TW201810042A - Memory management system and method thereof - Google Patents

Abstract

Disclosed is a memory management system and method thereof. The memory management system is configured on a mother board to assist a central processing unit to manage the memory accesses by a device system. The central processing unit is connected to a memory block. The device system comprises a plurality of host devices. The memory management system is configured to: set the number of the host devices of the device system; provide a device code for each host device; set the memory size for the memory block; and distribute the memory for each host device to access and accordingly generate an allocate table. When one host device sends a request to the central processing unit for accessing data, the central processing unit replies the host device according to the allocate table, such that the host device can read data from or write data in the memory block.

Description

Memory management system and method thereof

The present invention relates to a memory management system and method thereof, and more particularly to a memory management system and method for assisting a central processor in a server to perform memory management on a device system.

As the cost of computer hardware devices has decreased and the demand for large servers has increased, virtual devices have become more and more important. The virtual device can be regarded as an interface between a software and a computer system, and can provide various resources of the computer system to the software. This software is usually an operating system, so the operating system can access resources on the computer system through the virtual device. However, multiple virtual devices can be installed on a single computer system, and different operating systems can be installed on each virtual device. For example, Microsoft's Windows operating system and Linux operating system can be installed simultaneously on different virtual devices on a computer system. In short, a virtual device must be a host device. In order to meet the different work needs of users, different host devices may be used to perform different tasks. Therefore, how to effectively allocate resources (especially memory) on a computer system to different host devices according to various job requirements of different users is still a problem of discussion space.

Embodiments of the present invention provide a memory management method for assisting a central processing unit in a server to perform memory management on a device system. Central processor A memory block is connected, and the device system is provided with a plurality of host devices. The memory management method includes: setting a number of host devices installed in the device system; providing a device code to each host device; setting a memory capacity size of the memory block; and setting a memory usable by each host device Body capacity, and based on which to establish a resource allocation table. When the host device requests the central processor to perform data storage, the central processor responds to the host device according to the resource allocation table, so that the host device can use the memory block to perform data storage.

The embodiment of the present invention also provides a memory management system, which is disposed on a motherboard of the server to assist the central processor in the server to perform memory management on a device system. The central processing unit is connected to a memory block, and the device system is provided with a plurality of host devices. The memory management system includes a memory distribution module and an entity resource agreement module, and the physical resource agreement module is connected between the memory distribution module and a plurality of host devices of the device system. The memory distribution module is configured to: set the number of host devices installed in the device system; provide the device code to each host device; set the memory capacity of the memory block; and set the use of each host device The memory capacity and the establishment of a resource allocation table. When the host device sends a request signal for reading the data, the physical resource agreement module formats the request signal and transmits the signal to the central processing unit. Then, the central processing unit responds to the request signal of the host device according to the resource allocation table, so that the host device can use the memory block to perform data storage reading.

In summary, the main purpose of the memory management system and method of the present invention is to assist the central processing unit in using memory resources for host devices in the device system. At the same time, the memory management system is directly disposed on the motherboard. That is to say, after the memory management system is customized according to the requirements of the client/enterprise, as long as it is set on the client/ The server board on the enterprise side can assist the central processor to use the memory resources to the host device in the device system.

The detailed description of the present invention and the accompanying drawings are to be understood by the claims The scope is subject to any restrictions.

100‧‧‧ memory management method

S101~S106‧‧‧Steps

M‧‧‧ memory block

V‧‧‧ device system

VM1, VM2...VMn‧‧‧ host device

B‧‧‧ motherboard

3‧‧‧Memory Management System

30‧‧‧Central processor

32‧‧‧Memory Distribution Module

34‧‧‧Physical Resource Agreement Module

FIG. 1 is a flow chart of a memory management method according to an exemplary embodiment of the invention.

2 is a schematic diagram of a resource allocation table according to an exemplary embodiment of the present invention.

3 is a block diagram of a memory management system in accordance with an exemplary embodiment of the present invention.

Various illustrative embodiments are described more fully hereinafter with reference to the accompanying drawings. However, the inventive concept may be embodied in many different forms and should not be construed as being limited to the illustrative embodiments set forth herein. Rather, these exemplary embodiments are provided so that this invention will be in the In the figures, like numerals are used to indicate like elements.

[Embodiment of Memory Management Method]

The memory management method provided in this embodiment is mainly used to assist the central processing unit in the server of the client or the enterprise to perform memory management on the device system used by the client, which will be described in more detail below.

Please refer to FIG. 1. FIG. 1 is a flowchart of a memory management method according to an exemplary embodiment of the present invention. As shown in FIG. 1 , the memory management method 100 in this embodiment is mainly implemented by the following steps: Step S101: setting the number of host devices set in the device system; Step S102: providing device code to each host Device S203: setting the memory capacity size of the memory block; step S104: Dividing the memory block into a plurality of memory sub-blocks; Step S105: setting a memory capacity that can be used by each host device, and establishing a resource allocation table; and step S106: when the host device is directed to the central processor When requesting to perform data storage, the central processing unit responds to the host device according to the resource allocation table and the device code of the host device, to allow the host device to use one of the plurality of memory sub-blocks to perform data storage reading.

Further, in step S101, the number of host devices in the device system used by the client or the enterprise terminal is set. The device system described herein is a plurality of host devices set on a client or enterprise server, and each host device in the device system can be provided to a user. When the user uses the host device machine, the physical resource of the server needs to be accessed, and in this embodiment, the physical resource of the server is the memory.

Next, in step S102, each host device in the device system is given a device code. For example, if the number of host devices set in step S102 is 8, then in step S102, the eight host devices in the device system are given device codes of 0-7, but the present invention is for device codes. The implementation is not limited.

Next, in step S103, the memory capacity size of the memory block is set. This memory block is the physical resource of the aforementioned server. In this embodiment, the memory block is a Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM). For the convenience of the following description, in this step, the memory capacity of the memory block is set to 512 GB. In addition, in step S104, the memory block is further divided into a plurality of memory sub-blocks. That is to say, in this step, the memory block having a memory capacity of 512 GB will be further divided into a plurality of memory sub-blocks. The memory capacity of these memory sub-blocks can be 1 GB, 2 GB, 4 GB, 8 GB, 16 GB, etc., and so on. Thus, the memory sub-block There will be different memory capacity sizes, and it is possible to generate a plurality of memory sub-blocks of the same memory capacity.

The step S104 of dividing the memory block into a plurality of memory sub-blocks is mainly based on the fact that the client or the enterprise may separately use different host devices to perform different work contents. After the memory block is divided into a plurality of memory sub-blocks having a memory capacity of 1 GB, 2 GB, 4 GB, 8 GB, 16 GB, etc., in step S105, the memory usable by each host device is set. Body capacity. It should be noted that the memory capacity that can be used for each host device corresponds to the memory capacity of the memory sub-block. According to the above example, the memory capacity of each host device must be 1 GB. 2GB, 4GB, 8GB or 16GB...etc.

In step S105, a resource allocation table is further established, and the resource allocation table mainly records the correspondence between the device code of each host device and the memory capacity allocated by each host device. Please refer to FIG. 2. FIG. 2 is a schematic diagram of a resource allocation table according to an exemplary embodiment of the present invention. According to the resource allocation table shown in FIG. 2, the memory capacity allocated to the host device having the device code of "0" is 16 GB, and the memory capacity allocated to the host device having the device code "1" is 64 GB, and the device code is The memory capacity allocated to the "7" host device is 2 GB. However, the allocation manner of the memory in the resource allocation table shown in FIG. 2 is only used to exemplify the resource allocation table in the memory management method provided in this embodiment, and is not intended to limit the present invention. In addition, in this embodiment, the resource allocation table also records the memory capacity of the set memory block (ie, in FIG. 2, the memory capacity of the memory block set by the record is 512 GB. And the number of host devices in the device system (i.e., in Figure 2, the number of host devices recorded in the device system is 8).

Finally, in step S106, when the user operates a host device in the device system to operate, the host device requests the central processor to perform data storage. At this time, the central processor recognizes the device code of the host device to respond to the host device according to the foregoing resource allocation table and the identified device code. The response action referred to here is to allow the host device to use one of the plurality of memory sub-blocks. The data storage is read, wherein the memory capacity of the memory sub-block that the host device is allowed to use is determined by the content recorded in the foregoing resource allocation table. In addition, as described above, the memory sub-blocks may have different memory capacity sizes, but it is also possible to generate a plurality of memory sub-blocks of the same memory capacity. Therefore, in the content recorded in the resource allocation table, different device codes can be assigned to the same memory capacity size.

Notably, in this embodiment, the device system may be a physical device system or a virtual device system. That is to say, the memory management method provided by this embodiment can be applied to a virtual device system in which a plurality of virtual devices are established. The virtual device referred to herein is a virtual machine (VM) generally referred to as a virtual machine.

[An embodiment of a memory management system]

This embodiment provides a memory management system for performing the foregoing embodiments. Please refer to FIG. 3. FIG. 3 is a block diagram of a memory management system according to an exemplary embodiment of the present invention.

The memory management system 3 provided in this embodiment is mainly used to assist the central processing unit 30 of the server to perform memory management on the device system V used by the server. As shown in FIG. 3, the central processing unit 30 on the motherboard B is connected to a memory block M, and the device system V includes a plurality of host devices VM1, VM2, ... VMn, where n is a positive integer. In this embodiment, the memory block is also a Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM), and the number of host devices included in the device system V (ie, , the value of n) depends on the needs of different clients or enterprises.

The memory management system 3 provided in this embodiment includes a memory allocation module 32 and an entity resource agreement module 34. The memory allocation module 32 and the physical resource agreement module 34 are directly disposed on the motherboard B of a server. In this embodiment, the physical resource agreement module 34 can be a PCI-E bus, a single root I/O virtualization (SRIOV) bus, or a multipath I/O virtualization (Multi Root). I/O virtualization, MRIOV) bus, especially with single-path I/O virtualization (Single Root) I/O virtualization, SRIOV) bus is better. In addition, the memory distribution module 32 is implemented by a field-programmable gate array (FPGA) chip, and the FPGA chip is a reprogrammable chip. Therefore, the embodiment can provide a user-customized system. Memory management system.

Further, by designing the memory allocation module 32 (ie, the FPGA chip), the number of host devices VM1, VM2, . . . VMn disposed in the device system V can be set, and the device code is provided to each of the host devices VM1 and VM2. ...VMn. Then, according to the capacity of the memory block required by the client or the enterprise, the memory allocation module 32 sets the memory capacity of the memory block, and divides the memory block into a plurality of memory sub-blocks. (not shown). Finally, by designing the memory allocation module 32, the memory capacity that can be used by each of the host devices VM1, VM2, ... VMn can be set, and a resource allocation table can be established accordingly. The resource allocation table mainly records the correspondence between the device code of each host device and the memory capacity allocated by each host device. Please refer to FIG. 2 and the corresponding description in the foregoing embodiment, for the resource allocation table. The content will not be described again.

When the user operates a host device VM1, VM2 ... VMn in the device system V to operate, the host devices VM1, VM2, ... VMn request the central processor to perform data storage reading. At this time, the central processor recognizes the device code of the host device to respond to the host device according to the foregoing resource allocation table and the identified device code. The response action referred to herein is to allow the host device to use one of a plurality of memory sub-blocks (not shown) for data storage, wherein the host devices VM1, VM2, ... VMn are allowed to use the memory. The size of the memory capacity of the block is determined by the content recorded in the aforementioned resource allocation table.

As mentioned above, the memory sub-blocks will have different memory capacity sizes, but it is also possible to generate a plurality of memory sub-blocks of the same memory capacity. Therefore, in the content recorded in the resource allocation table, different device codes can be assigned to the same memory capacity size. In summary, in this embodiment, the memory capacity allocated to each host device VM1, VM2, ... VMn is necessarily equal to a plurality of memory sub-regions. The memory capacity of one of the blocks. Thus, the manner in which the memory block is cut into a plurality of memory sub-blocks is related to the memory capacity required or allocated by each host device VM1, VM2, ... VMn.

In addition, it should be noted that, in this embodiment, the device system V may be a physical device system or a virtual device system. That is to say, the memory management system provided in this embodiment can be used for memory management on a virtual device system in which a plurality of virtual devices are established. The virtual device referred to herein is a virtual machine (VM) generally referred to as a virtual machine. In recent years, more and more companies have adopted a virtual device system in order to achieve simplified management. According to the actual use requirements, each virtual machine in the virtual device system adopted by different enterprises is given different work contents, and the present invention can customize a set of memories for different practical needs of different enterprise users. The volume management system and the management method associated therewith enable each virtual device in the virtual device system employed by any enterprise user to be easily and efficiently allocated to usable memory resources.

[Possible effects of the examples]

In summary, the memory management method and system thereof provided by the present invention can assist a central processor of a server to perform memory management on the device system used by the server. The memory management method and system thereof provided by the present invention can be customized according to the needs of the user, and the memory management system provided by the present invention is directly disposed on the motherboard of the server, so that the user can A suitable memory management system is obtained for the host device of different working contents in the device system, and the memory management system is set to operate on the server board of the server.

The above description is only an embodiment of the present invention, and is not intended to limit the scope of the invention.

Claims (10)

A memory management method for assisting a central processing unit of a server to perform memory management of a device system, the central processing unit of the server is connected to a memory block, and the device system is provided with a plurality of hosts The device management method includes: setting a quantity of the host device disposed in the device system; providing a device code to each of the host devices; setting a memory capacity of the memory block; and setting each a memory capacity that can be used by the host device, and a resource allocation table is established; wherein, when the host device requests the central processor to perform data storage, the central processor responds according to the resource allocation table The host device enables the host device to use the memory block for data storage. The memory management method of claim 1, wherein the resource allocation table records a correspondence between the device code of each host device and a memory capacity allocated by each of the host devices. The memory management method of claim 1, further comprising: dividing the memory block into a plurality of memory sub-blocks; wherein, when the host device requests the central processor to perform data storage, The central processor responds to the host device according to the resource allocation table and the device code of the host device, to allow the host device to use one of the plurality of memory sub-blocks to perform data storage, and is allowed to use. The memory capacity of one of the plurality of memory sub-blocks is equal to the memory capacity allocated by the host device. The memory management method of claim 1, wherein the resource allocation table further records a memory capacity size of the memory block that is set, and a number of the host devices in the device system. The memory management method according to claim 1, wherein the device system is a virtual device system, and a plurality of virtual host devices are disposed. A memory management system is provided on a motherboard of a server for assisting a central processing unit of the server to perform memory management of a device system, wherein the central processing unit of the server is connected to a memory block And the device system is provided with a plurality of host devices, the memory management system includes: a memory distribution module, configured to: set the number of the host devices disposed in the device system; provide a device code for each a host device; setting a memory capacity of the memory block; setting a memory capacity that can be used by each of the host devices, and establishing a resource allocation table; and an entity resource agreement module The memory distribution module and the plurality of host devices of the device system; wherein, when the host device sends a request signal to perform data storage, the physical resource agreement module performs format conversion on the request signal Transmitting to the central processor, the central processor responds to the request signal of the host device according to the resource allocation table, so that the host device Using this data memory blocks were kept reading. The memory management system of claim 6, wherein the resource allocation table records a correspondence between the device code of each of the host devices and the memory capacity allocated by each of the host devices. The memory management system of claim 6, wherein the memory allocation module is further configured to: divide the memory block into a plurality of memory sub-blocks; wherein the host device sends the request signal to perform When the data is read, the physical resource agreement module converts the request signal into a format and transmits the data to the a central processing unit, the central processor responding to the request signal of the host device according to the resource allocation table and the device code of the host device, to allow the host device to use one of the plurality of memory sub-blocks for data storage The memory capacity of one of the plurality of memory sub-blocks that are read and allowed to be used is equal to the memory capacity allocated by the host device. The memory management system of claim 6, wherein the resource allocation table further records the size of the memory capacity of the memory block that is set, and the number of the host devices in the device system. The memory management system of claim 6, wherein the device system is a virtual device system, and a plurality of virtual host devices are disposed.