JP2003296157A - Data storage device, data processor, data processing method, and data processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data storage device, data processor, data processing method and a data processing program capable of reducing the number of retrieval steps to speed up the retrieval processing. <P>SOLUTION: This data storage device has a 2-n-ary tree data structure and stores a plurality of sets of range data. In this device, the range data are stored in a terminal storage area obtained by tracing branch points having, as an index, each n-bit value obtained by partitioning the start address of the range data from the top by n-bits in order from the top of the 2-n-ary tree. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing device having a search function or registration function for registered predetermined range data, a data processing method therefor, a data processing program, or the predetermined range data. Data storage device.

[0002]

2. Description of the Related Art Currently, there are various types of communication hardware for performing communication between computers, but in recent years, the processing by the CPU of the computer itself has been taken over to some extent to reduce the load on the CPU of the computer. The capable communication hardware has been widely used. An example is the InfiniBand system. InfiniBand specifies the address range on the computer's memory as the source or destination of data,
Band hardware has the ability to directly access that range of data. In order to use this function, "Infini
The attribute "Band hardware can access" must be attached to the appropriate address range in memory before communication.

An outline of processing of the system using such InfiniBand will be described. FIG. 13 shows an outline of functions in the system. In the figure, the memory 10 includes a program area 20, a storage area 30 of an accessible address range, and an area 40 storing transmission data. Further, details of the transmission operation are shown in (1) to (4) of the figure.

In the storage area indicated by the dotted line frame (1),
Data to be transmitted (transmission data 40) is stored, and an attribute that allows InfiniBand hardware to access is added to the stored address range. (2) The address range assigned in (1) is stored in the address range storage area 30. In (3), when a program running on this system communicates, the source or destination address range specified by this program must be included in the address range stored in (2). To confirm. (4) The specified source or destination address range is
If it is confirmed in (3) that the address range is attributed, the InfiniBand hardware is made to access the memory in this address range and communication is executed.

[0005]

However, in the search of (3) above, conventionally, a method has been adopted in which a set of registered range data is sequentially examined one by one.
Therefore, the number of range data to be checked is ½ of the total number of registrations on average, and all range data registered in the worst case must be checked. Therefore, when the number of registered range data increases, the search time also increases proportionately.

The present invention has been made in order to solve such a problem, and is capable of reducing the number of search steps and the search time, a data storage device, a data processing device, a data processing method, and a data processing device. The purpose is to provide a processing program.

[0007]

In order to solve the above problems, the present invention is a data storage device having a data structure of 2 ^ n-branch type and storing a plurality of first data sets. The first data is data indicating a predetermined address range and its attribute, and the first data is 2 ^ n.
A second tree associated with the first data in the branch tree
N obtained by separating the data of n from the upper bit by n bits
There is provided a data storage device characterized in that a branch point whose bit value is used as an index is stored in a storage area at an end obtained by sequentially tracing from a higher order of a 2 ^ n branch tree.

With such a configuration, when the first data is retrieved from the data storage device, in the present invention, the number of processings to follow the link and the comparison processing of the address range data are greatly reduced. Therefore, the search processing becomes faster and the search time is shortened as compared with the conventional data structure.

Further, the second data may be represented by a binary number.

With this configuration, when the first data is retrieved from the data storage device, the second data is used to facilitate the first data in the 2 ^ n-branch data structure. You can search for.

Further, the second data may be a start address of the predetermined address range. Further, the attribute may be presence or absence of an access right to the predetermined address range.

According to such a configuration, for example, in a system that performs communication by referring to an arbitrary attribute (access right) of an address range in the memory, the accessible address range is registered in advance. , The address range that can be referred to can be searched quickly, and the access right can be easily confirmed.

Further, the first data is stored at the end of the branch point corresponding to all the addresses included in the address range, in addition to the end of the branch point corresponding to the start address. You can also do it.

With this configuration, for example, when searching the first data, any of the areas corresponding to all the addresses in the address range can be searched in the 2 ^ n-tree type data structure. If there is a hit, the registration status of the address range can be confirmed, so that the search processing becomes efficient.

Further, when there is only one terminal having a storage area for storing the first data below a certain branch point, the certain branch point in the 2 ^ n-ary tree is It can also be characterized by being omitted.

According to such a configuration, for example, when the first data is searched while branching from the upper branch in the 2 ^ n-branch data structure, the number of branches is reduced, so the search processing is performed. Will be more efficient. Also, only the omitted branch points,
The storage capacity of the data storage device can be reduced.

Further, an index number of the omitted branch point may be stored in the branch point immediately below the omitted branch point.

According to such a configuration, even if the branch point is omitted, the omitted branch point can be specified, so that the route to the end storing the first data is clear in the data storage device. Becomes

Further, it is possible that the ends storing a plurality of the first data are connected to each other by a bidirectional link. Furthermore, the bidirectional link is
The first data at the end sorted by the start address of a predetermined address range stored in the storage area at the end is D
1, D2. ． ． When Dk, D1 and D2, D2 and D
3. ． ． , D (k−1) and Dk having the first data are connected to each other by a bidirectional link.

With such a configuration, for example, when searching for the first data, if the end that stores the first data is searched, the links before and after the end can be easily followed. Since it is possible to confirm the end where the other first data is stored, the search processing is dramatically improved in efficiency.
In addition, this makes it possible to reliably perform the search without storing the first data in the end area corresponding to all the addresses included in the address range within the 2 ^ n tree structure. The amount of data stored in the data storage device can be dramatically reduced.

Further, it has a data structure of 2 ^ n-branch type,
A data storage device for storing a set of a plurality of first data, wherein the first data is data indicating a predetermined address range and its attribute, and the branch of the 2 ^ n-branch of the m-th layer. In the storage area of the points, in the area of the index equal to the value of the m-th n-bit from the high-order of each n-bit obtained by dividing the second data associated with the first data by the n-bit from the most significant Stores the address of the branch point of the m + αth layer, which is directly below the mth layer and has an index equal to the value of the m + αth n-bit, and if the m + αth layer is the lowest layer, It is also possible to provide a data storage device characterized in that the first data is stored in an area of the lower storage area having an index equal to the value of the lower n bits of the second data. That.

Further, the second data may be represented by a binary number. Furthermore, the second data may be a start address of the predetermined address range.

Further, the attribute may be presence or absence of an access right to the predetermined address range.

Further, α is 1, and the first data is stored in the lowest layer storage area corresponding to all the addresses included in the address range in addition to the lowest layer storage area corresponding to the start address. It can also be characterized by being stored in.

Further, when there is only one index area in the lowest layer in which the first data is stored below the branch point of the m-th layer, the α is 2. You can also Further, in the storage area of the branch point of the (m + 1) th layer, m + 1 from the top of the second data.
The index area equal to the n-th bit value may be a non-storage area. Furthermore, in the storage area of the branch point of the (m + 2) th layer, the second
The data stored in the area of the index equal to the (m + 2) -th n-bit value from the highest-order data includes the (m + 1-th) n-bit value from the highest-order data of the second data. You can also

Further, in the storage area of the lowermost layer in which the plurality of first data are stored, mutual addresses are further stored so that the storage areas of the lowermost layer are connected to each other by a bidirectional link. It can also be characterized by being. Further, the bidirectional link stores the first data at the end sorted by the start address of the predetermined address range stored in the storage area at the end at D1, D2. ． ． Assuming that Dk is D1, D2 and D2 are stored in the lowermost storage area.
3. ． ． , D (k−1) and Dk of the first data may be stored so that their respective ends are linked to each other.

A data processing device for performing a predetermined process on the above-mentioned data storage device in order for a predetermined device to access predetermined data, wherein the first data is stored in a predetermined storage area of the data storage device. It is possible to provide a data processing device capable of storing the data of 1) so that a predetermined address within the address range of the first data can be referred to as a reference key.

With such a configuration, range data can be registered easily and retrievably in the data storage device having the data structure of 2 ^ n-ary tree.

The present invention is also a data processing device for performing a predetermined process on the above-mentioned data storage device in order for a predetermined device to access predetermined data, wherein the data indicates an address range of the predetermined data. And a processing unit that performs a predetermined process based on a search result of the search unit, and a data processing device that searches the third data in the data storage device. provide.

With such a configuration, it is possible to execute an appropriate process for the predetermined device to access the predetermined data based on the search result of the address range with respect to the data storage device. In addition, in the embodiment of the present invention, the search unit is configured by the search unit, and the processing unit is configured by the search result acquisition unit and the registration unit.

Further, the searching means obtains the fourth data related to the third data, the first obtaining means, and the fourth data obtained by the first obtaining means from upper bits to n bits. A second acquisition means for acquiring sooner, a search means for searching for a predetermined end storing the first data based on each value of n bits acquired by the second acquisition means, and A third acquisition means for acquiring the first data stored in the predetermined storage area based on the search result, and a comparison for comparing the data acquired by the third acquisition means with the predetermined data Means for determining whether or not the third data is stored in the data storage device based on the comparison result by the comparison means.

With this configuration, it is possible to perform a search process suitable for the data structure of the data storage device,
As a result, the number of processings required to follow links for the number of search processings and the comparison processing of address range data are dramatically reduced. Therefore, the search process is speeded up and the search time is shortened as compared with the conventional case.

FIG. 1 shows a retrieval process using a conventional technique,
It is the figure which showed the difference of the search process in this invention. The principle of the present invention will be briefly described with reference to this drawing. The left side of FIG. 1 is a diagram showing a search process by a conventional method, and the right side is a diagram showing a search process according to the present invention. For the address range showing the storage area of the transmission data 40 designated by the program 20 on the memory 10 of FIG.
0).

When the search keys (300, 400) are searched, as is apparent from the figure, according to the conventional method, the registered address range data is stored in the address range storage area 30 on the memory 10 of FIG. The areas are stored in order from the start, and they are sequentially compared with the search key one by one. The comparison process is sequentially performed until registration data that matches the search key is found.

According to the method of the present invention, the storage area 30 of the address range has a data structure of a 2 ^ n-branch tree, and the search processing determines the branch point of the 2 ^ n-branch tree from the upper hierarchy. The registration data is discovered by following the sequence. In addition,
In the embodiment of the present invention, the address range (300, 40
0) start data 300 represented by a binary number is divided into 2 bits from the high-order bit and divided into 2 bits.
The branch point having the bit as an index is traced, and the data is stored in the storage area of the traced destination.

In the embodiment of the present invention, the first
The acquisition unit is composed of the start address acquisition unit, the second acquisition unit is composed of the start address acquisition unit and the search unit, the search unit and the third acquisition unit are composed of the first search unit, and the comparison unit is composed of the comparison unit. To be done.

Further, the fourth data may be a start address of the address range designated for access.

According to such a configuration, for example, in a system for performing communication by giving an access right to an address range in the memory, an address range in the memory in which predetermined data to be transmitted is stored, that is, access By searching the address range registered in advance in the data storage device for the designated address range using the start address as a key, the referenceable address range can be quickly searched and the access right can be easily confirmed. .

Further, the first acquisition means has a conversion means for converting the fourth data into a binary number when the fourth data is not a binary number, and acquires the data converted by the conversion means. You can also

According to such a configuration, whatever value the fourth data has, it can be converted into a binary number, and the data format is suitable for searching the data structure of the 2 ^ n-ary tree. You can In addition, according to the embodiment of the present invention, the conversion unit is configured by the start address acquisition unit.

Further, the predetermined terminal searched by the searching means is to follow the branch points having the n-bit value acquired by the second acquiring means as an index in order from the higher order of the 2 ^ n branch tree. It can also be characterized in that it is a terminal (M) obtained by.

According to this structure, by searching the end of the 2 ^ n-ary tree corresponding to the fourth data, the first data and the predetermined data stored at the end can be easily obtained. Can be compared.

Further, when the first data is not stored in the storage area of the terminal (M), the searching means selects 2 ^ of the terminals storing the first data. It is also possible to search for an end near the end (M) in the n-ary tree and set the end near the end as a predetermined end.

According to this structure, when the first data is not stored in the terminal (M) of the data storage device, the predetermined data can be compared with the data of the other terminal in the vicinity. , The search result can be output reliably.

Further, when searching the end near the end (M), the searching means selects the second data from the ends storing the first data in the data storage device. It is also characterized in that the terminal that stores the first data having a value smaller than the data of 4 is searched for the terminal closest to the terminal (M) in the 2 ^ n-ary tree. it can.

According to this structure, the end storing the first data whose second data has a smaller value than the fourth data can be set as the predetermined end. And the first data stored at the predetermined end are compared, and it is easy to determine whether or not the first data corresponding to the third data is stored in the data storage device. Can be confirmed.

Further, the comparing means compares the start address (ai) of the predetermined address range of the first data with the start address (a) of the access-designated address range of the third data. 1 comparison means, and 2nd comparison means for comparing the end address (bi) of the predetermined address range of the first data with the end address (b) of the access-designated address range of the third data, When the comparison result by the first comparing means is ai <= a and the comparison result by the second comparing means is b <= bi, the searching means stores the third data in the data storage. It can also be characterized in that it is determined that it is stored in the device.

According to this structure, it is possible to easily determine whether or not the third data is included in the address range of the first data stored at the searched end. It is possible to easily determine whether or not the third data is stored in the data storage device.

Further, when the searching means determines that the third data is not stored in the data storage device, the processing means stores the third data in the data storage device. It is also possible to add an attribute that permits access to the predetermined data by the predetermined device.

According to this structure, it is confirmed whether or not the third device has the access right necessary for the predetermined device to access the predetermined data, depending on whether or not the third data is stored in the data storage device. Therefore, the access processing to the predetermined data by the predetermined device becomes easy. Further, by storing the third data in the data storage device, it is possible to access predetermined data that the predetermined device does not have an access right to.

Further, the processing means is based on a first judging means for judging whether or not the first data is stored in the storage area of the terminal (M), and a judgment result by the first judging means. And a storage unit for storing the third data in the data storage device.

According to such a configuration, for example, in a system that performs communication by referring to the access right of the address range in the memory, regarding the address range in the memory in which predetermined data to be transmitted is stored, Since an address range without access permission can be stored in the area corresponding to the start address of the address range,
An access right can be easily given to the address range. In addition, in the embodiment of the present invention, the first determination unit is configured by the first determination unit, and the storage unit is configured by the range data storage unit.

Further, when the storage means determines by the first determination means that the first data is not stored in the storage area at the end (M), the storage means stores the first data in the storage area. Of the end where the data of 1 is stored, the end of the first data in which the second data has a smaller value than the fourth data, and further, in the end of the 2 ^ n-ary tree, First searching means for searching a terminal closest to the terminal (M), and storing the third data in the data storage device based on a search result of the first searching means. Is. Further, the storage means stores the end (M) when the end searching the first data storing the first data having the second data smaller than the fourth data is not found by the first searching means. The third data may be stored in the area.

With such a configuration, for example, in a system in which communication is performed by referring to the access right of the address range in the memory, the address range of the predetermined data to be transmitted does not have access permission. However, since the address range can be efficiently stored in an appropriate area, the system can access the data in the address range. In addition, in the embodiment of the present invention, the first searching unit is configured by the second searching unit.

Further, the storage means, when a certain end is searched by the first searching means, sets the first data stored in the storage area at the searched end as the first data and the first data. It may be characterized in that the third data is replaced with the merged fourth data.

According to such a configuration, for example, when the address range A is stored in the area in which the data of the address range B having an address value smaller than the start address of the address range A is stored, Since it becomes possible to replace the data in a format in which a portion not included in the address range B is added to the address B, the storage processing becomes efficient and the data storage area can be minimized.

Further, the storage means, when the first judging means judges that the first data is stored in the storage area of the end (M), the storage area of the end (M) The first data already stored in the third data
It can also be characterized in that it is replaced with the data of.

According to this structure, when the third data has a wider range than the first data stored at the end (M), only the data range of the third data is stored. Thus, access permission can be given to the predetermined data, which is efficient.

Further, in the data storage device, when the first data is stored only at the end of the branch point corresponding to the start address of the first data, the processing means further includes the data storage. Of the terminals storing the first data in the apparatus, the terminal storing the first data having the second data having a larger value than the second data stored at the terminal stored by the storage means is stored. Second searching means for searching, second judging means for judging the inclusive relation between the first data stored in the end storage area searched by the second searching means, and the data stored by the storing means And the first judgment stored in the end storage area searched by the second searching means by the second judging means.
When the data stored in the storage unit is determined to be included in the data stored in the storage unit, the deletion unit deletes the data stored in the end storage area searched by the second searching unit. Can also be

According to such a configuration, by storing the third data, the range of the data stored in the other end region following the stored end in the 2 ^ n-ary tree is
When it overlaps with a part of the third data, the duplication can be deleted and the storage capacity of the data can be suppressed. In addition, in the embodiment of the present invention, the second searching unit is configured by the third searching unit, the second determining unit is configured by the second determining unit, and the deleting unit is configured by the stored data deleting unit. .

Further, when the plurality of first data stored in the data storage device are connected to each other by a bidirectional link, the searching means included in the searching means or the first searching means included in the processing means. Alternatively, the second searching means may be characterized by performing a searching process using the bidirectional link.

With such a structure, the retrieval and storage processing is speeded up.

Further, when the plurality of first data stored in the data storage device are connected to each other by a bidirectional link, the processing means further includes the first data in the data storage device. It is also possible to provide a link changing unit for changing the bidirectional link when the stored end is added or deleted.

According to this structure, the bidirectional link can be replaced correctly even after new data is stored. In the embodiment of the present invention,
The link changing means is composed of a bidirectional link setting unit.

The present invention is also a data processing method for performing a predetermined process on the above-mentioned data storage device in order for a predetermined device to access predetermined data, the data indicating an address range of the predetermined data. A third step of retrieving the third data in the data storage device,
A second step of permitting the predetermined device to access the predetermined data when it is determined as a result of the search in the first step that the third data is stored in the data storage device; As a result of the search in the first step, if it is determined that the third data is not stored in the data storage device, the third data is stored in the data storage device and the predetermined device is stored. And a third step of permitting access to the predetermined data.

Further, in the first step, the first acquisition step of acquiring the fourth data related to the third data, and the fourth data acquired by the first acquisition step from the upper bit to n A second acquisition step of acquiring bit soot, a search step of searching for a predetermined end storing the first data based on each value of n bits acquired by the second acquisition step, and the search step And a third acquisition step of acquiring the first data stored in the storage area at the predetermined end, and the data acquired by the third acquisition step and the predetermined data based on the search result of A comparing step, and determining whether or not the third data is stored in the data storage device based on a comparison result of the comparing step. Rukoto can also.

Further, in the second step, the storage of the end obtained by sequentially tracing the branch points having the n-bit value acquired in the second acquisition step as an index from the upper order of the 2 ^ n branch tree. The first determination step of determining whether or not the first data is stored in the area, and the third determination step based on the determination result of the first determination step.
And a storage step of storing the data in the data storage device.

Further, the present invention is a data processing program recorded on a computer-readable medium and performing a predetermined process on the above-mentioned data storage device for the predetermined device to access predetermined data, A first step of searching in the data storage device for third data, which is data indicating an address range of the predetermined data, and as a result of the search in the first step, the third data is the data storage device. A second step of permitting access to the predetermined data of the predetermined device when it is determined that the third data is stored in the storage device. When it is determined that the third data is not stored in the device, the third data is stored in the data storage device and stored in the predetermined data of the predetermined device. To provide a data processing program, characterized by a third step of granting access.

Further, in the first step, the first acquisition step of acquiring the fourth data related to the third data, and the fourth data acquired by the first acquisition step from the upper bit to n A second acquisition step of acquiring bit soot, a search step of searching for a predetermined end storing the first data based on each value of n bits acquired by the second acquisition step, and the search step And a third acquisition step of acquiring the first data stored in the storage area at the predetermined end, and the data acquired by the third acquisition step and the predetermined data based on the search result of A comparing step, and determining whether or not the third data is stored in the data storage device based on a comparison result of the comparing step. Rukoto can also.

Further, in the second step, the storage of the end obtained by sequentially tracing the branch points having the n-bit value acquired in the second acquisition step as an index from the upper order of the 2 ^ n branch tree. The first determination step of determining whether or not the first data is stored in the area, and the third determination step based on the determination result of the first determination step.
And a storage step of storing the data in the data storage device.

In the present invention, the computer-readable medium is a portable storage medium such as a CD-ROM, a flexible disk, a DVD disk, a magneto-optical disk, an IC card, a database holding a computer program, or , Other computers and their databases, and also transmission media on lines.

[0072]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings. It should be noted that, in all the embodiments described below, the registration method or the search method of the address range used in the InfiniBand system described above, and the data structure of the storage area used in those methods will be described in detail. That is, in communication between computers, the address range in the computer's memory is designated as the source or destination of data, and InfiniBand hardware directly accesses the data in the designated range.
It is premised that the attribute "finiBand hardware can access" is added by registering the address range in a predetermined area on the memory before communication. Less than,
Data indicating the address range and the above attributes are called range data.

Embodiment 1. In the present embodiment, the storage area 3 of the address range in the memory 10 shown in FIG.
Taking the case where 0 has a quadtree data structure as an example, the data structure and the registration method will be described in detail. FIG. 2 is a diagram showing an example of a quadtree data structure. For example, in a computer whose address is represented by 32 bits,
It is assumed that a quad-tree type (n = 2 in 2 ^ n-branch tree) data structure is adopted.

Indexes 00, 01, and
There are branches 10 and 11, and the branch point of the next layer exists at the end of the branch. The index can be considered as an index assigned to the next branch point. The access-permitted address range is stored in the final storage area.

In the area having such a data structure, the address range 0x124800000 to 0x1248f00 is set.
When registering range data 100 of 0 (0x is a hexadecimal number display), the start address of the address range is 0x124800
Start address "00010010" in which 00 is represented in binary notation
01001000. ． ． ． Is divided into 2 bits each from the higher order (00, 01, 00, 10, 01, 00, 10,
00. ． ． ). These 2 bits are "0th branch (or branch point)" and "1st" in order from the root of the quadtree structure.
"0th""2nd". ． ． Then, the range data 100 is stored in the storage area at the end (range data 20
0).

Embodiment 2. FIG. 3 is a diagram showing an example in which a branch point is omitted in the quadtree data structure. As shown in FIG. 3, FIG. 3A shows a branch state when the start address is traced as it is as in FIG. Here, at the branch point X in the figure, it is assumed that there is no data at the ends of branches other than the second (index 10). Here, as shown in FIG. 3B, the branch point X is omitted, and the omitted branch number (index) (10) is recorded at the next branch point. By adopting such a data structure, the required memory area can be reduced, and the number of search steps can be reduced during the search, so the search time can be shortened.

Embodiment 3. FIG. FIG. 4 is a diagram showing an example in which a quadtree structure is arranged on a memory. FIG. 4A shows the storage status of the data represented by the quadtree structure. In FIG. 4B, the branch point A, the branch point B, the branch point C, and the range data 200 stored at the end shown in FIG. 4B are arranged on the actual memory 10a (address range storage area 30a). is there.

As shown in FIG. 4B, the data storage area 301 at the branch point A has 00, 01 and 1 respectively.
The address of the next branch point on the quadtree is stored for each index of 0 and 11. Where index 00
With reference to, since the address of the branch point B is stored, the branch point B on the quadtree can be traced. Fork point B
The address of the next branch point is also stored in the data storage area 302 of the same as the data storage area 300 of the branch point A, and the address of the branch point C can be referred to by referring to the index 01. . Finally, data regarding the end range data is stored in the end storage area 300. As illustrated, as the range data 200, a start address 0x124800000, an end address 0x1248f000, and attribute data of this address range (that is, attribute data that is accessible) are stored.

The present embodiment is an example of the storage method in the memory area, and the storage method is not particularly limited as long as the data is stored so as to have the data structure of 2 ^ n-branch tree.

Fourth Embodiment FIG. 5 is a diagram showing an example in which a bidirectional link is added to range data in a quadtree data structure. As illustrated, the end range data 200a, 200b, and 200c are connected by a bidirectional link indicated by an arrow. That is, range data 2
00a and range data 200b, range data 200b and range data 200c. ． ． In this way, the range data of adjacent ends are linked by a bidirectional link. When performing a search, the time required for the search process is significantly reduced by using such a bidirectional link.

Embodiment 5. As described above, the data structure is mainly described in detail in Embodiments 1 to 4, but in the present embodiment and the following Embodiments 6 to 11, the data structure having such a data structure is shown in FIG.
The data processing device that performs processing such as search and registration in the address range storage area 30 will be described in detail. Here, it is assumed that the specified address range (referred to as a specified address range) directly accessed by the InfiniBand hardware described above is searched and registered. Further, as in the first to fourth embodiments, it is assumed that a quadtree data structure is adopted for the address range storage area 30 in a computer whose address is represented by 32 bits.

FIG. 6 is an example of a block diagram showing a functional configuration of the data processing device. This functional block diagram is common to the fifth to tenth embodiments. As illustrated, the data processing device 90 includes a search unit 50 that searches the address range storage area 30, a search result acquisition unit 60 that acquires the search result of the search unit 50 and distributes the transmission destination of the search result, and a search. Upon receiving the search result from the result acquisition unit 60, the corresponding address range is set to the address range storage area 30.
And a registration unit 70 for registering with.

Further, the search unit 50 is composed of a start address acquisition unit 51, a first search unit 52, and a comparison unit 53. The start address acquisition unit 51 acquires the start address of the specified address range from the range. The first search unit 52
Based on the start address acquired by the start address acquisition unit 51, the end where the range data is stored is searched, and the storage data of the end is acquired. The comparison unit 53 compares the range data acquired by the first search unit 52 with the designated address range, and outputs the comparison result to the search result acquisition unit 60.

Further, the registration unit 70 includes a first judgment unit 71,
The second searching unit 72, the range data storing unit 73, the third searching unit 74, the second judging unit 75, and the stored data deleting unit 76.
And the bidirectional link setting unit 77. The first determination unit 71 acquires the search result from the search result acquisition unit 60,
It is determined whether or not some range data is stored in the storage area at the end searched using the start address acquired by the start address acquisition unit 51 as a key. When the first determination unit 71 determines that no range data is stored, the second search unit 72 searches for the end where the range data is stored, from the end existing near the end. . The range data storage unit 73 includes the start address acquisition unit 5
The range data of the designated address range is stored in the terminal storage area searched by using the start address acquired by 1 as a key or the terminal storage area searched by the second searching unit 72.

The third searching unit 74 searches for an end having stored range data whose start address has a value larger than the end stored by the range data storage 73. Second
The determination unit 75 compares the end range data stored and processed by the range data storage unit 73 with the end range data searched by the third search unit 74 to determine the inclusion relationship. The stored data deletion unit 76 deletes the range data stored at the end based on the judgment result by the second judgment unit 75. The bidirectional link setting unit 77 uses the range data storage unit 7
3 and the stored data deletion unit 76 adds or deletes the end where the range data is stored, the bidirectional link is replaced.

Based on the above configuration, an example of the search process will be described first. FIG. 7 is a diagram specifically showing the flow of the search process in the present embodiment. In the storage area at the end of the illustrated quadtree structure, range data 200d (0x
12482000-0x1248f000) is registered, and (0x1248 ****-0
The range data represented by (x *********) has not been registered. In this state, the specified address range 100d (0
(x12482000-0x12483000) is used as the search key.

(1) Start address acquisition unit 5 of search unit 50
The start address of the designated address range is acquired by 1. As shown, the starting address is 0x12482
000, the start address acquisition unit 51 sets this to 2
Convert to decimal. As a result, "00010010010
01000. ． ． Is obtained. The first search unit 52 traces a branch (branch point) having this start address as an index. Specifically, up to the end of 32 bits (depth 16),
0th, 1st, 0th, 2nd, 1st, 0th, 2
The 0th, 0th, 0th, 2nd, 0th, 0th, 0th, 0th, 0th, 0th branch (branch point) will be traced in order.

(2) Range data 2 at the end of the traced destination
00d (0x12482000 to 0x1248f00
0) is stored. The first search unit 52 acquires the range data 200d.

(3) The comparing section 53 compares the range data 200d acquired by the first searching section 52 with the designated address range (search key) 100d. Since 0x12482000 <= 0x12482000 for the start address, the end addresses may be compared later. In this case, 0x1248f000> 0x12483000, so
The stored range data includes the specified address range.

(4) Therefore, the comparison unit 53 outputs the range data 200d to the search result acquisition unit 60. Based on the data received from the comparison unit 53, the search result acquisition unit 60
Range data 200d as a notification that the specified address range has been confirmed in the communication hardware as a search hit
Is output.

Sixth Embodiment In FIG. 8, range data 200e (0x1247f000-
0x12484000) is registered, and (0x1248 ***** to 0x *********) at the other end.
It is the figure which showed the search process in case the range data represented by *) are not registered. Here, the designated address range (search key) 100e to be searched is (0x12482000 to 0x1248300) as in the fifth embodiment.
0).

(1) Since the process up to the acquisition of the start address is the same as that of the fifth embodiment, its explanation is omitted. First search unit 52
Follows the 0th, 1st, 0th, 2nd, 1st, 0th branch in order up to 12 bits (depth 6) according to the acquired start address.

(2) At the branch point D, the range data 200e (0) is added to the end at the end of the first (index 01) branch.
x1247f000 to 0x12484000) exists, but range data does not exist at the end of the second (index 10) branch. Therefore, instead of the second route that should be originally traced, the first branch is traced. At the branch points thereafter, the branch having the largest number among the branches having the range data is traced.

(3) Following the branch, the range data 200e
It is assumed that the end where only (0x1247f000-0x12484000) exists is reached. This range data 20
0e may include the search key because the starting address is smaller than the search key. Moreover, it is the range data closest to the search key. The first search unit 52 acquires the range data 200e.

(4) The comparison unit 53 compares the acquired range data 200e with the end address of the search key 100e. In this case, 0x12484000> 0x1248
Since it is 3000, this range data 200e has search key 1
00e is included.

(5) Therefore, the comparison unit 53 outputs the range data 200e to the search result acquisition unit 60. Subsequent processing is the same as that of the fifth embodiment, and therefore the description is omitted.

Seventh Embodiment In FIG. 9, range data 200f (0x12490000-
0x12491000) is registered, and (0x1248 ***** to 0x *********) at the other end.
It is the figure which showed the search process in case the range data represented by *) are not registered. Here, the designated address range (search key) 100f to be searched is (0x12482000 to 0x1248300) as in the fifth embodiment.
0).

(1) The process up to the acquisition of the start address is the same as that of the fifth embodiment, and therefore its explanation is omitted. First search unit 52
Follows the 0th, 1st, 0th, 2nd, 1st, 0th, 2nd branch in order up to 14 bits (depth 7) according to the acquired start address.

(2) At the branch point E, the range data 200f (0 is set at the end at the end of the first (index 01) branch.
However, there is no range data at the end of the 0th (index 00) branch. Therefore, the first branch is traced instead of the 0th branch which should be traced. At the branch points thereafter, the branch having the smallest number among the branches having the range data is traced.

(3) The range data 200f is traced along the branch.
It is assumed that the end where only (0x124900000-0x12491000) exists is reached. This range data 20
0f does not include the search key 100f because the start address is larger than the search key 100f. However,
The range data is closest to the search key 100f.

(4) The first searching unit 52 uses the bidirectional link which is set at the end in advance, and the range data 20
The range data 200g immediately before 0f is acquired. The comparison unit 53 uses the acquired range data 200g and the search key 10
Compare with 0f. If the range data 200g includes the search key, the comparison unit 53 determines the range data 200g.
The g is output to the search result acquisition unit 60. Subsequent processing is the same as that of the fifth embodiment, and therefore the description is omitted.

Eighth Embodiment In the present embodiment and the following ninth and tenth embodiments, the registration processing performed when the search is not hit by the search processing will be described in detail. In the present embodiment, as shown in FIG. 10, range data 200h (0x12482) is stored in the storage area of the quadtree structure.
000 to 0x12482800) is registered, and (0x1248 *** to 0x ****) at the other ends.
The registration process when the range data represented by (***) is not registered will be described in detail. Also, the specified address range (search key, range data) to be searched and registered 100h
Is (0x12482000-0x12483000)
And

It will be described with reference to FIG. (1) The description up to the acquisition of the start address is omitted because it is the same as in the fifth embodiment. The first searching unit 52 uses the acquired start address up to the end of 32 bits (depth 16), 0th, 1st, 0th, 2nd, 1st, 0th, 2nd, 0th, 0. The 2nd, 2nd, 0th, 0th, 0th, 0th, 0th, 0th branch is traced in order.

(2) Range data 2 at the end of the traced destination
00h (0x12482000-0x1248280
0) is stored. The first search unit 52 acquires the range data 200h. The procedure up to this point is the fifth embodiment.
(1) and (2).

(3) This range data 200h is compared with the end address of the designated address range 100h. In this case, since 0x12482800 <0x12483000, the range data 100h includes the existing range data 200h. The comparison unit 53 outputs the result to the search result acquisition unit 60. From the output result, it is determined that the search did not hit.

(4) The search result acquisition unit 60 outputs the search result to the first judgment unit 71 of the registration unit 70. First judging unit 7
1 is range data 20 based on the output search result.
It is determined that 0h is stored in the end storage area traced by the start address of the designated address range 100h. Based on this determination, the range data storage unit 73 replaces the stored range data 200h with the new range data 1
Replace with 00h. Replaced range data is 20
0i.

Next, FIG. 10B will be described. (5) Since the new range data 200i inherits the bidirectional link that the existing range data 200h had, the setting of the bidirectional link in this part does not change.

(6) Next, the third searching unit 74 follows the bidirectional link from the newly inserted range data 200i to obtain the next range data 200j. The second determination unit determines the new range data 200i and the acquired range data 2
Compare 00j.

(7) If the new range data 200i does not include the range data 200j, the registration process is terminated and the communication hardware is notified. If the new range data 200i includes the range data 200j, the stored data deletion unit 76 deletes the range data 200j from the quadtree structure area, and the bidirectional link setting unit 77 deletes the deleted range data 200j. Replace the two-way link. Below, the third
The search unit 74 sequentially acquires existing range data by using the bidirectional link, and the second determination unit 75 acquires the new range data 2
The deletion of the existing range data and the replacement of the link are repeated until it is determined that 00i does not include the existing range data.

[Embodiment 9] In this embodiment mode, FIG.
As shown in, the range data 2 is stored in the storage area of the quadtree structure.
00m (0x1247f000-0x1248100
0) is registered, and (0x124
The registration processing in the case where the range data represented by 8 ***** to 0x ********** is not registered will be described in detail. The specified address range (search key, range data) 100k to be searched and registered is (0x12482000).
0x12483000).

It will be described with reference to FIG. (1) The description up to the acquisition of the start address is omitted because it is the same as in the fifth embodiment. The first search unit 52 uses the acquired start address up to 12 bits (depth 6).
The 0th, 1st, 0th, 2nd, 1st, 0th branch is traced in order.

(2) At the branch point F, the range data 200 m (0x1247) is located under the first (index 01) branch.
f000 to 0x12481000) exists, but range data does not exist at the tip of the second (index 10) branch that is the branch to be traced using the acquired start address as a key. Similar to the sixth embodiment, the first searching unit 52 traces another branch to search and acquire other range data, but the search result acquisition unit 60 determines that the search hits based on the comparison result of the comparison unit 53. If not, the search result is output to the first determination unit 71 of the registration unit 70. In the present embodiment, further detailed description of the search process will be omitted.

(3) Based on the output search result, the first judging section 71 judges that nothing is stored in the end storage area traced by the start address of the designated address range 100k. It Based on this determination, the second searching unit 72 causes the existing range data 200m (0x1247f000 to 0x124810) immediately before the end to exist.
00) is searched. The search result may be directly acquired from the search result of the first search unit 52 of the search unit 50, or may be acquired by the same process as the first search unit 52.

As a result, the searched range data 200
Since it is found that m does not include any address included in the range data 100k, the range data storage unit 7
3 extends the second branch (index 10) and the following branches to create a branch to the end that follows the start address of the range data 100k, and the range data 100 at this end.
Store k as is. 200k stored range data
And

Next, FIG. 11B will be described. (4) Since the existing range data 200m, which is one before the new range data 200k, is already known by the processing of the first search unit 52 or the second search unit 72, the bidirectional link setting unit 77 determines that the existing range data The bidirectional link between 200m and the next existing range data (not shown) is disconnected, and new range data 200k and existing range data 200m
Change the links so that and are connected by a bidirectional link.
Similarly, the existing range data next to the unlinked new range data 200k and the new range data 200k are connected by a bidirectional link.

(5) Subsequently, the third searching section 74, the second judging section, the stored data deleting section, and the bidirectional link setting section 77.
By doing so, the bidirectional link is replaced while deleting unnecessary range data. This processing is the same as (6) and (7) of the eighth embodiment, and therefore the description is omitted.

Embodiment 10. FIG. In the present embodiment, FIG.
As shown in FIG. 2, range data 200p (0x12490000 to 0x1249100) is stored in the storage area of the quadtree structure.
0) is registered, and (0x124
The search and registration processing in the case where the range data represented by 8 ***** to 0x *********** is not registered will be described in detail. The specified address range (search key, range data) 100n to be searched and registered is (0x1248).
2000-0x12483000).

It will be described with reference to FIG. (1) The description up to the acquisition of the start address is omitted because it is the same as in the fifth embodiment. The first search unit 52 uses the acquired start address up to 14 bits (depth 7).
0th, 1st, 0th, 2nd, 1st, 0th, 2
Follow the th branch in order.

(2) At the branch point G, the range data 200p (0x1249) is located under the first (index 01) branch.
0000 to 0x12491000) exists, but range data does not exist at the tip of the 0th (index 00) branch which is a branch to be traced using the acquired start address as a key. Similar to the sixth embodiment, the first searching unit 52 traces another branch to search and acquire other range data, but the search result acquisition unit 60 determines that the search hits based on the comparison result of the comparison unit 53. If not, the search result is output to the first determination unit 71 of the registration unit 70. In the present embodiment, further detailed description of the search process will be omitted.

(3) Based on the output search result, the first judging unit 71 judges that nothing is stored in the end storage area traced by the start address of the designated address range 100n. It Based on this determination, the second searching unit 72 determines that the existing range data 200p (0x12490000 to 0x124910) that is one after the end is present.
00) is searched. The search result may be directly acquired from the search result of the first search unit 52 of the search unit 50, or may be acquired by the same process as the first search unit 52.

As a result, the searched range data 200
Since it is found that p does not include any address included in the range data 100n, the range data storage unit 7
3 extends the 0th (index 00) branch and the following branches to create a branch to the end that follows the start address of the range data 100n.
Store n as it is. 200n of stored range data
And

Next, FIG. 12B will be described. (4) Since the existing range data 200p that is one after the new range data 200n is already known by the processing of the first search unit 52 or the second search unit 72, the bidirectional link setting unit 77 determines that the existing range data The bidirectional link between 200p and the existing range data (not shown) before this is removed, and the new range data 200n and the existing range data 200p
Change the links so that and are connected by a bidirectional link.
Further, the existing range data before the unlinked new range data 200n and the new range data 200n are connected by a bidirectional link.

(5) Subsequently, the third searching unit 74, the second judging unit, the stored data deleting unit, and the bidirectional link setting unit 77.
By doing so, the bidirectional link is replaced while deleting unnecessary range data. This processing is the same as (6) and (7) of the eighth embodiment, and therefore the description is omitted.

The embodiments of the present invention have been described above.
The present invention is not limited to the InfiniBand system described above, but can be applied to communication systems using various communication hardware. Further, in the embodiment, the data processing device is provided with both the search function and the registration function, but it is needless to say that the device may have only one function, and the form thereof is not particularly limited.

(Supplementary Note 1) A data storage device having a 2 ^ n (2 to the n-th power) branch tree type data structure and storing a set of a plurality of first data, wherein the first data is , Data indicating a predetermined address range and its attributes,
Is a branch point where each n-bit value obtained by dividing the second data associated with the first data by n bits from the higher order is an index in the 2 ^ n-ary tree. A data storage device characterized by being stored in a storage area at an end obtained by sequentially tracing from the upper part of a branch tree. (Supplementary note 2) The data storage device according to supplementary note 1, wherein the second data is represented by a binary number. (Supplementary note 3) The data storage device according to supplementary note 1 or 2, wherein the second data is a start address of the predetermined address range. (Supplementary note 4) The data storage device according to any one of supplementary notes 1 to 3, wherein the attribute is presence or absence of an access right to the predetermined address range. (Supplementary note 5) In the data storage device according to Supplementary note 3 or Supplementary note 4, the first data may include branches corresponding to all addresses included in the address range, in addition to an end of a branch point corresponding to the start address. A data storage device characterized by being stored at the end of a point. (Supplementary note 6) In the data storage device according to any one of supplementary notes 1 to 4, when there is only one terminal having a storage area in which the first data is stored under a certain branch point, The data storage device, wherein the certain branch point is omitted in the 2 ^ n-ary tree. (Supplementary note 7) The data storage device according to supplementary note 6, wherein an index number of the omitted branch point is stored at a branch point immediately below the omitted branch point. (Supplementary note 8) The data storage device according to any one of supplementary notes 1 to 7, characterized in that the terminals storing the plurality of first data are connected to each other by a bidirectional link. apparatus. (Supplementary note 9) In the data storage device according to supplementary note 8, the bidirectional link stores the first data at the end sorted by the start address of a predetermined address range stored in the storage area at the end at D1 and D2. ． ． ． When Dk, D
1 and D2, D2 and D3. ． ． , D (k-1) and Dk having first data respectively connected at their respective ends with a bidirectional link. (Supplementary note 10) A data storage device having a 2 ^ n-branch data structure and storing a set of a plurality of first data, wherein the first data has a predetermined address range and its attributes. N bits obtained by dividing the second data associated with the first data by n bits from the most significant bit in the storage area of the branch point of the mth layer of the 2 ^ n-ary tree. , The address of the branch point of the (m + α) th layer, which is directly below the mth layer and has an index equal to the (m + α) th n-bit value, in the area of the index equal to the mth n-bit value from the upper Is stored and the m + α layer is the bottom layer, the first data is stored in the storage area of the bottom layer having an index equal to the value of the lower n bits of the second data. is being done Data storage device according to claim. (Supplementary note 11) The data storage device according to supplementary note 10, wherein the second data is represented by a binary number. (Supplementary note 12) The data storage device according to supplementary note 10 or supplementary note 11, wherein the second data is a start address of the predetermined address range. (Supplementary note 13) The data storage device according to any one of supplementary notes 10 to 12, wherein the attribute is presence or absence of an access right to the predetermined address range. (Supplementary note 14) In the data storage device according to any one of supplementary notes 10 to 13, the α is 1, and the first
The data storage device is stored in a storage area in the lowermost layer corresponding to the start address and in a storage area in the lowermost layer corresponding to all addresses included in the address range. (Supplementary note 15) In the data storage device according to any one of supplementary notes 10 to 14, there is only one index area in the lowest layer in which the first data is stored below the branch point of the m-th layer. The data storage device according to claim 1, wherein α is 2. (Supplementary note 16) In the data storage device according to supplementary note 15, in the storage area of the branch point of the (m + 1) th layer, the second storage area
An area having an index equal to the (m + 1) th n-bit value from the most significant data is a non-storage area. (Supplementary Note 17) In the data storage device according to Supplementary Note 15 or Supplementary Note 16, in the storage area of the branch point of the (m + 2) th layer, an area having an index equal to the value of the n + th bit from the top of the second data The data storage device according to claim 1, wherein the data stored in includes the m + 1th n-bit value from the most significant bit of the second data. (Supplementary note 18) In the data storage device according to any one of supplementary notes 10 to 17, further, in the storage area of the lowermost layer in which the plurality of first data are stored, the storage areas of the lowermost layer are mutually adjacent. A data storage device, in which addresses of each other are stored so as to be connected by a bidirectional link. (Supplementary note 19) In the data storage device according to supplementary note 18, the bidirectional link may include the first data at the end sorted by the start address of a predetermined address range stored in the storage area at the end at D1 and D2. ． ． ． Assuming that Dk is D1, D2 and D2 are stored in the lowermost storage area.
3. ． ． , D (k-1) and Dk, the respective addresses having the first data of Dk are stored so that their respective ends are linked to each other. (Supplementary note 20) A data processing device that performs predetermined processing on the data storage device according to any one of supplementary notes 1 to 19 so that a predetermined device can access predetermined data. In the storage area of the first
The data processing device capable of memorizing the above data so that it can be referred to by using a predetermined address within the address range of the first data as a reference key. (Supplementary note 21) A data processing device that performs a predetermined process on the data storage device according to any one of supplementary notes 1 to 19 so that a predetermined device can access predetermined data, the address of the predetermined data. And a search means for searching the data storage device for the third data indicating the range and its attribute, and a processing means for performing a predetermined process based on the search result of the search means. Characteristic data processing device. (Supplementary note 22) In the data processing device according to supplementary note 21, the search unit is a fourth unit related to the third data.
First acquisition means for acquiring the above data, second acquisition means for acquiring n bits of the fourth data acquired by the first acquisition means from the upper bits, and n acquired by the second acquisition means. Search means for searching a predetermined end for storing the first data on the basis of each value of the bit, and the first storage stored in the storage area for the predetermined end on the basis of the search result of the search means. And a comparing means for comparing the data obtained by the third obtaining means with the predetermined data, and the third data is obtained based on the comparison result by the comparing means. A data processing device, which determines whether or not the data is stored in the data storage device. (Supplementary note 23) The data processing apparatus according to supplementary note 22, wherein the fourth data is a start address of the address range designated for access. (Supplementary Note 24) In the data processing device according to Supplementary Note 22 or Supplementary Note 23, the first acquisition unit has a conversion unit that converts the fourth data into a binary number when the fourth data is not a binary number, and A data processing device characterized by acquiring converted data. (Supplementary note 25) In the data processing device according to any one of supplementary notes 22 to 24, the predetermined terminal searched by the searching means is n obtained by the second acquiring means.
The branch point having each bit value as an index is 2n
End (M) obtained by sequentially tracing from the top of the branch tree
A data processing device characterized in that (Supplementary note 26) In the data processing device according to any one of supplementary notes 22 to 24, the search unit may be configured to execute the first operation if the first data is not stored in the storage area of the terminal (M). 1. A data processing device, characterized in that, among the ends storing 1 data, an end in the vicinity of the end (M) is searched in the 2 ^ n-ary tree and the end in the vicinity is made a predetermined end. . (Supplementary note 27) In the data processing device according to supplementary note 26, when the searching means searches for an end near the end (M), the end in which the first data is stored in the data storage device. Of the two, the second data is the end that stores the first data having a smaller value than the fourth data, and further the end that is closest to the end (M) in the 2 ^ n-ary tree is A data processing device characterized by searching. (Supplementary note 28) In the data processing device according to any one of supplementary notes 23 to 27, the comparison unit may include a start address (ai) of a predetermined address range of the first data,
First comparing means for comparing the start address (a) of the access-specified address range of the third data, and the end address (b) of the predetermined address range of the first data.
i) is compared with an end address (b) of the access-specified address range of the third data, and the comparison result by the first comparison means is ai <= a, and The comparison result by the second comparison means is b <=
The data processing device, wherein the search means determines that the third data is stored in the data storage device when the data is bi. (Supplementary note 29) In the data processing device according to any one of supplementary notes 21 to 28, when the search unit determines that the third data is not stored in the data storage device, the processing unit Is a data processing device, wherein the third data is stored in the data storage device, and an attribute that allows the predetermined device to access the predetermined data is added. (Supplementary note 30) In the data processing device according to supplementary note 29, the processing means may be arranged in the first (M) storage area.
First determination means for determining whether or not the data is stored, and storage means for storing the third data in the data storage device based on the determination result by the first determination means. A data processing device characterized by: (Supplementary note 31) In the data processing device according to supplementary note 30, when the storage unit determines by the first determination unit that the first data is not stored in the storage area of the end (M). Of the terminals storing the first data in the data storage device, the terminal storing the first data having a second data value smaller than the fourth data, and And a first searching means for searching a terminal closest to the terminal (M) in the 2 ^ n-ary tree, and the third data is stored in the data storage device based on a search result of the first searching means. A data processing device characterized by storing in a. (Supplementary note 32) In the data processing device according to supplementary note 31, the storage unit searches the end storing the first data by which the second data is smaller than the fourth data by the first search unit. If not, the data processing device stores the third data in the storage area of the terminal (M). (Supplementary note 33) In the data processing device according to supplementary note 31, when the certain searching end is searched for by the first searching means, the storage means stores the first data stored in the searched end storage area. Is replaced with fourth data obtained by merging the first data and the third data. (Supplementary note 34) In the data processing device according to any one of supplementary notes 31 to 33, the storage means stores the first data in the storage area of the terminal (M) by the first determination means. A data processing device, characterized in that, when it is determined that the data is present, the first data already stored in the end (M) storage area is replaced with the third data. (Supplementary note 35) In the data processing device according to any one of supplementary notes 30 to 34, in the data storage device, the first data is stored only at an end of a branch point corresponding to a start address of the first data. If so, the processing means further comprises the first means in the data storage device.
Second searching means for searching the end storing the first data having a value larger than the second data at the end stored by the storage means among the ends storing the data And the first data stored in the storage area at the end searched by the second searching means,
The second judgment means for judging the inclusion relation with the data stored by the storage means, and the first data stored in the terminal storage area searched by the second search means by the second judgment means Data comprising: a deletion unit that deletes the data stored in the end storage area searched by the second searching unit when it is determined that the data is stored in the data stored by the storage unit. Processing equipment. (Supplementary note 36) In the data processing device according to any one of supplementary notes 27 to 35, when the plurality of first data stored in the data storage device are connected to each other by a bidirectional link, the search means is The searching means or the processing means having the first searching means or the second
The data processing device, wherein the search means performs a search process using the bidirectional link. (Supplementary note 37) In the data processing device according to any one of supplementary notes 30 to 36, when the plurality of first data stored in the data storage device are linked to each other by a bidirectional link, the processing means The data processing device further comprises link changing means for changing the bidirectional link when an end of the data storage device storing the first data is added or deleted. (Supplementary note 38) A data processing method for performing a predetermined process on the data storage device according to any one of supplementary notes 1 to 19 in order for a predetermined device to access predetermined data, the address of the predetermined data. A first step of searching in the data storage device for third data, which is data indicating a range and its attributes, and as a result of the search in the first step, the third data is stored in the data storage device. If it is determined that the third data is stored in the data storage device as a result of the second step of permitting access to the predetermined data of the predetermined device and the search by the first step. If it is determined that the third data is not stored, the third data is stored in the data storage device to allow the predetermined device to access the predetermined data. Data processing method characterized by having a flop. (Supplementary note 39) In the data processing method according to supplementary note 38, the first step includes a first acquisition step of acquiring fourth data related to the third data, and the first step.
The first data is acquired based on the second acquisition step of acquiring n bits of the fourth data acquired by the acquisition step from the upper bits and each value of n bits acquired by the second acquisition step. A search step of searching for a predetermined end to be stored, a third acquisition step of acquiring the first data stored in the storage area of the predetermined end based on the search result of the search step, and the third A comparison step of comparing the data acquired in the acquisition step with the predetermined data is provided, and it is determined whether or not the third data is stored in the data storage device based on the comparison result of the comparison step. A data processing method characterized by the above. (Supplementary note 40) In the data processing method according to Supplementary note 38 or Supplementary note 39, in the second step, a branch point having each n-bit value acquired in the second acquisition step as an index is the 2 ^ n-ary tree. A first determination step of determining whether or not the first data is stored in the end storage area obtained by sequentially tracing from the upper order of
A storage step of storing the third data in the data storage device based on a determination result of the determination step. (Supplementary note 41) A data processing program recorded on a computer-readable medium and performing a predetermined process on the data storage device according to any one of supplementary notes 1 to 19 so that a predetermined device can access predetermined data. A first step of searching the data storage device for third data, which is data indicating an address range of the predetermined data and its attribute, and the third step as a result of the search by the first step. If it is determined that the data is stored in the data storage device, the second step of permitting access to the predetermined data of the predetermined device, and the search result by the first step, If it is determined that the data of No. 3 is not stored in the data storage device, the third data is stored in the data storage device and the predetermined data is stored. Data processing program, characterized by a third step of permitting access to said predetermined data location. (Supplementary note 42) In the data processing program according to supplementary note 41, the first step is obtained by a first obtaining step of obtaining fourth data related to the third data, and the first obtaining step. A second acquisition step of acquiring n bits of the fourth data from the upper bits,
Based on each value of n bits acquired by the second acquisition step, a search step of searching for a predetermined end that stores the first data, and the predetermined end based on the search result of the search step. A third acquisition step for acquiring the first data stored in the storage area of the storage device, and a comparison step for comparing the predetermined data with the data acquired by the third acquisition step, and the comparison by the comparison step. A data processing program, which determines whether or not the third data is stored in the data storage device based on a result. (Supplementary note 43) In the data processing program according to Supplementary note 41 or Supplementary note 42, in the second step, a branch point having each value of n bits acquired in the second acquisition step as an index is the 2 ^ n-ary tree. A first determination step of determining whether or not the first data is stored in the storage area at the end obtained by sequentially tracing from the higher order, and the third determination step based on the determination result of the first determination step. And a storage step of storing data in the data storage device.

[0126]

As described above in detail, according to the present invention, the number of processings for following a link and the comparison processing of address range data, which are performed at the time of searching, are significantly reduced. For example, in a computer whose address is represented by 32 bits,
Consider the case where 1000 pieces of range data are registered. In this case, with the conventional search method, an average of 500 pieces of range data will be examined, but with the search method of the present invention, at most 32 pieces of range data will be searched.
It is only necessary to examine one range data after the process of following the link for the number of times (or 16 times, 11 times ...). Therefore,
INDUSTRIAL APPLICABILITY The present invention has an effect that the search processing is faster and the search time is shorter than that of the conventional data structure.

[Brief description of drawings]

FIG. 1 is a diagram showing a difference between a search process using a conventional technique and a search process according to the present invention.

FIG. 2 is a diagram showing an example of a quadtree data structure according to the first embodiment.

FIG. 3 is a diagram showing an example in which a certain branch point is omitted in the quadtree data structure according to the second embodiment.

FIG. 4 is a diagram showing an example in which a quadtree structure according to a third embodiment is arranged on a memory.

FIG. 5 is a diagram showing an example in which a bidirectional link is added to range data of a quadtree type data structure in the fourth embodiment.

FIG. 6 is an example of a block diagram showing a functional configuration of a data processing device.

FIG. 7 is a diagram specifically showing a flow of search processing in the fifth embodiment.

FIG. 8 is a diagram specifically showing a flow of search processing in the sixth embodiment.

FIG. 9 is a diagram specifically showing a flow of search processing in the seventh embodiment.

FIG. 10 is a diagram specifically showing a flow of registration processing in the eighth embodiment.

FIG. 11 is a diagram specifically showing a flow of registration processing in the ninth embodiment.

FIG. 12 is a diagram specifically showing a flow of registration processing in the tenth embodiment.

FIG. 13 is a diagram showing an outline of functions of a system using InfiniBand.

[Explanation of symbols]

10, 10a memory, 30, 30a address range storage area, 50 search unit, 51 start address acquisition unit, 5
2 1st search part, 53 comparison part, 60 search result acquisition part, 70 registration part, 71 1st judgment part, 72 2nd search part, 73 range data storage part, 74 3rd search part, 75
Second determination unit, 76 Stored data deletion unit, 77 Bidirectional link setting unit, 90 Data processing device.

Claims (5)

[Claims] 1. A data storage device having a 2 ^ n (2 to the nth power) branch tree type data structure and storing a set of a plurality of first data, wherein the first data is a predetermined number. Data indicating the address range and the attribute thereof, the first data is n data obtained by dividing the second data associated with the first data by n bits from the upper order in a 2 ^ n tree. A data storage device characterized in that a branch point whose bit value is used as an index is stored in a storage area at an end obtained by sequentially tracing from a higher order of a 2 ^ n branch tree. 2. A data processing device for performing a predetermined process on a data storage device according to claim 1 in order for a predetermined device to access predetermined data, wherein an address range of the predetermined data and an attribute thereof. Data including a search unit that searches the data storage device for the third data that is the data indicating the above, and a processing unit that performs a predetermined process based on the search result of the search unit. Processing equipment. 3. The data processing apparatus according to claim 2, wherein the search unit is acquired by the first acquisition unit that acquires fourth data related to the third data, and the first acquisition unit. Second acquisition means for acquiring n bits of the fourth data from the upper bits, and a predetermined terminal for storing the first data based on each value of the n bits acquired by the second acquisition means. And a third means for acquiring the first data stored in the predetermined end storage area based on a search result of the searching means.
An acquisition unit and a comparison unit that compares the predetermined data with the data acquired by the third acquisition unit, and the third data is stored in the data storage device based on a comparison result by the comparison unit. A data processing device, characterized in that it determines whether or not it is present. 4. A data processing method for performing a predetermined process on a data storage device according to claim 1 in order for a predetermined device to access predetermined data, the address range of the predetermined data and an attribute thereof. A third step of searching in the data storage device for third data, which is data indicating that the third data is stored in the data storage device as a result of the search in the first step. If determined, the second step of permitting the predetermined device to access the predetermined data, and the third data is not stored in the data storage device as a result of the search by the first step. And the third step of storing the third data in the data storage device and permitting the predetermined device to access the predetermined data. A data processing method characterized by. 5. A data processing program recorded on a computer-readable medium and performing a predetermined process on a data storage device according to claim 1, for a predetermined device to access predetermined data. A first step of searching in the data storage device for third data, which is data indicating an address range of the predetermined data and its attribute; and as a result of the search in the first step, the third data is the When it is determined that the third data is stored in the data storage device, the second step of permitting access to the predetermined data of the predetermined device, and the result of the search by the first step, the third data is When it is determined that the third data is not stored in the data storage device, the third data is stored in the data storage device and the predetermined data of the predetermined device is stored. And a third step of permitting access to the data.