JPH0711652B2 - Light arithmetic - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a method of performing parallel logical operation at high speed using light.

[Prior art and its problems]

Computers that perform high-speed operations to process large-scale information are being researched, but the sequential processing method using electric circuits is approaching the performance limit. So, such as supercomputers and array processors
Research on parallel processing architectures that execute multiple operations simultaneously is in progress. On the other hand, light has a spatial expanse and its physical properties do not interfere with each other, so that operations using light are excellent in parallelism. Amplitude, phase, frequency, polarization, etc. are considered as means for modulating light, and spatial light modulators are being developed.

However, the conventional optical arithmetic methods are not practical because many of them are not suitable for parallel arithmetic or require complicated coding before arithmetic operation.

[Object of the Invention]

An object of the present invention is to provide an optical operation method for changing the light transmittance and executing logical operations in parallel and at high speed.

[Means for solving problems]

The optical arithmetic method of the present invention forms a plurality of input signals and a signal obtained by inverting the plurality of input signals, and a first light source that emits light according to the input signal and a second light source that emits light according to the inverted signal. A plurality of pairs of light sources are arranged to form a light emitting first surface and a light emitting second surface having the same structure as the light emitting first surface, and a light emitting pattern from the light emitting first surface and a light emitting second surface And the light emitted from the first light source and the second light source on the light emitting first surface and the first light source on the light emitting second surface and the second light source.
The first light emitted from the first light source on the first light emitting surface and the second light emitted from the second light source on the first light emitting surface are emitted so as to be adjacent to each other. Emission of the third light emitted from the first light source on the second surface The respective transmittances of the fourth light emitted from the second light source on the second surface are independently changed, and the first, second, and By changing the intensity of the 3rd and 4th lights and combining the 1st, 2nd, 3rd, and 4th lights after the change of the light intensity, the intensity of the output pattern obtained by photoelectric conversion is converted into a signal of 2 It is characterized in that the binarization is performed and logical operations on the plurality of input signals are spatially executed in parallel.

[Principle of Invention]

 The principle of the present invention will be described with reference to FIGS.

FIG. 2 is a diagram showing a relationship between input data and an input surface.
Two-dimensional binary input data 101 (a, b, c, d, e, f) to be calculated and data (,,,
,,) are arranged in an array to form the input surface 102.

Two sets of such input surfaces are prepared, and as shown in FIG.
If one of the input surfaces 103 and 104 of the set is placed in a state of being rotated by 90 ° and they are overlapped using the half mirror 105, the output surface 1
In 07, a pattern in which two sets of input patterns overlap is obtained.

FIG. 4 shows an example of the pattern. As shown in FIG. 4A, the data 108 on the input surface 103 is the upper left, the inverted data 109 is the upper right, the data 110 on the input surface 104 is the lower left, and the inverted data 111 is the lower right. 4 (b), (c),
In (d) and (e), the shaded portions indicate the portions not irradiated with light, and in (b), (c), (d) and (e), the input data (input surface 103, 104) is (0, 0). ), (0,
This corresponds to the cases of 1), (1,0), and (1,1).

At this time, the pattern of the mask 106 is, as shown in FIG. 5A, four first, second, third and fourth patterns 112, 113, 11
It is composed of 4,115, and the transmittance of each light is 0,1 / 4,1 / 2,3 /
It can be changed to 4,1. Since one of the input data 108, 110 and 109, 111 is irradiated, the light intensity after passing through the mask patterns 112, 113, 114, 115 is 100% as the strongest one, as shown in Table 1.
It is represented by 0,50,75,100%. If the transmitted light of one set of output patterns is collected and the light intensity is 100 or 75%, and 0 is set to 0 or 50%, the transmittance of the mask pattern is changed to set two sets of transmitted light. It is possible to execute a logical operation using the data on the input surface as an input.

For example, as shown in FIG. 5B, the mask pattern 11
If the transmittances of 2 and 114 are 1/2 and the transmittances of the mask patterns 113 and 115 are 0, the result of the AND operation is obtained on the output surface. Further, as shown in FIG. 5C, if the transmittance of the mask patterns 112 and 114 is 0 and the transmittance of the mask patterns 113 and 115 is 1/2, the result of the NOR operation can be obtained on the output surface. By setting the partial transmittance of the mask to 0, 1/4, 1/2, 3/4, 1 in this way, 14 kinds of logical operations shown in Table 1 can be executed. The numbers in the pattern column in the table represent the transmittance.
If the four mask patterns shown in FIG. 5 are combined with each input two-dimensional data, logical operation can be executed in parallel.

[Examples] Examples of the present invention will be described below.

FIG. 1 is a perspective view showing an example of an optical arithmetic device for realizing the optical arithmetic method of the present invention. This optical arithmetic unit comprises arrayed light sources 1 and 2 arranged at positions orthogonal to each other and composed of light sources capable of high-speed modulation such as LEDs, and a lens array 3 for collimating light beams emitted from these arrayed light sources. Half mirror 5 that combines collimated light with
And a spatial light modulator 6 such as a liquid crystal TV and the intensity of the pattern after passing through the spatial light modulator, for example, 2
A detector array 7 such as a three-dimensional CCD and an A / D converter 11 for binarizing a signal obtained by this detector array are provided. The arrayed light sources 1 and 2 are controlled by driving devices 8 and 9 each of which is composed of a circuit for applying a voltage to each light source, and the spatial light modulator 6 is controlled by a control device 10.
Controlled by.

In the optical arithmetic device having the above configuration, input data for blinking the arrayed light sources 1 and 2 are controlled by the driving devices 8 and 9. The light beams emitted from the arrayed light sources 1 and 2 are collimated by the lens arrays 3 and 4. The plurality of collimated lights are combined by the half mirror 5 and transmitted through the spatial light modulator 6. The spatial light modulator 6 by the controller 10
A desired operation can be executed by placing the opening of the above at a position determined by the type of logical operation. Spatial light modulator 6
The light transmitted through is received by the detector array 7, digitized by the A / D converter 11, and binarized. As described above, a desired logical operation can be executed.

〔The invention's effect〕

As described above in detail, logical operations can be executed in parallel and at high speed by using the optical operation method of the present invention.

[Brief description of drawings]

FIG. 1 is a perspective view showing an example of an optical arithmetic device used for implementing the present invention, FIG. 2 is a diagram showing a relationship between input data and an input surface, and FIG. 3 is a diagram explaining the principle of the optical arithmetic method. FIG. 4 is a diagram showing an output pattern, and FIG. 5 is a diagram showing a mask pattern. 1,2 …… Array light source 3,4 …… Lens array 5 …… Half mirror 6 …… Spatial light modulator 7 …… Detector array 8,9 …… Driving device 10 …… Control device 11 …… A / D Converter 101 …… Input data 102,103,104 …… Input surface 105 …… Half mirror 106 …… Mask 107 …… Output surface 108 …… Input surface 103 data 109 …… Input surface 103 inverted data 110 …… Input surface 104 data 111 ... Inverted data on the input surface 112 ... First pattern 113 ... Second pattern 114 ... Third pattern 115 ... Fourth pattern

Claims (1)

[Claims] 1. A plurality of first light sources that form a plurality of input signals and signals obtained by inverting the plurality of input signals, and a plurality of second light sources that emit light by the input signals and second light sources that emit light by the inverted signals. The light emitting first surface and the light emitting second surface having the same structure as the light emitting first surface are arranged in pairs to form a light emitting pattern from the light emitting first surface and a light emitting pattern from the light emitting second surface. And light emitted from the first light source and the second light source on the light emitting first surface and the first light source and the second light on the light emitting second surface.
Light emitted from the first light source on the first surface and the second light emitted from the second light source on the first surface are emitted. The respective transmittances of the third light emitted from the first light source of the second surface and the fourth light emitted from the second light source of the light emitting second surface are independently changed to change the first, second, and By changing the intensity of the 3rd and 4th lights and combining the 1st, 2nd, 3rd, and 4th lights after the change of the light intensity, the intensity of the output pattern obtained by photoelectric conversion is converted into a signal of 2 An optical arithmetic method characterized by performing binarization and spatially performing logical operations on the plurality of input signals.