CN1540473B - Automatic adjusting device and method for computer system efficiency - Google Patents

Abstract

An automatic adjusting device and method for system efficiency installed on computer motherboard. The device at least comprises a plurality of groups of efficiency detection elements which are respectively connected with the bus layout patterns of the mainboard elements so as to detect the working conditions of the mainboard elements through the data flow in the bus layout patterns. In addition, a efficiency control chip is respectively connected with the mainboard components and can readjust the working speed of the mainboard components according to the signals of the efficiency detection components, wherein the efficiency control chip can judge whether the working condition of a certain mainboard component is busy or not so as to determine whether to increase or decrease the working speed of the mainboard component or not.

Description

Computer system performance automatic adjustment device and method thereof

(1) Technical field

The present invention relates to a computer system performance automatic adjustment device and its method, in particular to an automatic adjustment device installed on the main board of the computer system, so as to detect the working status of each component on the main board, and perform Adjustment of resources to optimize the overall performance of the computer system.

(2) Background technology

With the continuous progress and development of electronic technology, multimedia computers combined with high picture quality and high sound quality are generally liked and used by consumers, and the degree of popularization of computers has been accelerated. The increase in consumption of computer-related products has also driven the development of related industries more vigorously and rapidly. As we all know, the performance of a computer system mainly depends on the motherboard used and the performance of various components installed on the motherboard such as central processing unit, memory, chipset, etc.

As far as the new generation of central processing unit is concerned, due to its more sophisticated and complex core structure, it can provide a relatively high execution clock and more powerful computing functions. As for the synchronous dynamic random access memory developed by PC66, PC100, and PC133, because it has the ability to process more data, the ability of the computer to process multimedia data is greatly improved, and the memory processing space of the personal computer is increased. In addition to the above-mentioned central processing unit and memory, other components on the motherboard, such as the north and south bridge chipsets, graphics accelerator cards, network cards, data cards, and other components are also becoming more and more detailed with the development of new technologies. Produce more powerful operating performance and meet the needs of consumers.

However, as the structures of various components become more and more complex, how to effectively integrate these components to produce better performance has become a rather difficult challenge. In particular, when the computer system is in operation, the components on the motherboard are often busy in different degrees depending on the programs being executed. For example, when the executed program mainly focuses on the calculation of the central processing unit, the central processing unit will be in a relatively busy working state; or when the executed program mainly focuses on the transmission and processing of image data, the graphics accelerator card Will be in a busy work situation.

In other words, when the program is actually executing, sometimes the central processing unit is busy, sometimes the memory is busy, and sometimes the peripheral components on the PCI interface card slot are busy, and even in many cases, some components will be in A situation where there is absolutely nothing to do. Therefore, if various components can be effectively integrated, and resources such as power supply and bandwidth can be allocated according to their working conditions, the overall operating performance of the computer system can obviously be further improved. For example, when the computer system executes a memory-based program, the resource supply enjoyed by the memory can be increased; or when the central processing unit has nothing to do, the speed of the central processing unit is slowed down and the power is reduced. Consumption, in order to improve the performance of the overall computer system.

It is worth noting that even in the traditional production line, manufacturers will make trade-offs and compromises based on operating experience, and preset the program execution speed, operating voltage, and some related settings of system performance for each component on the motherboard. However, since these setting values are not dynamically adjusted according to the busy status of each component when the computer system is in operation, or the allocation of supply resources is changed. In this way, it often results in the system setting the priority order of sharing resources, which is different from the working status of each component in actual operation. In this case, in addition to the failure to optimize the system performance, it also increases unnecessary power consumption and shortens the life of the product.

(3) Contents of the invention

In order to solve the above problems, the purpose of the present invention is to provide a system performance automatic adjustment device and method installed on the computer motherboard, so as to detect the working status of each motherboard component and adjust it to achieve the best performance of the computer system. change.

The device at least includes multiple sets of efficiency detection components, each connected to the bus of the corresponding motherboard components, so as to detect the working conditions of the motherboard components from the data traffic of these buses. In addition, a performance control chip is respectively connected to these motherboard components, and can respond to the signal of the performance detection component to judge whether these motherboard components are in a busy state, and adjust the motherboard components to use system resources according to the judgment result prioritization in order to optimize the overall performance of the motherboard.

In a preferred embodiment, the above-mentioned data traffic is the number of data access cycles or the number of instructions passing through the bus per unit time. And, above-mentioned bus at least comprises a PCI bus, is connected between the south bridge chip and the PCI slot; An AGP bus, is connected between the north bridge chip and the AGP slot; A memory bus, is connected between the north bridge chip and the memory ; and a CPU bus connected between the CPU and the north bridge chip. In addition, the motherboard components connected with the performance control chip include at least a central processing unit, a north bridge chip, a south bridge chip, an AGP slot, a PCI slot, and a motherboard power supply.

(4) Description of drawings

The preferred embodiment of the present invention will be described in more detail with the help of the following figures in the following explanatory text:

Fig. 1 shows the automatic adjustment device of the computer system installed on the motherboard among the present invention; and

FIG. 2 shows a method for automatically adjusting performance of a computer system disclosed in the present invention.

(5) specific implementation

Please refer to FIG. 1 , which is a schematic diagram of a motherboard. It should be noted that, for the convenience of understanding, here is a typical mainboard design on the market as an example for illustration. However, those who are familiar with this technology should be able to understand the characteristics and spirit of the present invention, and are not limited to the layout of the motherboard in this figure. As shown in the figure, the components on the motherboard 10 include a central processing unit (CPU) 12 installed on the processor socket, a north bridge chip 14 and a south bridge chip 16 . Among them, the central processing unit 12 has processing, control and storage circuits, so as to execute and control various computing functions of the computer system. As for the functions of the north bridge chip 14 and the south bridge chip 16, the instructions of the central processing unit 12 are communicated to other components on the motherboard for execution by the internal circuit, or the information on each component is passed back to the central processing unit 12 for supply Reference and use.

The components on the motherboard 10 include a memory 18 located under the north bridge chip 14 for storing data and related instructions before and after processing by the computer system. As for the left side of the north bridge chip 14 , there are AGP slots 20 and three PCI slots 22 . Wherein, the AGP slot 20 is used to install a graphics accelerator card, so as to process relevant image data faster and more perfectly. As for, the PCI slot 22 can be used to install other peripheral components such as network cards, data cards, etc. . In addition, above the north bridge chip 14 and the central processing unit 12 , there is a motherboard power supply 24 for supplying power required by each component on the motherboard 10 . There are a considerable number of bus layout patterns on the motherboard 10 to connect the above-mentioned components and various slots.

In order to detect the working status of each component on the mainboard 10 in real time and optimize the allocation of system resources, the present invention installs a system performance automatic adjustment device on the mainboard 10 . The system performance automatic adjustment device includes several performance detection components (Performance Monitor Circuit; PMC) 26 and a performance control chip (Performance Control Chip) 28 . Wherein, each performance detection component 26 is connected to a specific bus layout pattern on the motherboard 10, so as to detect the working condition of the specific component by the data flow in the bus layout pattern. Generally speaking, when monitoring data traffic, the number of data access cycles (data access cycles) passing through a specific bus layout per unit time can be used as the basis for measurement, or the number of instructions passing through a specific bus per unit time (command) As a basis for monitoring the working status of specific components.

For example, on the PCI bus between the south bridge chip 16 and the PCI slot 22, an performance detection element 26 can be installed, so as to detect the data flow of this part of the circuit layout, as the basis for judging the working state of the peripheral components; or, in An performance detection element 26 is also installed on the AGP bus between the north bridge chip 14 and the AGP slot 20, so that the working state of the graphics accelerator card is detected; in like manner, the performance detection element 26 can be installed between the north bridge chip 14 and the memory 18 on the memory bus to monitor the data flow of the memory, or installed on the CPU bus between the CPU 12 and the northbridge chip 14 to monitor the working status of the CPU 12.

As for the performance control chip 28, it is connected to each component on the motherboard 10 respectively, and compares the working conditions of each component according to the data flow transmitted by the performance detection component 26, and redistributes the system resources shared by each component. To adjust the working efficiency of each component and improve the overall performance of the motherboard. As shown in FIG. 1 , the performance control chip 28 is connected to each performance detection element 26 by a circuit so as to receive the data flow of the monitored element. In addition, the performance control chip 28 is connected to the central processing unit 12, the north bridge chip 14, the south bridge chip 16, the AGP slot 20, the PCI slot 22 and the motherboard power supply 24 respectively via the bus, so as to detect the components according to the performance 26 transmitted data, judge and compare the working status of each component, and carry out real-time adjustment actions.

In a preferred embodiment, the performance detection element 26 includes at least a counter (counter), so as to monitor the selected bus, so as to measure the number of instructions or the number of data accesses flowing through the bus in a unit time . For example, the unit time of measurement can be set as 100 microseconds (ms), so that the performance detection element 26 transmits the accumulated flow times to the performance control chip 28 every 100 microseconds. As for the performance control chip 28 mentioned above, it may include a comparator and a register. Wherein, the comparator compares the counting data sent by the performance detection element 26 with the flow setting value in the register. When the flow count of a certain component on the motherboard exceeds the flow setting value of the component, it is determined that the component is already in a busy state. For example, the setting value of the flow rate of a certain peripheral component on the PCI slot can be set to 3000 times/100 microseconds. Once the flow count monitored by the performance detection element 26 exceeds the time limit, it is determined that the peripheral element is in a busy state.

After determining the working status of each component on the motherboard, the performance control chip 28 can use the control signal to adjust the working speed of each component through the timing chip. That is, the working speed of each component is adjusted by adjusting the timing signal (clock) of each component on the motherboard. When a certain motherboard component is in a busy state, the working speed of this component can be accelerated by adjusting the timing signal; When the value is set, it can be determined that the component is not busy at the moment, and its working speed can be reduced by adjusting its timing signal.

Or, the performance control chip 28 can also send a control signal to modify the expected value in the temporary register of the north bridge chip 14 or the south bridge chip 16, so as to readjust the priority order (priority) of each component to use the system resources. Generally speaking, according to the original setting of the motherboard when it leaves the factory, the priority order of enjoying system resources is the central processing unit, the graphics accelerator card on the AGP slot, and the peripheral components on the PCI slot. However, when the system performance automatic adjustment device provided by the present invention is used, the prioritization of system resources will be re-ranked according to the traffic count monitored by the performance detection component 26 . For example, when the graphics accelerator card is in a busy state and the CPU is in a low workload state, the order of the AGP slots can be immediately increased to provide more system resources.

In addition to adjusting the priority of each component on the motherboard, the performance control chip 28 can also adjust the bandwidth to re-allocate system resources. For example, when a certain component on the motherboard is in a busy state, the performance control chip 28 can adjust the amount of data transmitted by the component per unit time through the control signal, that is, improve its ability to transmit data. In this way, by resetting the various components of the motherboard, the working efficiency of the busy components can be improved, while the working speed of the non-busy components can be reduced, so as to avoid unnecessary waste of system resources.

Moreover, since the performance control chip 28 is also connected to the motherboard power supply 24, the purpose of system optimization can also be achieved by adjusting the power supply to each component. For example, when a certain component is in a busy state, its operating voltage can be increased under the condition of specification limit (spec), so that the working speed of this component can be improved. Conversely, when a certain component is in a state of having nothing to do, its operating voltage is slightly lowered so as to reduce the working speed of the component. In this way, in addition to reducing the power consumption to achieve the purpose of power saving, it can also increase the cooling effect of the system, thereby increasing the service life of the components.

In order to facilitate the understanding of the related practice of adjusting the working performance of each component of the motherboard in real time in the present invention, as well as the optimization effect that can be achieved. Please refer to FIG. 2 , which provides a method for automatically adjusting performance of a computer system disclosed in the present invention. This method mainly includes the following steps. Firstly, execute a program based on at least one component on the motherboard (step 50), and measure the data traffic of the component in a busy state (step 52). According to the measured data flow, define the flow setting value of the component in busy state (step 54), and store this part of setting value in the temporary register of the above-mentioned performance control chip 28 . After setting the flow setting values of each component in sequence, the data flow of each component on the mainboard can be detected during the operation of the mainboard (step 56). When the data traffic of the first part of the components on the main board exceeds its corresponding flow setting value (step 58), then as before, the working speed of the first part of the components is increased (step 60). Conversely, when the data flow rate of the second part of the element is lower than its corresponding flow rate setting value, the working speed of the second part of the element can be reduced (step 62).

The present invention has considerable advantages:

(1) Since the performance detection components are used to monitor each component on the motherboard, the working status of each component can be reflected in real time, and the system resources of each component can be re-allocated accordingly, and the computer system can be obtained. The effect of performance optimization;

(2) Since the computer system can automatically adjust according to the current operating conditions of the program to optimize the performance of the system, it can reduce the operating noise of the computer by reducing the working speed of the currently less busy components. , at the same time can reduce unnecessary power consumption, thereby prolonging the service life of the product;

(3) Since the computer system can automatically adjust dynamically along with the working conditions of the components, the execution efficiency of the overall computer system will be faster than that of a traditional computer, and when the computer is in a rest state, the computer system using the device of the present invention will Can further save power resources.

The present invention is described above with preferred embodiments, and those skilled in the art can make various equivalent changes or equivalent replacements without departing from the spirit of the present invention, and the scope of patent protection should be based on the appended rights Requirements and their equivalent fields.

Claims (13)

1. A system performance automatic adjustment device is mounted on a computer motherboard, and is characterized in that it at least includes: Multiple groups of performance detection components are respectively connected to the bus of each motherboard component, so as to detect the working conditions of these motherboard components from the data flow in the bus; and The performance control chip is connected to the performance detection components by wires, and the performance control chip readjusts the working speed of the motherboard components according to the data flow detected by the performance detection component, wherein the performance control chip It can judge whether the working condition of a certain motherboard component is busy, so as to decide whether to increase or decrease the working speed of the motherboard component. 2. The device according to claim 1, wherein the bus comprises at least: PCI bus, connected between the south bridge chip and the PCI slot; AGP bus, connected between the north bridge chip and the AGP slot; a memory bus, connected between the north bridge chip and the memory; and The CPU bus is connected between the CPU and the north bridge chip. 3. The device according to claim 1, wherein the mainboard elements connected to the performance control chip include at least a central processing unit, a northbridge chip, a southbridge chip, an AGP slot, a PCI slot, and a mainboard power supply. 4. The device according to claim 1, wherein the performance detection element has a counter capable of measuring the number of instructions or the number of data accesses flowing through the selected bus layout pattern in a unit time. 5. The device according to claim 1, wherein the efficiency control chip further comprises: a temporary register for storing the flow setting values of these motherboard components; A comparator, which can compare the data flow transmitted by the performance detection component with the flow setting value in the temporary register, when the data flow of a certain motherboard component exceeds the flow setting of the motherboard component value, it is determined that the motherboard component is in a busy state. 6. The device according to claim 1, wherein the performance control chip is connected to a motherboard power supply, and can adjust the working speed of the motherboard components by adjusting the power supply to each of the motherboard components . 7. A system performance automatic adjustment device installed on a computer motherboard, characterized in that it at least includes: a plurality of counters, each connected to a bus of a corresponding motherboard component, capable of measuring the amount of data flowing through the selected bus per unit of time; and A performance control chip is respectively connected to these motherboard components, and the performance control chip includes: a temporary register for storing the flow setting values of these motherboard components; A comparator, which can compare the data flow transmitted by the counter with the flow setting value in the temporary register. When the data flow of a certain mainboard component exceeds the flow setting value of the mainboard component, Then it is determined that the motherboard component is in a busy state. 8. The device according to claim 7, wherein the bus comprises at least: PCI bus, connected between the south bridge chip and the PCI slot; AGP bus, connected between the north bridge chip and the AGP slot; a memory bus, connected between the north bridge chip and the memory; and The CPU bus is connected between the CPU and the north bridge chip. 9. The device according to claim 7, wherein the mainboard elements connected to the performance control chip at least include a central processing unit, a northbridge chip, a southbridge chip, an AGP slot, a PCI slot, and a mainboard power supply. 10. A computer system performance automatic adjustment method, which detects and adjusts its work performance for the data flow of each component on the mainboard, it is characterized in that, comprising the following steps: (1) Execute a program using at least one component on the motherboard; (2) Measure the data flow of the above-mentioned component in working condition; (3) Define the flow setpoint at which the component is in a busy state; Detect the data flow of all the components on the motherboard; When the first part of the data flow of the component on the motherboard exceeds its corresponding flow setting value, and the second part of the data flow of the component is lower than its corresponding flow setting value, increase the working speed of the component and decrease it. The operating speed of the second part of the element. 11. The method according to claim 10, wherein the data traffic is the number of data access cycles passing through the specific element per unit time. 12. The method according to claim 10, wherein the data flow is the number of instructions passing through the specific element per unit time. 13. The method according to claim 10, wherein the components on the motherboard at least include a central processing unit, a north bridge chip, a south bridge chip, an AGP slot, a PCI slot, and a motherboard power supply.

Images