the threshold frequency calculation. However, this implementation has the advantage of being very simple to design and provides immediate response to the operation of the actual system. Other methods include maintaining a penalty statistic such as by a histogram and thereby providing time-weighting to penalties such as renormalizing. Moreover, it is desirable to adjust other parameters such as penalty trip level, penalty decay rate, penalty magnitude conversion coefficients such as those used in equation (14), and limit values in gain elements. They can be dynamically adjusted either via an instruction from the computer 41 or when the cumulative penalty trip level is exceeded. It is also desirable that the various coefficients be adjusted based on a performance / energy target command.

It is desirable to inform the computer of how well the disk drive is meeting its assigned energy / performance goals. For example, this information can be passed in a status command. With this information, the computer can change the energy / performance target in some cases and inform the user that the designated target has not yet been achieved.

On the other hand, in the case of periodic access, it is also possible to improve the energy saving without affecting the performance. Periodic access is fairly common. For example, most word processing programs include an "autosave" function that saves the current working document on a disk drive at a user-specified time interval, e.g., every ten minutes. If this type of access pattern can be detected, it may enter the power saving mode more quickly and terminate the mode prior to the predicted access. This hides the power save mode return time from the user and actually saves more energy by entering the mode earlier. The access frequency method itself makes it possible to naturally sense periodic accesses. In this case, the effect on the Very Low Frequency (vlf) access is of concern. The vlf activity can be measured from the access density, but the target frequency range must be kept in mind. This is a longer time range than the mode return times. It is appropriate for the timing window to have a size of more than a few seconds to detect the autosave activity. The history buffer maintains the same vlf activity as the previous 3vlf occurrence, for example. If a predetermined pattern is detected in the history buffer, the cyclic mode is started. One simplest pattern is that the previous 3vlf occurrences are within a certain tolerance range, say 5%. Once the periodic pattern is started, the gain factors are reduced to enter the power saving mode sooner. It is assumed here that there is no performance penalty at all. These are restored when the periodic mode ends after the previous values are stored. The disk drive returns to the active state when the measured vlf is within a predetermined tolerance of the expected vlf. The tolerance is based on statistical confidence in the vlf value and also includes a power save mode return time to ensure that the disk drive is ready when vlf access occurs. If the pattern is not repeated, the cyclic mode is terminated. A simple detection mechanism for terminating the periodic mode is when access occurs within a predetermined tolerance of the expected vlf, i.e., either late or early.

The gain factors and other adjustable parameters are stored in the inactive storage so that when the power of the disk drive is turned off. This can be accomplished by using a semiconductor storage device such as a FLASH RAM or by recording the values on a disk. In the latter case, it must be completed before entering sleep mode.

Although the present invention has been particularly shown and described with reference to the preferred embodiments thereof, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the spirit and scope of the invention. And improvement can be made. Therefore, it is apparent that the above-described preferred embodiments are provided as an example of the present invention, not as a limitation of the present invention.

Claims (34)

A disk having data tracks, a plurality of electrically powered components, a data controller for accessing the components for reading and writing data on the disk, and a spindle motor for rotating the disk Disk drive components including a disk drive component, a spindle motor for rotating the disk, a head for reading data from or writing data to the disk, A method of managing power usage in a data recording disk drive having an actuator coupled to the head for moving the head, the method comprising: determining a frequency of accesses to one or more of the components; Selecting an access critical frequency from predetermined frequencies; And decreasing power to one or more of the components when the determined access frequency is less than the access threshold frequency. The method of claim 1, further comprising recovering power of a previously power-reduced component in response to an access from the controller. The disk drive of claim 1, wherein the disk drive further comprises a servo control electronics coupled to the actuator for holding the head on a track of the disk, wherein determining the frequency of accesses comprises: Wherein said step of reducing power comprises moving said actuator to a parked position to reduce power to said servo control electronics, characterized in that said step of reducing power comprises the step of moving said actuator to a parked position to reduce power to said servo control electronics A method for managing power usage in a disk drive. The method of claim 1, wherein determining the frequency of accesses comprises determining the frequency of accesses that cause the actuator to search across the disk, wherein the step of decreasing power comprises moving the actuator to a parked position roll And stopping the spindle motor. &Lt; Desc / Clms Page number 21 &gt; 2. The method of claim 1, wherein determining the frequency of accesses comprises estimating the access frequency from a density of accesses occurring in a time window. The method of claim 1, wherein the step of selecting an access critical frequency from predetermined access frequencies first determines whether a standard deviation of an access frequency pattern is less than a predetermined number of days of an average value of the pattern, And multiplying the power by the coefficient. &Lt; Desc / Clms Page number 21 &gt; 2. The method of claim 1, wherein the step of selecting an access critical frequency from predetermined access frequencies comprises: first determining whether a pattern of the access frequency is greater than a predetermined number of days of a predetermined average value of the pattern; And multiplying the minimum access frequency by a gain factor. 2. The method of claim 1, wherein said selecting an access critical frequency comprises calculating an access threshold frequency from a maximum and minimum access frequencies in a predetermined set of access frequencies. . 9. The method of claim 8, wherein selecting an access critical frequency comprises multiplying the maximum and minimum access frequencies by a predetermined gain factor. 10. The method of claim 9, further comprising adjusting the gain factor. 2. The method of claim 1, further comprising determining if the accesses are periodic patterns and recovering power of a power-reduced component prior to a next periodic access. A data controller for accessing the head for reading and writing data, a head controller for reading data from and writing data to the disk, An actuator coupled to the head for moving the head to a different track and moving it to a data-less parking position, a servo control electronics portion coupled to the actuator for holding the head on the data track during read / A method of managing power usage in a data recording disk drive operable in two power save modes, the method comprising: determining a frequency of read or write accesses; Storing a value of predetermined access frequencies; Calculating a first access threshold frequency from the stored values of predetermined access frequencies; Moving the actuator to a parked position when the read or write access frequency is less than the first threshold frequency and decreasing power to the servo control electronics to cause the disk drive to enter an operation in a first power save mode Wherein the at least one power saving mode comprises at least two power saving modes. 13. The method of claim 12, further comprising: calculating a second access threshold frequency from the stored value of predetermined access frequencies and decreasing power to the spindle motor when the frequency of the read or write accesses is less than the second threshold frequency, Further comprising the step of causing the drive to enter an operation in a second power save mode. &Lt; Desc / Clms Page number 21 &gt; 14. The method of claim 13, wherein the step of reducing power to the spindle motor includes moving the actuator to the parking position and essentially reducing power to the spindle motor, 2 power-saving mode of the at least two power-saving modes. 13. The method of claim 12, wherein the step of calculating a first access threshold frequency comprises calculating an access threshold frequency from the maximum and minimum access frequencies in a predetermined set of access frequencies. A method of managing power usage in a disk drive operable in a computer system. 16. The method of claim 15 wherein calculating the first access critical frequency comprises multiplying the maximum and minimum access frequencies by a gain factor. How to manage. 17. The method of claim 16, further comprising adjusting the gain factor. &Lt; Desc / Clms Page number 19 &gt; 18. A method of managing power usage in a disk drive operable in at least two power save modes. 18. The method of claim 17, wherein the step of adjusting the gain factor comprises selecting the gain factor within a predetermined range having a first limit value indicating a maximum power saving effect of the disk drive and a second limit value indicating a maximum performance of the disk drive Values as a performance factor having at least one of the values &lt; RTI ID = 0.0 &gt; of: &lt; / RTI &gt; A spindle motor for rotating the disk, a head for reading data from or writing data to the disk, a data controller for accessing the head for plotting and recording data, An actuator coupled to the head for moving to different tracks and to move to a dataless parking position, a servo control electronics connected to the actuator for maintaining the head on the data track during read / write of data, A method of managing power usage in a data recording disk drive capable of entry and exit operations to a storage medium, the method comprising: storing previously read or write accesses; Calculating an entry time to a power save mode that changes according to previously stored accesses after the most recent access from the stored previous ones; Selecting a performance factor having a first limit value indicating a maximum power saving effect in a disk drive and a predetermined value in a range of second limit values indicating a maximum performance of the disk drive; Wherein the mode entry action of the disk drive is adjusted as the selected performance factor using the selected performance factor to change the mode entry time calculation. . 21. The method of claim 19, wherein storing comprises storing values of previously determined access frequencies, wherein calculating the entry time into a power save mode comprises calculating an access threshold frequency from values of previously determined access frequencies And causing the power saving mode to enter when the read or write access frequency is less than the threshold frequency. 21. The method of claim 20, wherein calculating the access frequency comprises using a gain factor. 22. The method of claim 21, wherein the step of using selected performance factors to modify the mode entry time calculation comprises adjusting the gain factors by multiplying the selected performance factors by the gain factors. How to Manage Power Usage in Disk Drives. 20. The method of claim 19, wherein upon entering the power save mode, the actuator is moved to the parking position and power to the servo control electronics is reduced. The method according to claim 19, wherein, when the power saving mode is entered, the actuator is moved to the parking position and the spindle motor is stopped. 20. The method of claim 19, wherein the disk drive is capable of entering and exiting a second power save mode having less power consumption than the first power save mode, and may be varied from the stored previous accesses according to the previously stored access Calculating an entry time into a second power saving mode different from the first power saving mode entry time entering the first power saving mode; Further comprising the step of adjusting the second mode entry action of the disk drive as the selected performance factor using the performance factor selected to change the second mode entry time calculation How to manage power usage. 26. The method of claim 25, further comprising: causing the disk drive to enter the second power save mode without first entering the first power save mode. The method as claimed in claim 19, further comprising determining whether the stored read or write accesses represent a periodic pattern and ending the power save mode prior to the next periodic access. . A data controller for accessing said head for reading and writing said data; and a data controller for accessing said head for reading and writing said data, An actuator coupled to the head to move the head to different tracks and to move to a data-free parking position, a servo control electronics portion connected to the actuator to maintain the head on the data track during data read / write A method of managing power usage in a data recording disk drive capable of entering and exiting a power save mode, the method comprising: determining a frequency of read or write accesses; Storing a value of the previously determined access frequencies; Calculating an access threshold frequency from a stored value of the previously determined access frequencies; Entering the operation of the power saving mode when the read or write access frequency is less than the threshold frequency; Detecting the power saving mode entry and end times; If the power save mode duration from the detected power saver mode entry and end times is greater than the break-even time of the predetermined energy, which is the power saver mode duration required to save disk drive energy, which is usually the same as the disk drive energy required to exit the power saver mode Calculating an energy penalty if less; Accumulating the calculated energy penalties; And changing the access threshold frequency when the cumulative energy penalties exceed a predefined trip level to dynamically adjust the entry operation of the disk drive into the power save mode from the history of energy penalties. How to Manage Power Usage in Disk Drives. 29. The method of claim 28, further comprising: detecting times of the read and write accesses; If the time interval between consecutively detected access times without power saving mode entry is greater than the energy break-even time, calculating an opportunity loss penalty; Accumulating the calculated opportunity loss penalties; And changing the access threshold frequency when the accumulated opportunity loss penalties exceed a predetermined trip level to dynamically adjust the entry operation of the disk drive into the power save mode from the history of opportunity loss penalties Wherein the data recording disk drive is a disk drive. 29. The method of claim 28, wherein calculating the access critical frequency comprises calculating the access critical frequency using a gain factor from the maximum and minimum access frequencies in the set of stored access frequencies. How to manage power usage on disk drives. 30. The method of claim 28, wherein calculating the access critical frequency comprises adjusting a gain factor. 29. The method of claim 28, wherein upon activation of the power save mode, the actuator is moved to the parking position and power to the servo control electronics is reduced. 29. The method of claim 28, wherein the actuator is moved to the parked position and the spindle motor is stopped when the power saving mode is entered. 29. The method of claim 28, further comprising determining whether the accesses are periodic patterns and terminating the power save mode prior to the next periodic access accordingly. ※ Note: It is disclosed by the contents of the first application.