JP2003101415A - Data conversion device, data generation device, data conversion program, data generation program, data conversion method and data generation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a data conversion device, in which power consumption is reduced, data can be easily made correspondent with a straight binary form or complement form of '2' and conversion with the complement form of '2' can be performed by uniform and simple processing without depending on the bit length of data. SOLUTION: On the data input side of a memory 10, a numerical expression conversion part 20 is provided for converting the data in the straight binary form or complement form of '2' to data of a prescribed form. The numerical expression conversion part 20 divides the data of the straight binary form or complement form of '2' to two bit blocks each composed of continuous 8 bits and concerning the bit block having the most significant bit of '1', the bits of such a bit block are inverted except for the most significant bit.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for converting data in straight binary format or 2's complement format into data in a predetermined format, an apparatus for generating data in the predetermined format, a program for realizing them, and a program for implementing them. In particular, in order to reduce power consumption, it is easy to associate data with straight binary format or 2's complement format, and conversion with 2's complement format depends on bit length of data. TECHNICAL FIELD The present invention relates to a data conversion device, a data generation device, a data conversion program and a data generation program, and a data conversion method and a data generation method that can be performed by uniform and simple processing without performing the processing.

[0002]

2. Description of the Related Art As a format for representing a numerical value using a plurality of bits, for example, a straight binary format and a binary format are used.
The complement form of is widely known. The straight binary format is a numerical expression format for representing a positive integer, for example,
As shown in FIG. 9, the 16-bit bit pattern is regarded as a numerical expression, and from “00000000 00000000” to “11111111 1111”.
By "1111", numerical values from "0" to "65535" in decimal are sequentially assigned. FIG. 9 is a table showing bit allocation rules in the straight binary format.

The two's complement format is a numerical expression format for representing a signed integer including a negative number. Here, the two's complement expression means subtracting one from the radix and subtracting each digit from each digit. Value is subtracted, and the value obtained by adding 1 to the subtraction result is expressed as a negative value of the digit. Therefore, for example, as shown in FIG. 10, the 16-bit bit pattern is regarded as a numerical expression, and “00000000 00000000” to “0111111”
By 1 11111111 ", decimal numbers from" 0 "to" 3276
Numerical values up to 7 ”are sequentially assigned, and“ 10000000 0000000
From "0" to "11111111 11111111", the decimal number "-"
Numerical values from "32768" to "-1" are sequentially assigned. FIG. 10 is a table showing bit allocation rules in the two's complement format.

However, voice data such as PCM and M
Like color component data such as PEG (Moving Picture Experts Group), a signed integer that fluctuates positively and negatively around "0", or a median value (32768 when represented by 16 bits) () Is an unsigned integer that fluctuates up and down around (center), and when expressing the data that has the characteristic that its fluctuation distribution approximates a normal distribution, a signed integer in 2's complement notation , From 0 to a negative side, in an unsigned integer in straight binary, when it fluctuates up and down around the median, a large number of bits are inverted. For example,
In the case of a 16-bit bit pattern, the variation from “0” to the negative side (the maximum value in the straight binary format) is “00000000” in the straight binary format and the two's complement format.
Since it changes from "000000000000" to "11111111 11111111",
The number of inversion bits is 16.

Generally, audio data such as PCM and MPEG
Color component data such as is CMOS (Complementary Mental
-Oxide Semiconductor), and arithmetic processing is performed with input / output to / from the semiconductor memory device. Therefore, the bit value of the data is CMO.
Since it is expressed by turning on / off S, if bit inversion is frequently performed by fluctuating to the positive side and the negative side centering on "0", the on / off of CMOS is frequently repeated. , Power consumption increases. Therefore, in consideration of such data characteristics, an encoding format has been proposed in which the CMOS is not frequently turned on / off repeatedly.

Conventionally, as a typical encoding format,
There is a so-called signed absolute value format, and its details are disclosed in a document "Technical White Paper on Low-Power LSI for Nikkei Microdevices Edition by Nikkei BP," issued on October 28, 1994. Signed absolute value format uses the most significant bit of a bit string consisting of multiple bits as a sign bit for determining positive / negative, and expresses the absolute value of other bits in straight binary format. , Is a format that represents a signed integer value. For example, as shown in FIG. 11, considering a 16-bit bit pattern as a numerical expression, numerical values from “00000000 00000000” to “01111111 11111111” are sequentially assigned as decimal numbers “0” to “32767”,
Numerical values from "0" to "-32767" in decimal are sequentially assigned from "00000000" to "11111111 11111111". FIG. 11 is a table showing a bit allocation rule in the conventional signed absolute value format.

[0007]

However, in the signed absolute value format, since "0" in decimal is assigned to two bit patterns, it is compared with the straight binary format and the two's complement format. There is one less numerical range that can be expressed. Therefore, in order to make one-to-one correspondence between the straight binary format or the two's complement format and the signed absolute value format, one extra bit is reserved for the signed absolute value format data, or "0" is set. It is only possible to make an exceptional assignment to any of the bit patterns assigned to ". In the former case, the amount of memory used and power consumption increase, and in the latter case, it is necessary to add circuits and processing for exceptional allocation.

Further, in the signed absolute value format, the most significant bit is used as the sign bit, and therefore, in order to convert the data in the two's complement format into the data in the signed absolute value format, the two's complement number is used. It is necessary to always grasp the bit length of the format data. This is because in the case of 32-bit data in the 2's complement format, the conversion process differs depending on whether it is one data in 32-bit units or four data in 8-bit units. Therefore,
Circuits and processes for determining the bit length of data must be added. Further, in order to convert the data of the two's complement format into the data of the signed absolute value format, the positive numerical value may be left as it is, but the negative numerical value needs to undergo the bit inversion and the addition. It can't be done easily.

Therefore, the present invention has been made by paying attention to the unsolved problem of the conventional technique, and aims at reduction of the power consumption, and between the straight binary format and the two's complement format. A data conversion device, a data generation device, and a data conversion device capable of easily associating data with each other and performing conversion between two's complement format by uniform and simple processing independent of the bit length of data. It is an object to provide a program and a data generation program, and a data conversion method and a data generation method.

[0010]

In order to achieve the above object, a data conversion apparatus according to a first aspect of the present invention is a numerical value expressed by an absolute value of 2 for a bit pattern expressed by a plurality of bits. Is a device for converting straight binary format data that expresses a numerical value by allocating, and comprises data conversion means for converting the straight binary format data into predetermined format data, wherein the data conversion means is the straight binary format. Regarding the data of the format in which the most significant bit is "1", the bits other than the most significant bit of the data are inverted.

With such a configuration, when straight binary format data is given, the data converting means sets the most significant bit of the given data whose most significant bit is "1". Bits other than the bits are inverted. As a result, the straight binary format data is converted into the predetermined format data. here,
The data conversion means only needs to invert the bits other than the most significant bit, may invert all the bits other than the most significant bit, and may convert the bits other than the most significant bit. The parts may be reversed. From the viewpoint of reducing power consumption, it is preferable to reverse all. Hereinafter, the same applies to the data conversion device according to claims 2, 4, 5, 7 and 8 and the data conversion program according to claims 15 to 20.

Further, the data conversion apparatus according to a second aspect of the present invention is of a straight binary format in which a numerical value is expressed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. An apparatus for converting data, comprising data conversion means for converting the data in the straight binary format into data in a predetermined format, wherein the data conversion means is composed of a plurality of consecutive bits of the data in the straight binary format. Bit blocks are divided, and for some or all of these bit blocks whose most significant bit is "1", bits other than the most significant bit of the bit block are inverted.

With such a configuration, when data in the straight binary format is given, the given data is divided into bit blocks by the data converting means, and the most significant bit of some or all of the bit blocks is divided. If the value is "1", bits other than the most significant bit of the bit block are inverted. As a result, the straight binary format data is converted into the predetermined format data.

Furthermore, a data conversion apparatus according to a third aspect of the present invention is the data conversion apparatus according to any one of the first and second aspects, which further performs arithmetic processing on the straight binary format data. An arithmetic processing unit, a storage unit for storing the arithmetic result of the arithmetic processing unit, and an inverse data conversion for converting the data in the predetermined format into the data in the straight binary format by the reverse procedure of the conversion procedure by the data conversion unit. Means and the data conversion means is provided on the data input side of the storage means,
The data inverse conversion means is provided on the data input side of the arithmetic processing means.

With such a configuration, when the straight binary format data is stored in the storage means, the data conversion means is provided on the data input side of the storage means. The format data is stored after being converted into the format data. Then, when performing the arithmetic processing, the data of a predetermined format is read from the storage means, but since the data inverse conversion means is provided on the data input side of the arithmetic processing means, it is read by the data inverse conversion means. The converted data is converted into straight binary format data, and the arithmetic processing means performs arithmetic processing on the converted data. The calculation result is stored in the storage means, and in this case as well, the data in the straight binary format is similarly converted into the data in the predetermined format and then stored.

Further, according to a fourth aspect of the present invention, there is provided a data conversion apparatus of a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits. An apparatus for converting data, comprising data conversion means for converting the data in the two's complement format into data in a predetermined format, the data conversion means having the most significant bit of the data in the two's complement format being " Bits other than the most significant bit of the data of "1" are inverted.

With such a configuration, when the data of the two's complement format is given, the data converting means selects the most significant bit of the given data for which the most significant bit is "1". Bits other than the upper bits are inverted. As a result, the data in the 2's complement format is converted into the data in the predetermined format. Further, claim 5 according to the present invention
The data conversion device described is a device for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, Data conversion means for converting data in the complement format into data in a predetermined format is provided, and the data conversion means divides the data in the 2's complement format into bit blocks made up of a plurality of consecutive bits, For some or all of the most significant bits being "1", the bits other than the most significant bit of the bit block are inverted.

With such a configuration, when data in the two's complement format is given, the given data is divided into bit blocks by the data conversion means, and the highest part of all or part of these bit blocks is divided. Bit is "1"
, The bits other than the most significant bit of the bit block are inverted. As a result, the data in the 2's complement format is converted into the data in the predetermined format.

Furthermore, a data conversion apparatus according to a sixth aspect of the present invention is the data conversion apparatus according to any one of the fourth and fifth aspects, further comprising arithmetic processing on the data in the 2's complement format. Data for converting the data of the predetermined format into the data of the two's complement format by the reverse procedure of the conversion procedure by the data conversion means, the storage means for storing the calculation result of the calculation processing means to perform. Inverse conversion means, the data conversion means is provided on the data input side of the storage means, and the data inverse conversion means is provided on the data input side of the arithmetic processing means.

With such a configuration, when the data in the two's complement format is stored in the storage means, the data conversion means is provided on the data input side of the storage means. Is stored after being converted into data of a predetermined format. Then, when performing the arithmetic processing, the data in the predetermined format is read from the storage means, but since the data inverse conversion means is provided on the data input side of the arithmetic processing means, the data inverse conversion means
The read data is converted into data in the two's complement format, and the arithmetic processing means performs arithmetic processing on the converted data. The calculation result is stored in the storage means, and in this case as well, the data in the two's complement format is similarly converted into the data in the predetermined format by the data conversion means and then stored.

Further, according to the present invention, in the data converter of claim 7, a straight binary format is used in which a numerical value is expressed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. An apparatus for converting data, comprising data conversion means for converting the data in the straight binary format into data in a predetermined format, and the data conversion means has the most significant bit "0" in the data in the straight binary format. The bits other than the most significant bit of the data are inverted.

With such a configuration, when straight binary format data is provided, the data converting means sets the most significant bit of the provided data for which the most significant bit is "0". Bits other than the bits are inverted. As a result, the straight binary format data is converted into the predetermined format data. further,
A data conversion apparatus according to claim 8 of the present invention is an apparatus for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits. A data conversion means for converting the data of the two's complement format into the data of a predetermined format, wherein the data conversion means has the most significant bit of the data of the two's complement format being "0". For, the bits other than the most significant bit of the data are inverted.

With such a configuration, when the data of the two's complement format is given, the data converting means selects the most significant bit of the given data for which the most significant bit is "0". Bits other than the upper bits are inverted. As a result, the data in the 2's complement format is converted into the data in the predetermined format. On the other hand, in order to achieve the above object, a data generating device according to a ninth aspect of the present invention is a device for generating data, and is provided with a data generating means for generating data of a predetermined format, and the data generating means. Expresses a numerical value by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, and the most significant bit of the data expressed by the numerical expression is “0”. As it is, if the most significant bit is "1" in the data expressed by the numerical representation, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format.

With such a configuration, the data generating means performs numerical expression by assigning a numerical value to the bit pattern expressed by a plurality of bits by the absolute value expression of 2, and such numerical expression is performed. The data whose most significant bit is "0" remains as it is, and the data whose most significant bit is "1" among such numerically represented data is other than the most significant bit. Inverted bits are generated as data of a predetermined format.

Here, the data generating means is only required to generate the data of a predetermined format by inverting the bits other than the most significant bit, and inverting all the bits other than the most significant bit. It may be generated as data of a predetermined format, or may be generated by inverting a part of bits other than the most significant bit as data of a predetermined format. From the viewpoint of reducing power consumption, it is preferable to generate the inverted data as the predetermined format data. Hereinafter, the data generation device according to claims 10 to 14, and claims 21 to 2.
The same applies to the data generation program described in No. 6.

Further, a data generating device according to a tenth aspect of the present invention is a device for generating data, comprising a data generating means for generating data of a predetermined format, and the data generating means has a plurality of bits. The numerical expression is performed by assigning a numerical value with the absolute value expression of 2 to the bit pattern expressed by
It is divided into bit blocks made up of a plurality of consecutive bits, and the most significant bit of some or all of these bit blocks remains as it is, while the most significant bit of some or all of these bit blocks remains the same. "1"
In the above, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format.

With such a configuration, the data generation means performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. Data is divided into bit blocks, the most significant bit of some or all of these bit blocks is "0" as it is, and the most significant bit of some or all of these bit blocks. Is "1", bits other than the most significant bit are inverted to generate data of a predetermined format.

Further, a data generating device according to claim 11 of the present invention is a device for generating data, comprising a data generating means for generating data of a predetermined format, and the data generating means comprises a plurality of bits. Numerical expression is performed by assigning a numerical value in 2's complement notation to the bit pattern expressed by, and the numerical expression in which the most significant bit is "0" remains as it is. If the most significant bit is "1" in the performed data, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format.

With such a configuration, the data generating means performs numerical expression by assigning a numerical value in a two's complement expression to a bit pattern expressed by a plurality of bits. Of the performed data, the data with the most significant bit being "0" remains unchanged, and
In the data in which the numerical expression is performed, the data with the most significant bit being “1” is generated by inverting the bits other than the most significant bit as the data of the predetermined format.

Further, a data generating device according to a twelfth aspect of the present invention is a device for generating data, comprising a data generating means for generating data of a predetermined format, and the data generating means has a plurality of bits. Numerical expression is performed by assigning a numerical value in 2's complement notation to the bit pattern expressed by, and the numerically expressed data is divided into bit blocks consisting of a plurality of consecutive bits,
Some or all of these bit blocks whose most significant bit is "0" are unchanged, while some or all of these bit blocks whose most significant bit is "1" are not the most significant bit. Inverted bits are generated as the predetermined format data.

With such a configuration, the data generating means performs the numerical expression by assigning a numerical value in the two's complement expression to the bit pattern expressed by a plurality of bits. The performed data is divided into bit blocks, the most significant bit of some or all of these bit blocks is “0”, and the most significant bit of some or all of these bit blocks remains. The data of "1" is generated by inverting the bits other than the most significant bit as data of a predetermined format.

Further, a data generating device according to a thirteenth aspect of the present invention is a device for generating data, comprising a data generating means for generating data of a predetermined format, and the data generating means has a plurality of bits. Numerical expression is performed by assigning a numerical value with an absolute value expression of 2 to the bit pattern expressed by. If the most significant bit is "0" in the processed data, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format.

With such a configuration, the data generating means performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. The data whose most significant bit is "1" among the data that has been subjected to the above-mentioned data is unchanged, and the most significant bit whose data is represented by the most significant bit is "0" except for the most significant bit. Inverted bits are generated as data of a predetermined format.

Further, a data generating device according to a fourteenth aspect of the present invention is a device for generating data, comprising a data generating means for generating data of a predetermined format, and the data generating means has a plurality of bits. Numerical expression is performed by assigning a numerical value to the bit pattern expressed by 2 in the two's complement expression, and the numerical expression whose data has the most significant bit "1" is directly converted to the numerical expression. When the most significant bit is "0" in the performed data, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format.

With such a configuration, the data generating means performs numerical expression by assigning a numerical value in the two's complement expression to the bit pattern expressed by a plurality of bits. Among the performed data, the one with the most significant bit being “1” remains as it is,
Among the data in which the numerical representation is performed, the data in which the most significant bit is "0" is generated by inverting the bits other than the most significant bit as the data in the predetermined format.

On the other hand, in order to achieve the above object, a data conversion program according to a fifteenth aspect of the present invention assigns a numerical value by an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. A program for converting straight binary format data for performing numerical expression, which is a program for causing a computer to execute processing realized as data conversion means for converting the straight binary format data into data of a predetermined format, The data converting means inverts bits other than the most significant bit of the straight binary format data having the most significant bit “1”.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 1
An effect equivalent to that of the described data converter is obtained. Furthermore, a data conversion program according to a sixteenth aspect of the present invention converts straight binary format data that performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. Is a program for causing a computer to execute a process realized as data conversion means for converting the data in the straight binary format into data in a predetermined format, wherein the data conversion means is in the straight binary format. Data is divided into bit blocks consisting of multiple consecutive bits, and for some or all of these bit blocks, the most significant bit of which is "1", the bits other than the most significant bit of the bit block are inverted. It is supposed to do.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 2
An effect equivalent to that of the described data converter is obtained. Furthermore, a data conversion program according to a seventeenth aspect of the present invention is for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits. Is a program for causing a computer to execute processing realized as data conversion means for converting data in the 2's complement format into data in a predetermined format, the data conversion means being the 2's complement format. Regarding the data of the format in which the most significant bit is "1", the bits other than the most significant bit of the data are inverted.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 4
An effect equivalent to that of the described data converter is obtained. Further, the data conversion program according to claim 18 of the present invention converts the data in the 2's complement format for performing the numerical expression by assigning a numerical value in the 2's complement expression to a bit pattern expressed by a plurality of bits. Is a program for causing a computer to execute processing realized as data conversion means for converting data in the 2's complement format into data in a predetermined format, the data conversion means being the 2's complement format. Format data is divided into bit blocks consisting of a plurality of consecutive bits, and for some or all of these bit blocks whose most significant bit is "1", bits other than the most significant bit of the bit block Is designed to be reversed.

With such a configuration, the program is read by the computer, and when the computer executes processing in accordance with the read program, the method according to claim 5
An effect equivalent to that of the described data converter is obtained. Furthermore, a data conversion program according to a nineteenth aspect of the present invention converts straight binary format data that performs numerical expression by assigning a numerical value by an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. Is a program for causing a computer to execute a process realized as data conversion means for converting the data in the straight binary format into data in a predetermined format, wherein the data conversion means is in the straight binary format. For data whose most significant bit is "0", bits other than the most significant bit of the data are inverted.

With such a configuration, the program is read by the computer, and when the computer executes processing in accordance with the read program, the method according to claim 7
An effect equivalent to that of the described data converter is obtained. Further, a data conversion program according to a twentieth aspect of the present invention converts data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits. Is a program for causing a computer to execute processing realized as data conversion means for converting data in the 2's complement format into data in a predetermined format, the data conversion means being the 2's complement format. With respect to data of the format in which the most significant bit is "0", bits other than the most significant bit of the data are inverted.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 8
An effect equivalent to that of the described data converter is obtained. on the other hand,
In order to achieve the above-mentioned object, a data generation program according to a twenty-first aspect of the present invention is a program for generating data, and causes a computer to execute processing realized as data generation means for generating data in a predetermined format. The data generation means performs a numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits,
Of the data expressed by the numerical expression, the data whose most significant bit is “0” is unchanged, and the data whose numerical expression is expressed by the most significant bit is “1”. The inverted data is generated as the data in the predetermined format.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 9
An operation equivalent to that of the described data generating device can be obtained. Further, a data generation program according to a twenty-second aspect of the present invention is a program for generating data, and is a program for causing a computer to execute a process realized as a data generation means for generating data of a predetermined format. The data generating means performs a numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, and the numerically expressed data is converted from a plurality of consecutive bits. And the most significant bit of some or all of these bit blocks is "0" as it is, and the most significant bit of some or all of these bit blocks is "1" Is designed to generate inverted data of the bits other than the most significant bit as the data of the predetermined format. .

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 1
An operation equivalent to that of the data generation device described in 0 is obtained. Further, a data generating program according to a twenty-third aspect of the present invention is a program for generating data, and is a program for causing a computer to execute a process realized as a data generating means for generating data of a predetermined format. The data generating means performs numerical expression by assigning a numerical value in a two's complement expression to a bit pattern expressed by a plurality of bits, and the most significant bit of the data expressed by the numerical expression is "0". The data having the most significant bit of "1" in the numerically expressed data is generated by inverting bits other than the most significant bit as the data of the predetermined format. ing.

With such a configuration, the program is read by the computer, and when the computer executes processing in accordance with the read program, the method according to claim 1
An effect equivalent to that of the data generating device described in 1 is obtained. Further, a data generating program according to a twenty-fourth aspect of the present invention is a program for generating data, and is a program for causing a computer to execute a process realized as a data generating means for generating data of a predetermined format. The data generating means performs a numerical expression by assigning a numerical value in a 2's complement expression to a bit pattern expressed by a plurality of bits, and the numerically expressed data is composed of a plurality of consecutive bits. If the most significant bit of some or all of the bit blocks is “0”, the most significant bit of some or all of the bit blocks is “1”. Inverted bits other than the most significant bit are generated as the data in the predetermined format.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 1
The same operation as the data generating device described in 2 is obtained. Further, a data generating program according to a twenty-fifth aspect of the present invention is a program for generating data, and is a program for causing a computer to execute a process realized as a data generating means for generating data of a predetermined format. The data generating means performs a numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, and the most significant bit of the data expressed by the numerical expression is "1". The data having the most significant bit of “0” in the numerically represented data is generated by inverting the bits other than the most significant bit as the data of the predetermined format. Has become.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 1
An effect equivalent to that of the data generator described in 3 is obtained. Further, a data generating program according to claim 26 of the present invention is a program for generating data, and is a program for causing a computer to execute a process realized as a data generating means for generating data of a predetermined format. The data generating means performs a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern expressed by a plurality of bits, and the most significant bit of the data expressed by the numerical expression is "1". The data with the most significant bit being “0” in the data expressed by the numerical expression is generated by inverting the bits other than the most significant bit as the data of the predetermined format. ing.

With such a configuration, the program is read by the computer, and when the computer executes the process in accordance with the read program, the method according to claim 1
An effect equivalent to that of the data generating device described in 4 is obtained. On the other hand, in order to achieve the above object, claim 2 according to the present invention
The data conversion method described in No. 7 is a method for converting straight binary format data in which a numerical value is expressed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. A data conversion step of converting binary format data into data of a predetermined format, wherein the data conversion step regards the most significant bit of the straight binary format data with the most significant bit being “1”. Invert bits other than.

Here, in the data conversion step, all bits other than the most significant bit may be inverted, all the bits other than the most significant bit may be inverted, or some of the bits other than the most significant bit may be inverted. You may. From the viewpoint of reducing power consumption, it is preferable to reverse all. The same applies to the data conversion method described in claims 28, 30, 31, 33 and 34 and the data generation method described in claim 35.

Further, in the data conversion method according to a twenty-eighth aspect of the present invention, a straight binary format in which a numerical value is expressed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. A method of converting data, comprising a data conversion step of converting the straight binary format data into a predetermined format data, wherein the data converting step comprises the straight binary format data consisting of a plurality of consecutive bits. It is divided into bit blocks, and for some or all of these bit blocks whose most significant bit is "1", the bits other than the most significant bit of the bit block are inverted.

Further, a data conversion method according to a twenty-ninth aspect of the present invention is the data conversion method according to any one of the twenty-seventh and twenty-eighth aspects, further comprising a data reading step of reading the data in the predetermined format from the storage means. When,
A data reverse conversion step of converting the data read in the data read step into the data in the straight binary format by a reverse procedure of the conversion procedure of the data conversion step, and an arithmetic process on the data converted in the data reverse conversion step. And a data storing step of storing the calculation result obtained through the calculation processing step in the storage means, and converting the data processed by the calculation processing step through the data conversion step. After that, the data is stored through the data storing step.

Furthermore, a data conversion method according to a thirtieth aspect of the present invention is a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits. A method of converting data, comprising a data conversion step of converting the data of the two's complement format into data of a predetermined format, wherein the data conversion step is such that the most significant bit of the data of the two's complement format is " Bits other than the most significant bit of the data of "1" are inverted.

Further, in the data conversion method according to a thirty-first aspect of the present invention, a numerical value representation is performed by assigning a numerical value by a two's complement expression to a bit pattern expressed by a plurality of bits. A method of converting data, comprising a data conversion step of converting the data in the 2's complement format into data in a predetermined format, the data converting step comprising converting the data in the 2's complement format into a plurality of consecutive bits. The bit blocks are divided into the following bit blocks, and for some or all of these bit blocks whose most significant bit is "1", the bits other than the most significant bit are inverted.

Furthermore, a data conversion method according to a thirty-second aspect of the present invention is the data conversion method according to any one of the thirty-third and thirty-first aspects, further comprising a data reading step of reading the data in the predetermined format from the storage means. When,
A data reverse conversion step of converting the data read in the data read step into the data of the two's complement format by the reverse procedure of the conversion procedure of the data conversion step;
An arithmetic processing step of performing arithmetic processing on the data converted in the data reverse converting step; and a data storing step of storing the arithmetic result obtained through the arithmetic processing step in the storage means. After the data that has been subjected to arithmetic processing is converted through the data conversion step, it is stored through the data storage step.

Furthermore, a data conversion method according to a thirty-third aspect of the present invention is a straight binary format in which a numerical value is expressed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits. A method for converting data, comprising a data conversion step of converting the data in the straight binary format into data in a predetermined format, wherein the data converting step has the most significant bit of the straight binary format data being "0". For those that are, the bits other than the most significant bit of the data are inverted.

Furthermore, a data conversion method according to a thirty-fourth aspect of the present invention is a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits. A method of converting data, comprising a data conversion step of converting the data of the two's complement format into data of a predetermined format, wherein the data conversion step is such that the most significant bit of the data of the two's complement format is " With respect to "0", the bits other than the most significant bit of the data are inverted.

On the other hand, in order to achieve the above object, a data generating method according to a thirty-fifth aspect of the present invention is a method for generating data, which includes a data generating step of generating data of a predetermined format, In the data generating step, a numerical expression is performed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, and the most significant bit of the data for which the numerical expression is performed is "0". For some data, the most significant bit "1" in the numerically expressed data is generated by inverting the bits other than the most significant bit as the data of the predetermined format.

Here, in the data generation step, it is only necessary to generate the data of a predetermined format by inverting the bits other than the most significant bit, and the data of the predetermined format by inverting all the bits other than the most significant bit. It may be generated, or may be generated by inverting a part of bits other than the most significant bit as data of a predetermined format. From the viewpoint of reducing power consumption, it is preferable to generate the inverted data as the predetermined format data. Hereinafter, claim 36
The same applies to the data generation methods described in Nos. 40 to 40.

A data generating method according to a thirty-sixth aspect of the present invention is a method for generating data, which includes a data generating step of generating data of a predetermined format, and the data generating step comprises a plurality of bits. Numerical expression is performed by assigning a numerical value to the bit pattern expressed by 2 with the absolute value expression, and the data expressed by the numerical expression is divided into bit blocks consisting of a plurality of consecutive bits, and these bit blocks are divided. If some or all of the bit blocks have the most significant bit "0", and if some or all of the bit blocks have the most significant bit "1", invert the bits other than the most significant bit. The generated data is generated as the data in the predetermined format.

A data generating method according to a thirty-seventh aspect of the present invention is a method of generating data, which includes a data generating step of generating data of a predetermined format, and the data generating step comprises a plurality of bits. Numerical expression is performed by assigning a numerical value in a two's complement notation to the bit pattern expressed by, and the data in which the most significant bit is "0" in the numerically expressed data is as it is,
If the most significant bit is "1" in the data represented by the numerical value, the bits other than the most significant bit are inverted to generate the data of the predetermined format.

Further, a data generating method according to a thirty-eighth aspect of the present invention is a method of generating data, which includes a data generating step of generating data of a predetermined format, and the data generating step comprises a plurality of bits. Numerical expression is performed by assigning a numerical value to the bit pattern expressed by 2 in the two's complement expression, and the data expressed by the numerical expression is divided into bit blocks consisting of a plurality of consecutive bits, Some or all of the most significant bits are "0", while some or all of the bit blocks have the most significant bit of "1", the bits other than the most significant bit are inverted. Objects are generated as the data in the predetermined format.

Furthermore, a data generating method according to a thirty-ninth aspect of the present invention is a method of generating data, which includes a data generating step of generating data of a predetermined format, and the data generating step comprises a plurality of bits. Numerical expression is performed by assigning a numerical value with an absolute value expression of 2 to the bit pattern expressed by. In the data for which the most significant bit is “0”, the data other than the most significant bit is inverted and generated as the data of the predetermined format.

A data generating method according to claim 40 of the present invention is a method for generating data, which includes a data generating step for generating data of a predetermined format, and the data generating step comprises a plurality of bits. Numerical expression is performed by assigning a numerical value in 2's complement notation to the bit pattern expressed as
If the most significant bit is "0" in the data represented by the numerical value, the bits other than the most significant bit are inverted to generate the data of the predetermined format.

[0064]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings. 1 to 4 are diagrams showing a first embodiment of a data conversion device, a data generation device, a data conversion program and a data generation program, and a data conversion method and a data generation method according to the present invention.

In this embodiment, the data conversion device, the data generation device, the data conversion program and the data generation program, and the data conversion method and the data generation method according to the present invention are, as shown in FIG. This is applied to the case where data in the two's complement format is input to the memory 10 and arithmetic processing is performed by an ALU (Arithmetic Logic Unit) 50.

First, the arithmetic processing unit 10 to which the present invention is applied.
The configuration of 0 will be described with reference to FIG. FIG. 1 is a block diagram showing the configuration of an arithmetic processing device 100 to which the present invention is applied. The processor 100, as shown in FIG.
Memory 1 for storing data to be processed
0, a numeric expression conversion unit 20 provided between the memory 10 and an external device and for converting data in a straight binary format or two's complement format provided from the external device into data of a predetermined format, and read from the memory 10. Register file 30 for holding the stored data, and register file 3
Pipeline registers 42 and 44 that hold the data read from 0, and an ALU 50 that performs arithmetic processing on the data held in the pipeline registers 42 and 44
And a pipeline register 46 for holding data processed by the ALU 50. The register file 30 and the pipeline registers 42 to 46
Uses a CMOS as a storage element.

The register file 30 has multiplexers 32, 34 and 36 and a plurality of internal registers (not shown), and transfers the data read from the pipeline register 46 or the memory 10 to either of the internal registers via the multiplexer 32. The data stored and stored in the internal register is output to the pipeline register 42 via the multiplexer 34 and to the pipeline register 44 via the multiplexer 36, respectively.

Next, FIG. 2 shows the configuration of the numerical expression conversion unit 20.
It will be described in detail with reference. Figure 2 is a numerical representation conversion unit
It is a circuit diagram which shows the structure of 20. Numerical representation conversion unit 20
Shows 16-bit data in parallel as shown in FIG.
0 to 6 bits out of 16 data signal lines transmitted to
Bit signal line x corresponding to ₀~ X₆Each one day
Exclusive OR circuit ex connected to the input side₀~ Ex₆When,
Bit corresponding to the 8th to 14th bits of the data signal line
Signal line x₈~ X₁₄To each data input side
Exclusive OR circuit ex₇~ Ex₁₃Consists of and
It Here, the 15th bit is the most significant bit.

A bit signal line x ₇ corresponding to the 7th bit of the data signal lines is connected to the other data input side of each exclusive OR circuit ex _{0 to} ex ₆ , and each exclusive OR circuit. A bit signal line x ₁₅ corresponding to the 15th bit of the data signal lines is connected to the other data input side of the circuits ex _{7 to} ex ₁₃ . With this configuration, when the 7th bit of the data on the data signal line is "1",
The exclusive OR circuits ex _{0 to} ex ₆ make the data 0 to 0
6 bit is inverted, and when the 15th bit is "1", the 8 to 14 bit data is inverted by the exclusive OR circuit ex ₇ ~ex _13.

Therefore, when the data given from the outside is in the straight binary format, the data is converted as shown in FIG. FIG. 3 is a table showing bit allocation rules in the encoding format according to the present invention. As for the data in the straight binary format, as shown in FIG. 3, a numerical value is expressed by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by 16 bits, and the numerically expressed data is It is divided into two consecutive 8-bit bit blocks, and each bit block whose most significant bit is "0" remains as it is, and each bit block whose most significant bit is "1" is its most significant bit. Converted to data with bits other than the upper bits inverted. By converting to such a predetermined format, 1
For data having a characteristic that fluctuates up and down around the median value of 6 bits, the number of bits to be inverted becomes small.

If the externally applied data is in the 2's complement format, the data is converted as shown in FIG. FIG. 4 is a table showing bit allocation rules in the encoding format according to the present invention. As for the data in the two's complement format, as shown in FIG. 4, the numerical expression is performed by assigning a numerical value in the two's complement expression to a bit pattern expressed by 16 bits, It is divided into two consecutive 8-bit bit blocks, and each bit block whose most significant bit is "0" remains as it is, and each bit block whose most significant bit is "1" is its most significant bit. Converted to data with bits other than the upper bits inverted. By such conversion into the predetermined format, for data having a characteristic that its value fluctuates toward the positive side and the negative side around "0", the number of bits to be inverted is reduced even if it fluctuates toward the positive side and the negative side. .

Next, the operation of the first embodiment will be described. When data in straight binary format or 2's complement format is given from the outside, the given data is divided into two bit blocks of upper 8 bits and lower 8 bits by the numerical expression conversion unit 20, and each bit block of each bit block For those bits whose most significant bit is "1", bits other than the most significant bit of the bit block are inverted. As a result, the data in the straight binary format or the two's complement format is converted into the data in the predetermined format and stored in the memory 10.

The data in the memory 10 is passed through the register file 30 and the pipeline registers 42 and 44 to AL.
Data processed by U50 has a characteristic that the data given from the outside fluctuates up and down around the 16-bit median value, or the value fluctuates positively and negatively around "0". In the case of the data having the characteristic, the register file 30 and the pipeline registers 42 and 44 have a small number of bits to be inverted, so that the number of times the CMOS is turned on / off is reduced.

As described above, in the present embodiment, the memory 1 is provided with the numerical expression conversion unit 20 for converting straight binary format or 2's complement format data into predetermined format data.
Provided on the data input side of 0, the numerical expression conversion unit 20 divides the data of straight binary format or 2's complement format data into two continuous bit blocks of 8 bits, and the highest bit of each bit block. Bits other than the most significant bit of the bit block are inverted with respect to the bit having "1".

As a result, data having a characteristic of fluctuating up and down around the median value and data having a characteristic of fluctuating positively and negatively around "0" can be converted into a predetermined format. Since the number of bits to be inverted is small, the power consumption can be relatively reduced as compared with the straight binary format or the two's complement format.
In addition, conversion to a predetermined format only involves inverting specific bits of data in straight binary format or 2's complement format, so there is no need to reserve an extra bit or perform exceptional allocation. Corresponding of data between the straight binary format or the two's complement format becomes relatively easy as compared with the value format. Further, since the conversion to the predetermined format can be performed only by the inversion operation of the specific bit, the circuit and the processing for the conversion can be made relatively simple as compared with the signed absolute value format. Further, it is possible to deal with data in units of arbitrary bit lengths by a uniform circuit or processing without depending on the bit lengths of data.

In the first embodiment, the numerical expression conversion unit 20 corresponds to the data conversion means described in claim 1, 2, 4 or 5, and the conversion process by the numerical expression conversion unit 20 is defined in claim 27. , 28, 30 or 31 corresponding to the data conversion step. Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 5 is a diagram showing a second embodiment of a data conversion device, a data generation device, a data conversion program and a data generation program, and a data conversion method and a data generation method according to the present invention. Only the parts different from the first embodiment will be described below, and the same parts will be denoted by the same reference numerals and description thereof will be omitted.

In this embodiment, the data conversion device, the data generation device, the data conversion program and the data generation program, and the data conversion method and the data generation method according to the present invention are, as shown in FIG. The present invention is applied to the case where the data in the 2's complement format is input to the memory 10 and the arithmetic processing is performed by the ALU 50. What is different from the first embodiment is the straight binary format or the 2's complement format and the present invention. Mutual conversion of data to and from the encoding format according to ALU
This is also done when inputting / outputting to / from 50.

First, the arithmetic processing unit 10 to which the present invention is applied.
The configuration of 0 will be described with reference to FIG. FIG. 5 is a block diagram showing the configuration of the arithmetic processing unit 100 to which the present invention is applied. In addition to the configuration according to the first embodiment, the arithmetic processing unit 100 includes an image output unit 60 that outputs data in the memory 10 to an image display unit (not shown) and an audio output unit (not shown), as shown in FIG. And an audio output unit 70 for outputting the data of the memory 10, and further between the pipeline register 42 and the ALU 50, between the pipeline register 44 and the ALU 50,
The numerical expression conversion unit 20 is provided between the pipeline register 46 and the register file 30, between the image output unit 60 and the image display device, and between the audio output unit 70 and the audio output device.

Next, the operation of the second embodiment will be described. When the ALU 50 performs the arithmetic processing, the arithmetic data converted by the numerical value display conversion unit 20 is read from the memory 10, passes through the register file 30 and the pipeline registers 42, 44, and is converted by the numerical expression conversion unit 20.
The data is converted back to the original format and given to the ALU 50. Then, the ALU 50 performs arithmetic processing on the given data, and the numerical expression conversion unit 20 reconverts it into a predetermined format and holds it in the pipeline register 46. Therefore, in the case of data having a characteristic that fluctuates up and down around the median value or data that has a characteristic that fluctuates positively and negatively around "0", the register file 30 and the pipeline register In the cases of 42 and 44, the number of bits to be inverted is small even if it changes to the positive side and the negative side, so that the number of times the CMOS is turned on / off is reduced. Further, even if the ALU 50 performs arithmetic processing in the straight binary format or the two's complement format, it can be used as it is without changing the content of the arithmetic processing.

When displaying an image on the image display device, the image data converted by the numerical display conversion unit 20 is read from the memory 10, passed through the image output unit 60, and then converted by the numerical expression conversion unit 20. The data is converted back to the original format and output to the image display device. Therefore, in the case of data having a characteristic that vertically fluctuates around a median value or color component data such as MPEG having a characteristic that the value fluctuates positively and negatively around "0", image output Since the number of bits to be inverted is small in the unit 60, the CMOS
The number of on / off switchings is reduced. Further, even if the image output unit 60 performs the output processing in the straight binary format or the two's complement format, it can be used as it is without changing the content of the output processing.

When outputting a voice with the voice output device, the voice data converted by the numerical display conversion unit 20 is read from the memory 10, passed through the audio output unit 70, and then converted by the numerical expression conversion unit 20. The data is converted back to the original format and output to the audio output device. Therefore, in the case of data having a characteristic of fluctuating up and down around the median value or voice data such as PCM having a characteristic of fluctuating positive and negative sides around "0", the audio output unit In 70, since the number of bits to be inverted is small, the number of times CMOS is switched on / off is reduced. In addition, even if the audio output unit 70 performs output processing in a straight binary format or a two's complement format,
It can be used as it is without changing the contents of the output process.

As described above, the present embodiment is provided with the ALU 50 for performing the arithmetic processing on the data in the straight binary format or the two's complement format, and the pipeline register 46 for storing the arithmetic result of the ALU 50. ,
The numerical expression conversion unit 20 is provided on the data input side of the pipeline register 46, and the numerical expression conversion unit 20 is provided on the data input side of the ALU 50.

As a result, data of a predetermined format is stored in the pipeline register 46 and data of the straight binary format or the two's complement format is given to the ALU50. Therefore, the ALU50 is in the straight binary format or the two's complement format. Even if the arithmetic processing is performed, the power consumption in the pipeline register 46 can be reduced without changing the content of the arithmetic processing. The same applies to the image output unit 60 and the audio output unit 70.

In the second embodiment, the numerical expression conversion unit 20 corresponds to the data conversion unit according to claim 1 or 6 or the data inverse conversion unit according to claim 3 or 6, and corresponds to the numerical expression conversion unit. The conversion step according to 20 corresponds to the data conversion step described in claims 27 to 32 or the data inverse conversion step described in claim 29 or 32. The ALU 50 corresponds to the arithmetic processing means according to claim 3 or 6, and is a pipeline register 46.
Corresponds to the storage means according to claim 3, 6, 29 or 32.

In the first and second embodiments, the data of the straight binary format or the two's complement format is converted into the data of the predetermined format by the exclusive OR circuits ex _{0 to} ex _13. However, the present invention is not limited to this, and a function equivalent to this may be realized by a computer program. In this case, the numerical expression conversion unit 20 corresponds to the data conversion means described in claims 15 to 18.

In the first and second embodiments, the conversion is performed by inverting the bits other than the most significant bit with respect to the data in which the most significant bit of each bit block is "1". However, the configuration is not limited to this, and the data in which the most significant bit of each bit block is “0” may be configured to be inverted by inverting the bits other than the most significant bit. Even with such a configuration, the same effect as that of the first and second embodiments can be obtained. In addition, the data conversion is the first
It may be performed by hardware as in the second embodiment or by a computer program.

In this case, the numerical expression conversion unit 20
Corresponds to the data conversion means described in claim 7, 8, 19 or 20, and the conversion process by the numerical expression conversion unit 20 corresponds to the data conversion step described in claim 33 or 34. Further, in the above-mentioned first and second embodiments, the configuration is such that straight binary format or 2's complement format data is converted into predetermined format data.
The present invention is not limited to this, and the data in the predetermined format may be directly generated without conversion from the data in the straight binary format or the data of two's complement format. The data may be generated by hardware as in the first and second embodiments, or may be generated by a computer program. Further, the data having the most significant bit “1” may be converted by inverting the bits other than the most significant bit, or the data having the most significant bit “0” may be converted into data other than the most significant bit. The conversion may be performed by inverting the bits.

In the first and second embodiments, the data in straight binary format or 2's complement format is divided into two consecutive 8-bit bit blocks, and each bit block is divided into two bit blocks. Although it is configured to perform determination of the most significant bit and inversion of other bits, the present invention is not limited to this, and in this case, 16 bits are grouped together and the determination of the most significant bit and other bits are performed. May be configured to be reversed.

In the case where data conversion or generation is realized by a computer program, CPU, ROM
In a computer in which the RAM and the RAM are connected to the bus, the program may be read into the RAM from a storage medium in which the computer program is stored and executed. Here, the storage medium is a semiconductor storage medium such as RAM or ROM, a magnetic storage type storage medium such as FD or HD, or C.
An optical read type storage medium such as D, CDV, LD, and DVD, and a magnetic storage type / optical read type storage medium such as MO, regardless of reading methods such as electronic, magnetic, and optical. Any storage medium that can be read by a computer is included.

Further, in the first and second embodiments, the data conversion device, the data generation device, the data conversion program and the data generation program, and the data conversion method and the data generation method according to the present invention will be described with reference to FIG. And, as shown in FIG. 5, data in straight binary format or 2's complement format is input to the memory 10 and AL
It was applied to the case of performing arithmetic processing by U50,
The present invention is not limited to this, and can be applied to other cases without departing from the gist of the present invention.

In the first and second embodiments described above, the structure of the internal register of the register file 30 was not particularly described, but in the case of handling the encoding format according to the present invention, the internal register is Can be stored as follows. In 2's complement format, when performing a signed load of byte or half word data,
Sign extension processing is performed in which a code bit is repeatedly copied and filled in a vacant bit in a register. However, in the encoding format according to the present invention, as shown in FIG. The sign bit is copied only to the most significant bit of and the rest is filled with "0". on the other hand,
In the case of no code, as shown in FIG. 7, all empty bits are filled with “0”. In this way, most of the bits to be filled in the vacant bits are the same “0” regardless of whether they are signed or unsigned, and the number of inverted bits is small. 6 and 7 are diagrams for explaining a method of storing data in a register when a code is added.

In the first and second embodiments, the coding format according to the present invention is applied to the case of handling integer type data, but the present invention is not limited to this, and the coding format according to the present invention is not limited to this. Can also be applied to the case of handling floating point type data, as shown in FIG. FIG. 8 is a diagram for explaining a case where the encoding format according to the present invention is applied to floating point type data.

The encoding format according to the present invention is provided with its own simple floating-point arithmetic unit in order to facilitate the arithmetic operation with secured precision. This is based on fixed-point arithmetic and adds a mechanism for managing the position of the decimal point. In FIG. 8, the 24-bit mantissa F is represented by a signed binary in the encoding format according to the present invention. Therefore, the 23rd bit indicates a sign bit, and this bit is always copied to the 31st bit. As a result, it is possible to quickly determine whether it is greater than, less than, or equal to zero. The exponent E is represented by 2's complement using 7 bits from the 24th bit to the 30th bit. However, 31
When the sign bit of the bit is "1", the exponent E is all inverted. This is because it is necessary to maintain compatibility. The floating point to which the encoding format according to the present invention is applied is
It is represented by the following formula (1). As you can see from the formula,
In the simple floating point to which the encoding format according to the present invention is applied, the decimal point position is managed only in 8-bit units.
This is to reduce the burden of decimal point alignment in addition and subtraction. F × 256 ^E (1)

[0094]

As described above, according to the data conversion device of the first to third or seventh aspects of the present invention,
For data that has a characteristic that it fluctuates up and down around the median value, and data that has a characteristic that its value fluctuates positively and negatively around "0", the number of bits to be inverted by conversion to a predetermined format is Since the number is small, it is possible to obtain an effect that the power consumption can be relatively reduced as compared with the straight binary format. In addition, conversion to a predetermined format only involves inverting specific bits of data in straight binary format or 2's complement format, so there is no need to reserve an extra bit or perform exceptional allocation. Compared with the value format, there is an effect that data can be relatively easily associated with the straight binary format or the two's complement format. Furthermore, since conversion to a predetermined format can be performed only by inverting specific bits, it is possible to make the conversion circuit and processing relatively simple compared to the signed absolute value format. The effect is also obtained. Further, it is possible to obtain the effect that data having a unit of an arbitrary bit length can be handled by a uniform circuit or processing without depending on the bit length of the data.

Further, according to the data conversion apparatus of the third aspect of the present invention, since the data in the predetermined format is stored in the storage means and the data in the straight binary format is given to the arithmetic processing means, the straight binary format is provided. Even in the conventional arithmetic processing device that performs arithmetic processing in the form, it is possible to obtain the effect that the power consumption in the storage unit can be reduced without changing the content of the arithmetic processing.

Further, according to the data conversion device of the present invention, the data having a characteristic that its value fluctuates positively and negatively around "0" is predetermined. The conversion to the form reduces the number of bits to be inverted even if it fluctuates to the positive side and the negative side, so that it is possible to obtain an effect that the power consumption can be relatively reduced as compared with the two's complement form. . In addition, since conversion to a prescribed format only inverts specific bits of 2's complement format data, there is no need to reserve an extra bit or make exceptional allocation, and to compare with the signed absolute value format. As a result, it is possible to obtain an effect that the data can be relatively easily associated with the two's complement format. Furthermore, since conversion to a predetermined format can be performed only by inverting specific bits, it is possible to make the conversion circuit and processing relatively simple compared to the signed absolute value format. The effect is also obtained. Further, it is possible to obtain the effect that data having a unit of an arbitrary bit length can be handled by a uniform circuit or processing without depending on the bit length of the data.

Further, according to the data conversion apparatus of the sixth aspect of the present invention, the data in the predetermined format is stored in the storage means, and the data in the two's complement format is given to the arithmetic processing means. Even in the conventional arithmetic processing device that performs the arithmetic processing in the complement format, the effect that the power consumption in the storage means can be reduced without changing the content of the arithmetic processing is also obtained.

On the other hand, according to the data generating apparatus of the ninth aspect, the tenth aspect or the thirteenth aspect of the present invention, the same effect as that of the data converting apparatus according to the first aspect can be obtained. Furthermore, according to the data generation device of the eleventh, twelfth or fourteenth aspect of the present invention, the same effect as that of the data conversion device of the fourth aspect can be obtained. On the other hand, according to the data conversion program of the fifteenth, sixteenth or nineteenth aspect of the present invention, the same effect as that of the data conversion device of the first aspect can be obtained.

Further, according to the data conversion program of the seventeenth, eighteenth or twentieth aspect of the present invention, the same effect as that of the data conversion device of the fourth aspect can be obtained. On the other hand, according to the data generation program of the twenty-first, twenty-second, or twenty-fifth aspect of the present invention, an effect equivalent to that of the data generation device of the ninth aspect can be obtained. Further, according to the data generation program of the twenty-third, twenty-fourth or twenty-sixth aspect of the present invention, an effect equivalent to that of the data generation device of the eleventh aspect can be obtained.

On the other hand, claims 27 to 29 according to the present invention.
According to the data conversion method described in Item 33, the same effect as that of the data conversion device described in Item 1 can be obtained. further,
According to the data conversion method of the 29th aspect of the present invention, the same effect as that of the data conversion apparatus of the 3rd aspect can be obtained. Furthermore, according to the data conversion method of the thirtieth to thirty-two or thirty-fourth aspects of the present invention, an effect equivalent to that of the data conversion apparatus of the fourth aspect can be obtained.

Further, according to the data conversion method of the thirty-second aspect of the present invention, the same effect as that of the data conversion device of the sixth aspect can be obtained. On the other hand, claim 3 according to the present invention
According to the data generation method described in 5, 36 or 39, the same effect as that of the data generation device according to claim 9 can be obtained.
Furthermore, according to the data generating method of the thirty-seventh, thirty-eighth or fortyth aspect of the present invention, the same effect as that of the data generating apparatus of the eleventh aspect can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an arithmetic processing unit 100 to which the present invention is applied.

FIG. 2 is a circuit diagram showing a configuration of a numerical expression conversion unit 20.

FIG. 3 is a table showing bit allocation rules in the encoding format according to the present invention.

FIG. 4 is a table showing bit allocation rules in the encoding format according to the present invention.

FIG. 5 is a block diagram showing a configuration of an arithmetic processing unit 100 to which the present invention is applied.

FIG. 6 is a diagram for explaining a method of storing data in a register when a code is added.

FIG. 7 is a diagram for explaining a method of storing data in a register when a code is added.

FIG. 8 is a diagram for explaining a case where the encoding format according to the present invention is applied to floating-point type data.

FIG. 9 is a table showing bit allocation rules in the straight binary format.

FIG. 10 is a table showing bit allocation rules in the two's complement format.

FIG. 11 is a table showing a bit allocation rule in a conventional signed absolute value format.

[Explanation of symbols]

100 arithmetic processing unit 10 memory 20 Numerical value conversion unit 30 register file 32-36 multiplexer 42-46 pipeline registers 50 ALU 60 Image output section 70 Audio output section

Claims (40)

[Claims] 1. A device for converting straight binary format data, which performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein Data conversion means for converting data into data of a predetermined format is provided, wherein the data conversion means is a bit other than the most significant bit of the data in which the most significant bit of the straight binary format data is "1". A data conversion device, which is configured to invert. 2. A device for converting straight binary format data, which performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein Data conversion means for converting data into data of a predetermined format is provided, and the data conversion means divides the data of the straight binary format into bit blocks made up of a plurality of consecutive bits, and a part or all of these bit blocks. A data conversion device, wherein bits other than the most significant bit of the bit block of the one having the most significant bit of "1" are inverted. 3. The arithmetic processing means according to claim 1, further comprising arithmetic processing means for performing arithmetic processing on the data in the straight binary format, and storage means for storing an arithmetic result of the arithmetic processing means. A data reverse conversion means for converting the data in the predetermined format into the data in the straight binary format by a reverse procedure of the conversion procedure by the data conversion means, and the data conversion means is provided on the data input side of the storage means, A data conversion device, wherein the data reverse conversion means is provided on the data input side of the arithmetic processing means. 4. An apparatus for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Data converting means for converting the data of the above-mentioned data into data of a predetermined format, wherein the data converting means is the data of the two's complement format in which the most significant bit is "1" except for the most significant bit of the data. The data conversion device is characterized in that the bit of is inverted. 5. An apparatus for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Data conversion means for converting the data of 2 to the data of a predetermined format, wherein the data conversion means converts the data of the 2's complement format,
It is divided into bit blocks made up of a plurality of consecutive bits, and for some or all of these bit blocks whose most significant bit is "1", the bits other than the most significant bit of the bit block are inverted. Data conversion device characterized by 6. The arithmetic processing means according to claim 4, further comprising arithmetic processing means for performing arithmetic processing on the data of the two's complement format, and storage means for storing the arithmetic result of the arithmetic processing means. And a data inverse conversion means for converting the data in the predetermined format into the data in the two's complement format by the reverse procedure of the conversion procedure by the data conversion means, and the data conversion means is provided on the data input side of the storage means. A data conversion device, wherein the data reverse conversion means is provided on the data input side of the arithmetic processing means. 7. An apparatus for converting straight binary format data, which performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein the straight binary format data is converted. Data conversion means for converting data into data of a predetermined format is provided, wherein the data conversion means is a bit other than the most significant bit of the data for which the most significant bit of the straight binary format data is "0". A data conversion device, which is configured to invert. 8. An apparatus for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Data conversion means for converting the data of the above format into data of a predetermined format, wherein the data conversion means is the data of the two's complement format whose most significant bit is "0" except for the most significant bit of the data. A data conversion device characterized by inverting the bits of. 9. An apparatus for generating data, comprising data generating means for generating data of a predetermined format, wherein the data generating means expresses an absolute value of 2 with respect to a bit pattern expressed by a plurality of bits. Numerical expression is performed by assigning a numerical value with, and the data with the most significant bit “0” in the numerically expressed data is as it is,
When the most significant bit is “1” in the data represented by the numerical expression, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format. Data generation device. 10. An apparatus for generating data, comprising data generating means for generating data of a predetermined format, wherein the data generating means expresses an absolute value of 2 with respect to a bit pattern expressed by a plurality of bits. The numerical representation is performed by allocating a numerical value with, and the data subjected to the numerical representation is divided into bit blocks consisting of a plurality of consecutive bits, and the most significant bit of some or all of these bit blocks is "0". , The most significant bit of all or part of those bit blocks is “1”.
The data generating device is characterized in that bits other than the most significant bit are inverted and are generated as the data of the predetermined format. 11. An apparatus for generating data, comprising data generating means for generating data of a predetermined format, wherein the data generating means uses a two's complement representation for a bit pattern represented by a plurality of bits. A numerical expression is performed by assigning a numerical value, and the most significant bit of the data that has been numerically expressed is "0" as it is, and the most significant bit of the data that has been numerically expressed is "1".
The data generating device is characterized in that bits other than the most significant bit are inverted and are generated as the data of the predetermined format. 12. An apparatus for generating data, comprising data generating means for generating data of a predetermined format, wherein the data generating means uses a 2's complement representation for a bit pattern represented by a plurality of bits. Numerical expressions are performed by assigning numerical values, and the numerically expressed data is divided into bit blocks consisting of a plurality of consecutive bits, and the most significant bit of some or all of these bit blocks is "0". Some of the bit blocks have the most significant bit of "1" among some or all of the bit blocks, and the bits other than the most significant bit are inverted to generate the data in the predetermined format. A data generation device characterized in that. 13. An apparatus for generating data, comprising data generating means for generating data of a predetermined format, wherein the data generating means expresses an absolute value of 2 for a bit pattern expressed by a plurality of bits. Numerical expression is performed by assigning a numerical value with, and the data in which the most significant bit is “1” in the numerically expressed data is as it is,
When the most significant bit is "0" in the data expressed by the numerical value, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format. Data generation device. 14. A device for generating data, comprising data generating means for generating data of a predetermined format, wherein the data generating means uses a two's complement representation for a bit pattern represented by a plurality of bits. A numerical expression is performed by assigning a numerical value, and the most significant bit of the data that has been numerically expressed is "1", and the most significant bit of the data that has been numerically expressed is "0".
The data generating device is characterized in that bits other than the most significant bit are inverted and are generated as the data of the predetermined format. 15. A program for converting straight binary format data, which performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein A program for causing a computer to execute processing realized as data conversion means for converting data into data of a predetermined format, wherein the data conversion means has the most significant bit of the straight binary format data being “1”. A data conversion program for inverting a bit of data other than the most significant bit of the data. 16. A program for converting straight binary format data for performing numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein the program is of the straight binary format. A program for causing a computer to execute processing realized as data conversion means for converting data into data of a predetermined format, wherein the data conversion means converts the data in the straight binary format into bits formed by a plurality of consecutive bits. It is characterized in that it is divided into blocks, and for some or all of these bit blocks whose most significant bit is "1", bits other than the most significant bit of the bit block are inverted. Data conversion program. 17. A program for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Is a program for causing a computer to execute processing realized as data conversion means for converting the data of the above into data of a predetermined format, wherein the data conversion means is such that the most significant bit of the data of the two's complement format is "1". The data conversion program is characterized in that bits other than the most significant bit of the data are inverted. 18. A program for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Is a program for causing a computer to execute a process realized as a data conversion unit that converts the data of 1 to data of a predetermined format, wherein the data conversion unit converts the data of the 2's complement format,
It is divided into bit blocks made up of a plurality of consecutive bits, and for some or all of these bit blocks whose most significant bit is "1", the bits other than the most significant bit of the bit block are inverted. A data conversion program characterized by 19. A program for converting straight binary format data, which performs numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein the program is of the straight binary format. A program for causing a computer to execute a process realized as data conversion means for converting data into data of a predetermined format, wherein the data conversion means is such that the most significant bit of the straight binary format data is "0". A data conversion program for inverting a bit of data other than the most significant bit of the data. 20. A program for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Is a program for causing a computer to execute processing realized as data conversion means for converting the data of the above format into data of a predetermined format, wherein the data conversion means is such that the most significant bit of the data in the two's complement format is “0”. The data conversion program is characterized in that bits other than the most significant bit of the data are inverted. 21. A program for generating data, which is a program for causing a computer to execute a process realized as data generating means for generating data of a predetermined format, wherein the data generating means has a plurality of bits. A numerical expression is performed by assigning a numerical value with an absolute value expression of 2 to the expressed bit pattern, and the data in which the most significant bit is "0" among the data expressed by the numerical expression is as it is.
When the most significant bit is “1” in the data expressed by the numerical value, the data other than the most significant bit is inverted and is generated as the data of the predetermined format. Data generator to do. 22. A program for generating data, which is a program for causing a computer to execute a process realized as data generating means for generating data of a predetermined format, wherein the data generating means has a plurality of bits. A numerical expression is performed by assigning a numerical value to the expressed bit pattern with an absolute value expression of 2, and the data that has been subjected to the numerical expression is divided into bit blocks consisting of a plurality of consecutive bits, Some or all of the most significant bits are "0", and some or all of the bit blocks have the most significant bit of "1".
Is a data generation program in which bits other than the most significant bit are inverted to generate the data in the predetermined format. 23. A program for generating data, which is a program for causing a computer to execute a process realized as data generating means for generating data of a predetermined format, wherein the data generating means has a plurality of bits. Numerical representation is performed by assigning a numerical value to the represented bit pattern in a two's complement representation, and the numerical representation is performed on the data whose most significant bit is "0" as it is. The most significant bit of the data is "1"
Is a data generation program in which bits other than the most significant bit are inverted to generate the data in the predetermined format. 24. A program for generating data, which is a program for causing a computer to execute a process realized as data generating means for generating data of a predetermined format, wherein the data generating means has a plurality of bits. Numerical representation is performed by assigning a numerical value to the represented bit pattern in a two's complement representation, and the data subjected to the numerical representation is divided into bit blocks made up of a plurality of consecutive bits. Parts or all of which have the most significant bit of "0" as they are, while those of some or all of the bit blocks that have the most significant bit of "1" have the bits other than the most significant bit inverted. Is generated as data of the predetermined format. 25. A program for generating data, which is a program for causing a computer to execute a process realized as data generating means for generating data of a predetermined format, wherein the data generating means has a plurality of bits. A numerical expression is performed by assigning a numerical value to the expressed bit pattern with an absolute value expression of 2, and the data in which the most significant bit is "1" among the data expressed by the numerical expression is as it is.
When the most significant bit is “0” in the data represented by the numerical expression, the bit other than the most significant bit is inverted and is generated as the data of the predetermined format. Data generator to do. 26. A program for generating data, which is a program for causing a computer to execute a process realized as data generating means for generating data of a predetermined format, wherein the data generating means has a plurality of bits. A numerical expression is performed by assigning a numerical value to the expressed bit pattern in a two's complement expression, and the numerical expression is performed as it is for the data whose most significant bit is "1" in the numerically expressed data. The most significant bit of the data is "0"
Is a data generation program in which bits other than the most significant bit are inverted to generate the data in the predetermined format. 27. A method of converting straight binary format data for performing numerical expression by assigning a numerical value by an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein the straight binary format data is converted. A data conversion step of converting the data into data of a predetermined format, wherein the data conversion step is for bits other than the most significant bit of the data for which the most significant bit of the straight binary format data is "1". A data conversion method characterized by inverting. 28. A method of converting straight binary format data for performing numerical expression by assigning a numerical value with an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein the straight binary format data is converted. The method includes a data conversion step of converting data into data of a predetermined format, wherein the data conversion step divides the straight binary format data into bit blocks made up of a plurality of consecutive bits, and a part or all of the bit blocks. A data conversion method, in which the bit other than the most significant bit of the bit block is inverted for the one whose most significant bit is "1". 29. The data read step according to claim 27, further comprising: a data read step of reading the data of the predetermined format from a storage means; and a read step of the data read step, which is a reverse procedure of a conversion procedure of the data conversion step. The data inverse conversion step of converting the data into the straight binary format data, the arithmetic processing step of performing arithmetic processing on the data converted in the data inverse conversion step, and the arithmetic result obtained through the arithmetic processing step A step of storing data in a storage means, wherein the data processed by the operation processing step is converted through the data conversion step and then stored through the data storage step. Data conversion method. 30. A method for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Data conversion step of converting the data of the above format into data of a predetermined format, wherein the data conversion step is for data of which the most significant bit is "1" among the data of the two's complement format except for the most significant bit of the data. A data conversion method characterized by inverting the bits of. 31. A method for converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Data conversion step of converting the data of 1 to data of a predetermined format, wherein the data conversion step divides the data of the 2's complement format into bit blocks made up of a plurality of consecutive bits, and a part of the bit blocks. Alternatively, the data conversion method is characterized by inverting the bits other than the most significant bit of the bit block with respect to all of the most significant bits being “1”. 32. The data reading step according to claim 30 or 31, further comprising: a data reading step of reading the data in the predetermined format from a storage means; and a step of reversing a conversion procedure of the data converting step. A data inverse conversion step of converting the data into the data of the two's complement format, an operation processing step of performing an operation process on the data converted in the data inverse conversion step, and an operation result obtained through the operation processing step. A step of storing data in the storage means, wherein the data processed by the operation processing step is converted through the data conversion step, and then stored through the data storage step. Data conversion method. 33. A method of converting straight binary format data for performing numerical expression by assigning a numerical value by an absolute value expression of 2 to a bit pattern expressed by a plurality of bits, wherein the straight binary format data is converted. A data conversion step of converting the data into data of a predetermined format, wherein the data conversion step is for bits other than the most significant bit of the data for which the most significant bit of the straight binary format data is "0". A data conversion method characterized by inverting. 34. A method of converting data in a two's complement format for performing a numerical expression by assigning a numerical value in a two's complement expression to a bit pattern represented by a plurality of bits, wherein the two's complement format is used. Data conversion step of converting the data of the above into data of a predetermined format, wherein the data conversion step is for data of which the most significant bit is "0" among the data of the two's complement format except for the most significant bit of the data. A data conversion method characterized by inverting the bits of. 35. A method of generating data, comprising a data generating step of generating data of a predetermined format, wherein the data generating step represents an absolute value of 2 for a bit pattern expressed by a plurality of bits. The numerical representation is performed by allocating a numerical value with, and the most significant bit of the data represented by the numerical expression is “0”, and the most significant bit of the data represented by the numerical expression is “1”. A data generating method is characterized in that a bit other than the most significant bit is inverted to generate the data in the predetermined format. 36. A method of generating data, comprising a data generating step of generating data of a predetermined format, wherein the data generating step expresses an absolute value of 2 for a bit pattern expressed by a plurality of bits. The numerical representation is performed by allocating a numerical value with, and the data subjected to the numerical representation is divided into bit blocks consisting of a plurality of consecutive bits, and the most significant bit of some or all of these bit blocks is "0". Of the bit blocks, the most significant bit of all or part of the bit block is "1". The bits other than the most significant bit are inverted to generate as the data of the predetermined format. Characteristic data generation method. 37. A method of generating data, comprising a data generating step of generating data of a predetermined format, wherein the data generating step is a two's complement representation for a bit pattern represented by a plurality of bits. Numerical expression is performed by assigning a numerical value, and the most significant bit of the numerically expressed data is “0”, and the most significant bit of the numerically expressed data is “1”. The data generating method is characterized in that a bit other than the most significant bit is inverted to generate the data of the predetermined format. 38. A method of generating data, comprising a data generating step of generating data of a predetermined format, wherein the data generating step is a two's complement representation for a bit pattern represented by a plurality of bits. Numerical expressions are performed by assigning numerical values, and the numerically expressed data is divided into bit blocks consisting of a plurality of consecutive bits, and the most significant bit of some or all of these bit blocks is "0". Some of the bit blocks have the most significant bit of "1" among some or all of the bit blocks, and the bits other than the most significant bit are inverted to generate the data of the predetermined format. Data generation method. 39. A method of generating data, comprising a data generating step of generating data of a predetermined format, wherein the data generating step expresses an absolute value of 2 for a bit pattern expressed by a plurality of bits. The numerical representation is performed by allocating a numerical value with, and the most significant bit of the data represented by the numerical representation is "1" as it is, and the most significant bit of the data represented by the numerical representation is "0". A data generating method is characterized in that a bit other than the most significant bit is inverted to generate the data in the predetermined format. 40. A method of generating data, comprising a data generating step of generating data of a predetermined format, wherein the data generating step is a two's complement representation for a bit pattern represented by a plurality of bits. A numerical expression is performed by assigning a numerical value, and the most significant bit of the data that has been numerically expressed is "1", and the most significant bit of the data that has been numerically expressed is "0". The data generating method is characterized in that a bit other than the most significant bit is inverted to generate the data of the predetermined format.