CN104461928B - Divide the method and device of cache - Google Patents

Abstract

The embodiment of the invention discloses a method and device for dividing a high-speed cache, relates to the field of computer technology, and can improve the performance of a virtual machine when the virtual machine is started and copied. The method of the present invention includes: when none of the cache data in the cache group accessed by the first virtual machine hits, judging the virtual machine management operation state VMS of the first virtual machine; when the VMS of the first virtual machine is the first state, replace the least recently used cache data belonging to the second virtual machine in the cache group. The invention is suitable for virtualization environment.

Description

Method and device for dividing cache memory

technical field

The invention relates to the field of computer technology, in particular to a method and device for dividing cache memory.

Background technique

Virtualization technology is one of the popular technologies in today's enterprises, and it has been widely used in research fields such as server integration, system migration, and load isolation. The current virtualization technology can strictly divide physical resources such as multi-core processors, memory, and input/output I/O, eliminate competition for physical resources among virtual machines, and ensure the isolation of virtual machine functions and performance. However, the current virtualization technology does not divide shared hardware resources such as on-chip multi-level caches, and implicit competition among virtual machines for these hardware resources will lead to performance degradation of some virtual machines. For example, some virtualization applications on a virtual machine are sensitive to the capacity of the LLC (Last-Level Cache, last-level shared cache). When the LLC capacity available to the virtual machine decreases due to competition, the performance of the virtual machine lowered.

In order to change the above situation and improve the performance of the virtual machine, the researchers proposed a shared cache partition mechanism. In the prior art, the modified LRU (Least Recently Used, least recently used algorithm) and page coloring method are used to divide the cache to limit the shared cache capacity used by different virtual machines, thereby reducing competition conflicts on the shared cache.

The prior art has at least the following problems: In a multi-tenant data center, the effects of different management operations performed by each user on virtual machines will be superimposed, for example, in the process of starting virtual machines and performing replication operations using virtual machines, when a large number of virtual machines When the virtual machines are started at the same time in a short period of time, or the disk image copied by the virtual machine is large and takes a long time, due to the superposition of operation effects, the performance of the virtual machine will be severely degraded. However, the cache division method in the prior art cannot improve the performance of the virtual machine well when the virtual machine is started and copied, and the performance of the virtual machine is low.

Contents of the invention

Embodiments of the present invention provide a method and device for dividing a cache, which can solve the problem of low performance of a virtual machine when the virtual machine is started and copied.

In order to achieve the above object, embodiments of the present invention adopt the following technical solutions:

In a first aspect, embodiments of the present invention provide a method for dividing a cache, including:

When none of the cache data in the cache group accessed by the first virtual machine hits, determine the virtual machine management operation state VMS of the first virtual machine;

When the VMS of the first virtual machine is in the first state, replace the least recently used cache data belonging to the second virtual machine in the cache group.

With reference to the first aspect, in a first possible implementation manner, before determining the virtual machine management operation status VMS of the first virtual machine, the method further includes:

A virtual machine VM register is added, and the data structure of the VM register includes: the VMS and a virtual machine identifier VMID.

Further, before determining the virtual machine management operation status VMS of the first virtual machine, the method further includes:

When the first virtual machine performs a startup operation or a copy operation, set the VMS of the first virtual machine to a first state; when the first virtual machine does not perform a startup operation or a copy operation, set the first The VMS of the virtual machine is in the second state.

Optionally, when the VMS of the first virtual machine is in the first state, before replacing the least recently used cache data belonging to the second virtual machine in the cache group, the method further includes:

judging whether the cache failure rate of the second virtual machine is greater than a failure rate threshold;

If the cache failure rate of the second virtual machine is greater than the failure rate threshold, replace the least recently used cache data in the cache group;

If the cache failure rate of the second virtual machine is not greater than the failure rate threshold, then determine the number of cache blocks belonging to the second virtual machine in the cache group.

With reference to the first possible implementation, in the second possible implementation, after adding the virtual machine VM register, the method further includes:

The flag bit of the cache address is extended, and the VMID is added to the flag bit, and the cache data in the cache address belongs to the virtual machine corresponding to the VMID.

Optionally, after the determination of the number of cache blocks belonging to the second virtual machine in the cache group, the method further includes:

judging whether the number of cache blocks of the second virtual machine is less than a number threshold;

If the number of cache blocks of the second virtual machine is less than the number threshold, replace the least recently used cache data in the cache group;

If the number of cache blocks of the second virtual machine is not less than the number threshold, replace the least recently used cache data belonging to the second virtual machine in the cache group.

With reference to the first aspect or the first possible implementation of the first aspect, in a third possible implementation, when the VMS of the first virtual machine is in the second state, replace the least recent The cached data to use.

With reference to the first aspect or any possible implementation manner, in a fourth possible implementation manner, the second virtual machine is used to control the physical input and output I/O resources of the virtual environment, and the first virtual machine Interact and start the first virtual machine.

In a second aspect, an embodiment of the present invention provides a device for dividing a cache, including:

A judging unit, configured to judge the virtual machine management operation state VMS of the first virtual machine when none of the cache data in the cache group accessed by the first virtual machine hits;

The replacement unit is configured to replace the least recently used cache data belonging to the second virtual machine in the cache group when the VMS of the first virtual machine is in the first state.

With reference to the second aspect, in a first possible implementation manner, the device further includes:

The adding unit is used to add a virtual machine VM register, and the data structure of the VM register includes:

The VMS and virtual machine identifier VMID.

Further, the apparatus further includes: a setting unit, configured to set the VMS of the first virtual machine to a first state when the first virtual machine performs a startup operation or a copy operation; When the first virtual machine does not execute the start operation or the copy operation, set the VMS of the first virtual machine to the second state.

Optionally, the judging unit is further configured to judge whether the cache failure rate of the second virtual machine is greater than a failure rate threshold; the replacement unit is also configured to determine whether the cache failure rate of the second virtual machine is greater than the failure rate threshold. rate threshold, replace the least recently used cache data in the cache group; the device also includes:

A determining unit, configured to determine the number of cache blocks belonging to the second virtual machine in the cache group when the cache failure rate of the second virtual machine is not greater than the failure rate threshold.

With reference to the first possible implementation manner, in a second possible implementation manner, the device further includes:

The extension unit is configured to extend the flag bit of the cache address, and add the VMID added by the adding unit to the flag bit, and the cache data in the cache address belongs to the virtual machine corresponding to the VMID.

Optionally, the judging unit is further configured to judge whether the number of cache blocks of the second virtual machine is smaller than the number threshold determined by the determining unit; When the number of cache blocks is less than the number threshold determined by the determination unit, replace the least recently used cache data in the cache group; the replacement unit is also configured to when the number of cache blocks of the second virtual machine is not less than the specified When the number threshold determined by the determining unit is exceeded, replace the least recently used cache data belonging to the second virtual machine in the cache group.

With reference to the second aspect or the first possible implementation manner of the second aspect, in a third possible implementation manner, the replacement unit is further configured to: when the VMS of the first virtual machine is in the second state, Replaces the least recently used cache data in the cache group.

With reference to the second aspect or any possible implementation manner, in a fourth possible implementation manner, the second virtual machine is used to control the physical input and output I/O resources of the virtual environment, and the first virtual machine Interact and start the first virtual machine.

The method and device for dividing the last-level shared cache provided by the embodiments of the present invention are similar to the problems in the prior art that the performance of the virtual machine cannot be well improved when the virtual machine is started and copied, and the performance of the virtual machine is low. In contrast, in the present invention, when the virtual machine is started and copied, the last-level shared cache between the first virtual machine and the second virtual machine is divided to reduce the cache capacity occupied by the second virtual machine. Since the second virtual machine is not sensitive to the cache capacity, even if the cache failure rate is high, its performance is not greatly affected, so by reducing the cache capacity occupied by the second virtual machine, increase the cache occupied by the first virtual machine Capacity, to reduce the contention for the cache among the virtual machines, thereby improving the performance of the virtual machine startup or replication.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings that need to be used in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For Those of ordinary skill in the art can also obtain other drawings based on these drawings without making creative efforts.

Fig. 1 is a flow chart of a method provided by an embodiment of the present invention;

Fig. 2 is the flow chart of the method provided by another embodiment of the present invention;

FIG. 3 is a memory access address structure diagram provided by another embodiment of the present invention;

FIG. 4 is a cache structure diagram provided by another embodiment of the present invention;

FIG. 5 is a cache structure diagram provided by another embodiment of the present invention after using traditional LRU data replacement;

FIG. 6 is a cache structure diagram provided by another embodiment of the present invention after using improved LRU data replacement;

Fig. 7 and Fig. 8 are schematic diagrams of the device structure provided by another embodiment of the present invention;

Fig. 9 is a schematic structural diagram of a device provided by another embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

An embodiment of the present invention provides a method for dividing high-speed cache, which is used in a virtualization environment. The first virtual machine and the second virtual machine run on the same virtual platform, and the second virtual machine is a privilege running on a virtual machine monitor. A virtual machine, which can control physical I/O resources and interact with a first virtual machine at the same time. The first virtual machine is a guest virtual machine, and multiple first virtual machines can be started by multiple second virtual machines. As shown in Figure 1, the method includes:

101. When none of the cached data in the cache group accessed by the first virtual machine is hit, the cache controller judges a VMS (Virtual Management State, virtual machine management operation state) of the first virtual machine.

It should be noted that, in the processor, a virtual machine VM register is added for each processor core, and the VM register structure includes: VMS and VMID (Virtual Machine Identifier, virtual machine identifier). Then, the flag bit of the cache address is extended, and the VMID is added to the flag bit, wherein the cache data in the cache address belongs to the virtual machine corresponding to the VMID.

Wherein, the virtual machine monitor responds to the management operation of the virtual machine, and sets the state of the VMS according to the operation content of the virtual machine. When the first virtual machine performs a startup operation or a copy operation, the virtual machine monitor sets the VMS at this time as the first state; when it is not the first virtual machine that performs a startup operation or a copy operation, the virtual machine monitor sets the current VMS VMS is the second state.

Furthermore, when the first virtual machine accesses the cache, if the first virtual machine accesses the cache, none of the cached data in the cache group is hit. At this time, it is necessary to access the memory, find the corresponding data, and then replace it with the data in the cache , the cache controller judges the state of the VMS at this time.

It should be noted that the cache group mentioned in the embodiment of the present invention is a cache group determined according to a memory access address when a virtual machine performs a memory access operation.

102. When the VMS of the first virtual machine is in the first state, the cache controller replaces the least recently used cache data belonging to the second virtual machine in the cache group.

It should be noted that the cache controller judges the state of the VMS, and when it is judged that the VMS is in the second state, replaces the least recently used cache data in the cache group.

Wherein, when the VMS of the first virtual machine is in the first state, the cache controller obtains the cache miss rate of the second virtual machine, and then determines whether the cache miss rate of the second virtual machine is greater than a miss rate threshold. If the cache failure rate of the second virtual machine is greater than the failure rate threshold, replace the least recently used cache data in the cache group; if the cache failure rate of the second virtual machine is not greater than the failure rate threshold, the cache controller determines the cache group The number of cache blocks belonging to the second virtual machine in .

Further, after determining the number of cache blocks belonging to the second virtual machine in the cache group, the cache controller determines whether the number of cache blocks of the second virtual machine is smaller than a threshold value. If the number of cache blocks of the second virtual machine is less than the number threshold, replace the least recently used cache data in the cache group; if the number of cache blocks of the second virtual machine is not less than the number threshold, then replace the cache data belonging to the second virtual machine The least recently used cached data.

It should be noted that the failure rate threshold of the second virtual machine and the threshold of the number of cache blocks belonging to the second virtual machine in the cache group are all set in advance, and the values can be determined according to the specific operating conditions of the virtual machine and the needs of users. Setting; the embodiment of the present invention does not limit the implementation of the cache controller to obtain the cache failure rate of the second virtual machine, and can be any implementation known to those skilled in the art, for example, through the second virtual machine's MPKI (Misses Per1KInstructions, the miss rate per 1000 instructions) to calculate.

Compared with the problem in the prior art that the performance of the virtual machine cannot be well improved and the performance of the virtual machine is low when the virtual machine is started and copied, in the embodiment of the present invention, when the virtual machine is started and copied , dividing the last-level shared cache between the second virtual machine and the first virtual machine, reducing the capacity of the cache occupied by the second virtual machine. Since the second virtual machine is not sensitive to the cache capacity, even if the cache failure rate is high, its performance is not greatly affected, so by reducing the cache capacity occupied by the second virtual machine, increase the cache occupied by the first virtual machine Capacity, to reduce the competition among the virtual machines for the cache, and solve the problem in the prior art that the competition between the virtual machines for the cache reduces the performance of the virtual machine. By solving the above technical problems, the performance of virtual machine startup or replication can be improved.

Yet another embodiment of the present invention provides a method for dividing a last-level shared cache, which is used in a virtualization environment. In the virtual environment, on the virtual platform of the present embodiment, the maximum number of virtual machines that can run is N=8, and Domain0 is a privileged virtual machine running on the virtual machine monitor hypervisor, which can control physical I/O resources, and At the same time, it interacts with the guest virtual machine. The guest virtual machine needs to be started through Domain0. The cache block size in the virtual machine shared cache is 8 bytes, 4-way group associative, and a total of 16 groups. As shown in Figure 2, the method includes:

201. Add a virtual machine VM register.

Wherein, the processor includes multiple processor cores, and a virtual machine VM register is added to each processor core, and the data structure in the VM register includes: a virtual machine management operation state VMS and a virtual machine identifier VMID.

It should be noted that VMS has two states of 0 and 1, which are used to distinguish the type of operation performed by the guest virtual machine at this time: if VMS=0, it means that the guest virtual machine is not performing a startup operation or a copy operation at this time; if VMS=1 , indicating that the guest virtual machine is performing a startup operation or a copy operation at this time. The VMID is used to identify different virtual machines, and the number of digits of the VMID is determined by the maximum number of runnable virtual machines on the platform.

For example, if the maximum number of virtual machines that can run on the virtual platform is N, the number of bits of the VMID is log2(N). In the embodiment of the present invention, the maximum number of virtual machines that can run is N=8, then the number of digits of the VMID is log2(N)=3, and the VMID of Domain0 is 000. This embodiment does not limit the value of N, and the method of this embodiment is also applicable when N is other values.

202. Extend the flag bit of the cache address.

Wherein, the VMID is added to the tag bit of the cache address. Here, the VMID is used to mark which virtual machine the cache block in the cache address belongs to. When the cache block data is replaced, the VMID value in the register at this time can be written into the VMID in the cache address.

203. The virtual machine monitor sets a failure rate threshold of Domain0 and a threshold of the number of cache blocks belonging to Domain0 in the cache group.

Among them, the failure rate threshold of Domain0 is obtained according to the experimental results. When a guest virtual machine performs startup or replication operations, although Domain0 can maintain good performance when the failure rate is high, when the failure rate exceeds a certain value, the performance of Domain0 may degrade. In order to ensure that Domain0 can maintain good performance, users can set the failure rate threshold by themselves according to the experimental results. The threshold for the number of cache blocks belonging to Domain0 in the cache group is also set by the user, and the maximum value cannot exceed the value of the cache blocks in the cache group.

For example, in the embodiment of the present invention, if the number of cache blocks in each group is 4, then the maximum value of the cache block number threshold is 4.

It should be noted that the cache group mentioned in this step is the cache group determined according to the memory access address when the virtual machine performs a memory access operation.

204. The virtual machine monitor sets the state of the virtual machine management operation state VMS.

Wherein, the virtual machine monitor sets the state of the VMS according to the operation content of the virtual machine. When the virtual machine monitor responds to the guest virtual machine performing a startup operation or a copy operation, set VMS=1; when the virtual machine monitor responds not to the guest virtual machine performing a startup operation or a copy operation, set VMS=0.

205. The cache controller determines whether all the cached data in the cache group accessed by the client virtual machine misses, and if none of them hits, execute step 206; if not, end the process.

Among them, after the virtual machine monitor sets the state of the VMS, the virtual machine accesses the cache according to the access address. If the virtual machine accesses the cache, the cache data of the cache group is not hit. At this time, it is necessary to access the memory to find the corresponding Then replace the data with the data in the cache, and the cache controller judges the state of the VMS at this time; if there is a cache data hit in the cache group when the virtual machine accesses the cache, data replacement is not required, and the process ends.

It should be noted that, in the embodiment of the present invention, the memory access address is 16 bits, so the structure of the memory access address is as shown in Figure 3, wherein the tag bit Tag is the first nine bits of the memory access address, when the tag bit in the cache block When it matches the flag bit in the memory access address and the effective position in the cache block is 1, it means a cache hit; the cache group bit Set is the tenth to thirteenth bits of the memory access address, which is used to determine the cache group to be accessed , the offset bit Offset is the last three bits of the memory access address, which is used to determine the byte to be accessed in the cache block.

For example, Xen responds to a virtual machine command with VMID=101, and the virtual machine monitor sets the state of the VMS. Then, the virtual machine with VMID=101 accesses the cache, and the memory access address of the virtual machine this time is set to 0001000111010011. Then, the group number Set=1010 to access the cache can be found from the memory access address, as shown in Figure 4 . The extended tag bits of the four cache blocks in the cache group 1010 are: 000000001111, 010000000011, 000000000001, and 011000100001, where the first three digits of the tag bits are the extended VMID, indicating that the cache block belongs to The last nine bits are the Tag bits of the address of each cache block. The Tag bit of the memory access address of the virtual machine this time is 000100011. It can be seen that none of the four cache blocks are hit. At this time, it is necessary to access the memory to find the corresponding data, and then replace it with the data in the cache, go to step 206.

206. The cache controller determines whether the VMS status is 1, and if the VMS status is 1, execute step 207; otherwise, use the traditional LRU replacement policy.

It should be noted that the status of the VMS has been set in step 204, and it is judged at this time, so as to determine the replacement strategy to be used when replacing data.

Wherein, the replacement strategy includes a traditional LRU replacement strategy and an improved LRU replacement strategy in the embodiment of the present invention. When the VMS status is 1, use the improved LRU replacement strategy in the embodiment of the present invention, that is, perform steps 207 to 211; when the VMS status is 0, use the traditional LRU replacement strategy, that is, replace the least recently used cache in the cache group data.

For example, in the embodiment of the present invention, when the virtual machine with VMID=101 accesses the cache, none of the cache blocks of the cache group to be accessed by the client virtual machine is hit. If the VMS status is 1 at this time, step 207 is performed; if the VMS status is 0 at this time, a traditional LRU replacement strategy is executed for data replacement. As shown in Figure 4, suppose that in the 1010th cache group at this time, the 4th cache block is the least recently used cache block, and its VMID=011 shows that this cache block belongs to the virtual machine with VMID=011, and the judgment result If the VMS state is 0, select the fourth cache block and replace the data in the cache block with the data accessed from the memory by the virtual machine with VMID=101. The result is shown in Figure 5.

207. The cache controller acquires the cache miss rate of Domain0 at this time.

Wherein, the cache failure rate is equal to the number of cache misses divided by the number of cache accesses. The embodiment of the present invention does not limit the implementation of the cache controller to obtain the cache failure rate of Domain0, and may be any implementation known to those skilled in the art. For example, MPKI via Domain0.

208. The cache controller determines whether the cache failure rate is greater than the failure rate threshold, and if not greater than the failure rate threshold, executes step 209; otherwise, uses the traditional LRU replacement strategy.

Wherein, if the cache failure rate of Domain0 is greater than the failure rate threshold at this time, then use the traditional LRU replacement strategy, that is, replace the least recently used cache data in the cache group; if the cache failure rate of Domain0 is not greater than the failure rate threshold, then execute Step 209.

209. The cache controller determines the number of cache blocks belonging to Domain0 in the cache group.

Wherein, the cache controller can obtain which virtual machine each cache block in the cache group belongs to at this time according to the VMID extended in the cache tag bit, so as to determine the number of cache blocks belonging to Domain0.

For example, in the embodiment of the present invention, as shown in FIG. 4, in cache group 1010, according to the VMID of each cache block, it can be obtained that: the first and third cache blocks belong to Domain0, and the second cache block belongs to VMID=010 For a virtual machine, if the fourth cache block belongs to the virtual machine with VMID=011, it can be determined that the number of cache blocks belonging to Domain0 is 2.

210. The cache controller determines whether the number of cache blocks is less than the threshold number, and if it is smaller than the threshold number, execute step 211; otherwise, use the traditional LRU replacement policy.

Wherein, if the number of cache blocks belonging to Domain0 in the cache group is less than the number threshold, the traditional LRU replacement strategy is used, that is, the least recently used cache data in the cache group is replaced; if the number of cache blocks belonging to Domain0 in the cache group is not less than the number threshold, go to step 211.

211. The cache controller selects the least recently used cache block belonging to Domain0 in the cache group for replacement.

For example, in the embodiment of the present invention, as shown in FIG. 4 , at this time, in the 1010th group, the first and third cache blocks belong to Domain0, assuming that the third cache block is the least recently used cache belonging to Domain0 block, select the third cache block, and replace the data in the cache block with the data accessed from the memory by the virtual machine with VMID=101. The result is shown in Figure 6.

Compared with the problem in the prior art that the performance of the virtual machine cannot be well improved and the performance of the virtual machine is low when the virtual machine is started and copied, in the embodiment of the present invention, when the virtual machine is started and copied , divide the last-level shared cache between Domain0 and the guest virtual machine, and reduce the capacity of the cache occupied by Domain0. Since Domain0 is not sensitive to the cache capacity, even if the cache failure rate is high, its performance is not greatly affected. Therefore, by reducing the cache capacity occupied by Domain0 and increasing the cache capacity occupied by guest virtual machines, the virtual machine The competition for the cache among the virtual machines solves the problem in the prior art that the performance of the virtual machines decreases due to the competition for the cache among the virtual machines. By solving the above technical problems, the performance of virtual machine startup or replication can be improved.

Another embodiment of the present invention provides a device 30 for dividing a cache. As shown in FIG. 7, the device 30 includes:

A judging unit 31, configured to judge the virtual machine management operation status VMS of the first virtual machine when none of the cached data in the cache group accessed by the first virtual machine hits;

The replacement unit 32 is configured to replace the least recently used cache data belonging to the second virtual machine in the cache group when the VMS of the first virtual machine is in the first state.

Further, as shown in FIG. 8, the device 30 may also include:

The adding unit 33 is configured to add a virtual machine VM register, the data structure of the VM register includes: the VMS and a virtual machine identifier VMID.

Further, as shown in FIG. 8, the device 30 may also include:

The extension unit 34 is configured to extend the flag bit of the cache address, and add the VMID added by the adding unit 33 to the flag bit, and the cache data in the cache address belongs to the virtual machine corresponding to the VMID.

Further, as shown in FIG. 8, the device 30 may also include:

A setting unit 35, configured to set the VMS of the first virtual machine to a first state when the first virtual machine performs a startup operation or a copy operation; when the first virtual machine does not perform a startup operation or a copy operation , setting the VMS of the first virtual machine to a second state.

Further, the replacement unit 32 is also used for:

When the VMS of the first virtual machine is in the second state, replace the least recently used cache data in the cache group.

Further, the judging unit 31 is further configured to judge whether the cache failure rate of the second virtual machine is greater than a failure rate threshold;

The replacement unit 32 is further configured to replace the least recently used cache data in the cache group when the cache failure rate of the second virtual machine is greater than the failure rate threshold; as shown in FIG. 8 , the device 30 Can also include:

The determining unit 36 is configured to determine the number of cache blocks belonging to the second virtual machine in the cache group when the cache failure rate of the second virtual machine is not greater than the failure rate threshold.

Further, the judging unit 31 is further configured to judge whether the number of cache blocks of the second virtual machine is smaller than the number threshold determined by the determining unit 36;

The replacement unit 32 is further configured to replace the least recently used cache data in the cache group when the number of cache blocks of the second virtual machine is less than the number threshold determined by the determination unit 36;

The replacing unit 32 is further configured to replace the least recently used cache block belonging to the second virtual machine in the cache group when the number of cache blocks of the second virtual machine is not less than the number threshold determined by the determining unit 36. Cache data.

Compared with the problem in the prior art that the performance of the virtual machine cannot be well improved and the performance of the virtual machine is low when the virtual machine is started and copied, in the embodiment of the present invention, when the virtual machine is started and copied , dividing the last-level shared cache between the second virtual machine and the first virtual machine, reducing the capacity of the cache occupied by the second virtual machine. Since the second virtual machine is not sensitive to the cache capacity, even if the cache failure rate is high, its performance is not greatly affected, so by reducing the cache capacity occupied by the second virtual machine, increase the cache capacity occupied by the first virtual machine , to reduce the competition among the virtual machines for the cache, and solve the problem in the prior art that the performance of the virtual machine is degraded due to the competition for the cache between the virtual machines. By solving the above technical problems, the performance of virtual machine startup or replication can be improved.

Yet another embodiment of the present invention provides a device 40 for dividing a cache. As shown in FIG. 9 , the device 40 includes:

The processor 41 is configured to determine the virtual machine management operation state VMS of the first virtual machine when none of the cache data in the cache group accessed by the first virtual machine hits; and, when the first virtual machine's When the VMS is in the first state, replace the least recently used cache data belonging to the second virtual machine in the cache group.

Further, the processor 41 is further configured to add a virtual machine VM register, and the data structure of the VM register includes: the VMS and a virtual machine identifier VMID.

Further, the processor 41 is further configured to extend the flag bit of the cache address, and add the VMID to the flag bit, and the cache data in the cache address belongs to the virtual machine corresponding to the VMID.

Further, the processor 41 is further configured to set the VMS of the first virtual machine to the first state when the first virtual machine executes a startup operation or a copy operation; and, when the first virtual machine Set the VMS of the first virtual machine to the second state when the machine is not performing the startup operation or the copy operation.

Further, the processor 41 is further configured to replace the least recently used cache data in the cache group when the VMS of the first virtual machine is in the second state.

Optionally, the processor 41 is further configured to determine whether the cache miss rate of the second virtual machine is greater than a miss rate threshold; and, when the cache miss rate of the second virtual machine is greater than the miss rate threshold , replace the least recently used cache data in the cache group; and, when the cache invalidation rate of the second virtual machine is not greater than the invalidation rate threshold, determine that the cache data in the cache group belongs to the second The number of cache blocks for the virtual machine.

Optionally, the processor 41 is further configured to determine whether the number of cache blocks of the second virtual machine is less than a number threshold; and, if the number of cache blocks of the second virtual machine is less than the number threshold, then replacing the least recently used cache data in the cache group; and, if the number of cache blocks of the second virtual machine is not less than the number threshold, replacing the least recently used cache data belonging to the second virtual machine in the cache group The cached data to use.

Compared with the problem in the prior art that the performance of the virtual machine cannot be well improved and the performance of the virtual machine is low when the virtual machine is started and copied, in the embodiment of the present invention, when the virtual machine is started and copied , dividing the last-level shared cache between the second virtual machine and the first virtual machine, reducing the capacity of the cache occupied by the second virtual machine. Since the second virtual machine is not sensitive to the cache capacity, even if the cache failure rate is high, its performance is not greatly affected, so by reducing the cache capacity occupied by the second virtual machine, increase the cache capacity occupied by the first virtual machine , to reduce the competition among the virtual machines for the cache, and solve the problem in the prior art that the performance of the virtual machine is degraded due to the competition for the cache between the virtual machines. By solving the above technical problems, the performance of virtual machine startup or replication can be improved.

An apparatus for accessing a divided cache provided by an embodiment of the present invention can implement the method embodiment provided above. For specific function implementation, please refer to the description in the method embodiment, and details will not be repeated here. The method and device for dividing a cache provided by the embodiments of the present invention may be applicable to a virtualization environment, but is not limited thereto.

It should be noted that the cache mentioned in the embodiment of the present invention is applicable to the last-level shared cache, but is not limited thereto.

Each embodiment in this specification is described in a progressive manner, the same and similar parts of each embodiment can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for relevant parts, please refer to part of the description of the method embodiment.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the programs can be stored in a computer-readable storage medium. During execution, it may include the processes of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM), and the like.

The above is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Anyone skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present invention. All should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (14)

1. A method for dividing cache, characterized in that, comprising: When none of the cache data in the cache group accessed by the first virtual machine hits, determine the virtual machine management operation state VMS of the first virtual machine; When the VMS of the first virtual machine is in the first state, replace the least recently used cache data belonging to the second virtual machine in the cache group, wherein the first state is used to represent the first virtual machine The state in which the start operation or the copy operation is performed, the second virtual machine is used to control the physical input and output I/O resources of the virtual environment, interact with the first virtual machine, and start the first virtual machine. 2. The method according to claim 1, wherein before the judging the virtual machine management operation state VMS of the first virtual machine, the method further comprises: A virtual machine VM register is added, and the data structure of the VM register includes: the VMS and a virtual machine identifier VMID. 3. The method according to claim 2, wherein, after adding the virtual machine VM register, the method further comprises: The flag bit of the cache address is extended, and the VMID is added to the flag bit, and the cache data in the cache address belongs to the virtual machine corresponding to the VMID. 4. The method according to claim 1, wherein before the judging of the virtual machine management operation state VMS of the first virtual machine, the method further comprises: When the first virtual machine performs a startup operation or a copy operation, set the VMS of the first virtual machine to a first state; when the first virtual machine does not perform a startup operation or a copy operation, set the first The VMS of the virtual machine is in the second state. 5. according to the described method of claim 1 or 4, it is characterized in that, described method also comprises: When the VMS of the first virtual machine is in the second state, replace the least recently used cache data in the cache group. 6. The method according to claim 1, wherein when the VMS of the first virtual machine is in the first state, before replacing the least recently used cache data belonging to the second virtual machine in the cache group , the method also includes: judging whether the cache failure rate of the second virtual machine is greater than a failure rate threshold; If the cache failure rate of the second virtual machine is greater than the failure rate threshold, replace the least recently used cache data in the cache group; If the cache failure rate of the second virtual machine is not greater than the failure rate threshold, then determine the number of cache blocks belonging to the second virtual machine in the cache group. 7. The method according to claim 6, wherein after the determination of the number of cache blocks belonging to the second virtual machine in the cache group, the method further comprises: judging whether the number of cache blocks of the second virtual machine is less than a number threshold; If the number of cache blocks of the second virtual machine is less than the number threshold, replace the least recently used cache data in the cache group; If the number of cache blocks of the second virtual machine is not less than the number threshold, replace the least recently used cache data belonging to the second virtual machine in the cache group. 8. A device for dividing cache memory, comprising: A judging unit, configured to judge the virtual machine management operation state VMS of the first virtual machine when none of the cache data in the cache group accessed by the first virtual machine hits; A replacement unit, configured to replace the least recently used cache data belonging to the second virtual machine in the cache group when the VMS of the first virtual machine is in a first state, wherein the first state is used to represent the The state in which the first virtual machine executes the startup operation or the copy operation, the second virtual machine is used to control the physical input and output I/O resources of the virtual environment, interact with the first virtual machine, and start the First virtual machine. 9. The device according to claim 8, further comprising: The adding unit is configured to add a virtual machine VM register, the data structure of the VM register includes: the VMS and a virtual machine identifier VMID. 10. The device according to claim 9, further comprising: The extension unit is configured to extend the flag bit of the cache address, and add the VMID added by the adding unit to the flag bit, and the cache data in the cache address belongs to the virtual machine corresponding to the VMID. 11. The device according to claim 8, further comprising: A setting unit, configured to set the VMS of the first virtual machine to a first state when the first virtual machine performs a startup operation or a copy operation; and, when the first virtual machine does not perform a startup operation or During the copy operation, set the VMS of the first virtual machine to the second state. 12. The device according to claim 8 or 11, wherein the replacement unit is also used for: When the VMS of the first virtual machine is in the second state, replace the least recently used cache data in the cache group. 13. The device according to claim 8, wherein the judging unit is further configured to judge whether the cache failure rate of the second virtual machine is greater than a failure rate threshold; The replacement unit is further configured to replace the least recently used cache data in the cache group when the cache failure rate of the second virtual machine is greater than the failure rate threshold; the device further includes: A determining unit, configured to determine the number of cache blocks belonging to the second virtual machine in the cache group when the cache failure rate of the second virtual machine is not greater than the failure rate threshold. 14. The device according to claim 13, wherein the judging unit is further configured to judge whether the number of cache blocks of the second virtual machine is smaller than a number threshold determined by the determining unit; The replacement unit is further configured to replace the least recently used cache data in the cache group when the number of cache blocks of the second virtual machine is smaller than the number threshold determined by the determination unit; The replacement unit is further configured to replace the least recently used cache data belonging to the second virtual machine in the cache group when the number of cache blocks of the second virtual machine is not less than the number threshold determined by the determination unit .