CN111427539A - Random data stream computing system and computing control method based on siganmin - Google Patents
Random data stream computing system and computing control method based on siganmin Download PDFInfo
- Publication number
- CN111427539A CN111427539A CN202010200108.0A CN202010200108A CN111427539A CN 111427539 A CN111427539 A CN 111427539A CN 202010200108 A CN202010200108 A CN 202010200108A CN 111427539 A CN111427539 A CN 111427539A
- Authority
- CN
- China
- Prior art keywords
- random
- data
- module
- random number
- data stream
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F7/00—Methods or arrangements for processing data by operating upon the order or content of the data handled
- G06F7/58—Random or pseudo-random number generators
- G06F7/588—Random number generators, i.e. based on natural stochastic processes
Abstract
The invention provides a random data stream computing system and a computing control method based on a siganmin, wherein the system comprises: the random number generation module, the data storage module and the random operation module are sequentially connected; the random number generation module generates a true random data stream according to input data; the data storage module is used for storing the true random data stream and/or storing the operation result output by the random operation module; the random operation module performs logic operation on the data in the data storage module and sends an operation result to the data storage module and/or outputs the operation result; at least one of the random number generation module, the data storage module and the random operation module comprises the skullet, so that the size of a random data flow calculation system based on the skullet is smaller, and the power consumption and the area of the system are reduced.
Description
Technical Field
The invention relates to the technical field of spin electronics, in particular to a random data stream computing system and a computing control method based on a siganmin.
Background
At present, the application of big data such as internet of things and artificial intelligence is emerging, the computing performance of an end-side intelligent scene on hardware is higher and higher, especially when facing a large amount of matrix operations, in order to complete matrix multiply-add operations, a conventional solution is to use a L U (arithmetric and logic unit) including logical operation units such as multipliers and adders, however, as the data volume increases and the network scale expands, the number of a L U to be used is more and more, so that the area occupied by a data processing module in a chip is too large, and effective configuration cannot be obtained on end-side hardware facing to the intelligent internet of things (AIoT) with limited computing resources.
The random calculation is that a random factor is added on the basis of the traditional accurate calculation, namely, the next operation of a random number selection algorithm is carried out. An essential feature of random computation is that a number can be represented by a bit stream processed by very simple circuitry, and the number itself is interpreted as a probability, i.e. the probability that each bit in the string of bits is 1, as in conventional floating point 7/10, can be represented by a string 10101111011 of binary numbers. According to bernoulli's law of numbers, the probability can be estimated in terms of frequency, i.e., the probability of 1 per bit can be expressed in terms of the fraction of 1's in the bitstream in the string. Random computation recodes information in a probability domain, and converts the information into a series of 01 sequences, and the ratio of the number of 1 in the sequence in the whole sequence represents the probability of channel information.
At present, the principle of random computing hardware in the field of integrated circuits is mostly process variation or thermal disturbance occurring in CMOS-based processes. With the further shrinking of the process size, the power consumption and area ratio of the system increase and the area cannot be further reduced due to the random computing hardware based on the CMOS.
Disclosure of Invention
The invention provides a random data stream computing system and a computing control method based on a siganmin, which can at least partially solve the problems in the prior art.
In order to achieve the purpose, the invention adopts the following technical scheme:
in a first aspect, a skynerger-based random data stream computing system is provided, comprising: the random number generation module, the data storage module and the random operation module are sequentially connected; the random number generation module generates a true random data stream according to input data; the data storage module is used for storing the true random data stream and/or storing the operation result output by the random operation module; the random operation module performs logic operation on the data in the data storage module and sends an operation result to the data storage module and/or outputs the operation result;
at least one of the random number generation module, the data storage module and the random operation module comprises a skyrmion.
Further, the random number generation module includes: the system comprises a Sgimenk random number generator, a comparator and an input cache unit;
the input buffer unit is used for buffering the input data, the Sjg random number generator is used for generating random numbers, and the comparator compares the input data with the random numbers to generate the true random data stream.
Further, the skynting random number generator comprises: from the bottom up according to heavy metal layer, ferromagnetic free layer, tunnel layer and the ferromagnetic pinned layer that sets up, this tunnel layer and this ferromagnetic pinned layer form two magnetic tunnel junctions including the same two parts of size for the magnetic resistance in the perception ferromagnetic free layer changes.
Further, the data storage module includes: a siglec sub-track memory.
Further, the sightline sub-track memory comprises: a plurality of tracks, each track comprising: the system comprises a Sgermin writing end, a Sgermin reading end, a nano track, a data input end and a data output end;
the skyrmion writing end generates skyrmions, the skyrmion reading end detects the skyrmions, the nano-track is used for completing distribution and movement of a plurality of skyrmions, the data input end receives data to be stored, and the data output end reads the stored data.
Further, the random operation module is a boolean logic operation unit, including: a combination of one or more of a sigmin and gate, a sigmin or gate, a sigmin nor gate.
Further, the random operation module comprises: a sigmin adder and/or a sigmin multiplier.
Further, the skynerger-based random data stream computing system further comprises: and the control module is connected with the random number generation module, the data storage module and the random operation module.
In a second aspect, a random data stream calculation control method based on a sigramin is provided, which includes:
transmitting the input data to an input cache unit to finish data caching;
transmitting the input data buffered in the input buffer unit to a comparator, and controlling a Skeleton random number generator to generate a random number and transmitting the random number to the comparator so that the comparator compares the input data with the random number to generate a true random data stream;
transmitting the true random data stream to a data storage module for data storage;
transmitting the data in the data storage module to a random operation module for logic operation;
and the operation result output by the random operation module is sent to the data storage module and/or is output to an external microprocessor.
The invention provides a random data stream computing system and a computing control method based on a siganus, wherein the system comprises: the random number generation module, the data storage module and the random operation module are sequentially connected; the random number generation module generates a true random data stream according to input data; the data storage module is used for storing the true random data stream and/or storing the operation result output by the random operation module; the random operation module performs logic operation on the data in the data storage module and sends an operation result to the data storage module and/or outputs the operation result; at least one of the random number generation module, the data storage module and the random operation module comprises a skyrmion. The skutters have topological discontinuity in energy, so that the skutters have higher stability and lower system power consumption and area, and are easy to move and not influenced by lattice pinning, so that the size of a random data flow computing system based on the skutters is smaller, and the system power consumption and area are reduced.
In order to make the aforementioned and other objects, features and advantages of the invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts. In the drawings:
FIG. 1 is a block diagram of a random data flow computing system based on Stegimen according to an embodiment of the present invention;
FIG. 2 is a block diagram of a random data flow computing system based on the Scramble according to an embodiment of the present invention;
FIG. 3 is a block diagram illustrating a random number generation module in a Scotch-based random data stream computing system provided by an embodiment of the present invention;
FIG. 4 is a circuit block diagram of a random number generation module in a random data stream Scotch-based computing system provided by an embodiment of the present invention;
FIG. 5 is a block diagram of the skamming random number generator of FIG. 4;
FIG. 6 is a circuit schematic diagram of the Skeleton random number generator of FIG. 4;
FIG. 7 illustrates a top view of the skamming random number generator of FIG. 4;
FIG. 8 is a block diagram illustrating a data storage module in a Scotch-based random data flow computing system provided by an embodiment of the present invention;
FIG. 9 is a schematic diagram of the operation of the skybook memory of FIG. 8;
FIG. 10 is a block diagram illustrating a random operation module in a random Stegming based data stream computing system according to an embodiment of the present invention;
FIG. 11 shows a schematic diagram of a SgimenK AND gate in an embodiment of the present invention;
FIG. 12 illustrates a schematic diagram of a Schlemm OR gate in an embodiment of the present invention;
fig. 13 (a) shows a schematic diagram of a logical and, logical or operation device based on the skynman;
fig. 13 (b) shows a timing chart of a logical and, logical or operation device based on the sgrminal;
fig. 14 shows a flowchart of a random data stream calculation control method based on the skynerger in the embodiment of the present invention.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The detailed features and advantages of the present invention are described in detail in the following embodiments, which are sufficient for anyone skilled in the art to understand the technical content of the present invention and to implement the present invention, and the related objects and advantages of the present invention can be easily understood by anyone skilled in the art from the disclosure, the claims and the drawings of the present specification. The following examples further illustrate aspects of the present invention in detail, but are not intended to limit the scope of the invention in any way.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
Fig. 1 is a first block diagram of a random data flow computing system based on a skynman according to an embodiment of the present invention. As shown in fig. 1, the skammomum-based random data stream computing system includes: the random number generating module 100, the data storage module 200 and the random operation module 300 are connected in sequence.
Wherein, the random number generation module 100 generates a true random data stream according to the input data; the data storage module 200 is used for storing true random data streams and/or storing operation results output by the random operation module; the random operation module 300 performs logical operation on the data in the data storage module, and sends the operation result to the data storage module 200 and/or outputs the operation result;
at least one of the random number generation module 100, the data storage module 200 and the random operation module 300 includes a skyrmion.
Specifically, at least one or two or all of the generation of the random number, the saving of the random data stream and the logical operation of the random data are performed by the skynolve. For example, the random number generation module 100 uses a thermal perturbation based true random number generator to generate random numbers using the brownian motion of the siganus, and/or a nano-multilayer film based random logic device to implement a basic boolean logic function using the motion of the siganus in the magnetic multilayer film.
It should be noted that the random operation module 300 performs logic (and, or, not, etc.) operation on two or more random data streams in the data storage module according to corresponding operation rules, and delivers the operation result to the data storage module for storage, or directly outputs the operation result without passing through the data storage module.
According to the random data flow computing system based on the sGermin, which is provided by the embodiment of the invention, the sGermin has topological discontinuity on energy, so that the system has higher stability and lower system power consumption and area, and the sGermin is easy to move and is not influenced by lattice pinning, so that the size of the random data flow computing system based on the sGermin is smaller, the system power consumption and area are reduced, and a new hardware support is provided for the deployment of random computing on end-side equipment.
In an alternative embodiment, referring to fig. 2, the skullet-based random data stream computing system may further include: and the control module 400, wherein the control module 400 is connected with the random number generation module 100, the data storage module 200 and the random number operation module 300, and is used for controlling the modules to coordinate and cooperate.
FIG. 3 is a block diagram illustrating a random number generation module in a Scotch-based random data stream computing system provided by an embodiment of the present invention; as shown in fig. 3, the random number generation module includes a skamming random number generator, and under the control of the control module, generates a true random number using input binary data, and outputs the true random number as a random data stream.
The Schlemn random number generator is adopted, the thermal disturbance-based true random number generator is adopted, and Brownian motion of the Schlemn is utilized to generate random numbers, so that the characteristics of small size, high stability and low power consumption of the Schlemn random number generator can be effectively utilized, and the power consumption and the area of a system are reduced.
In an alternative embodiment, referring to fig. 4, the random number generation module 100 may specifically include: the system comprises a Sgimenk random number generator, a comparator and an input cache unit;
the input buffer unit is used for buffering input data (generally binary data), the Sjgren random number generator is used for generating random numbers, and the comparator compares the input data with the random numbers to generate the true random data stream. The input and output of the random number generation module 100 are controlled by a control module.
Specifically, under the control of the control module, the random number generation module 100 may complete the generation and output of the n-bit random data stream. The binary data is input from outside, firstly enters the input buffer unit, enters the x end of the comparator when reaching n bits, and simultaneously enters the y end of the comparator through the n-bit random number generated by the Sgmuion random number generator. The two data are compared in a comparator, and if y > x, the output is 1, y < x, and the output is 0. Of course, it can be understood by those skilled in the art that the output of the comparator is only an exemplary illustration, and it may also be that if y > x, the output is 0, y < x, and the output is 1, which is not limited in this embodiment of the present invention. The clock control is used for controlling the Schlemn random number generator, so that the time of inputting the buffered data into the comparator by the input buffer unit is synchronous with the time of inputting the random number into the comparator by the Schlemn random number generator.
By adopting the effective cooperation of the random number generator of the sjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj.
In an alternative embodiment, referring to fig. 5, the sgmm random number generator is composed of four layers of films, from bottom to top, a Heavy metal layer (Heavy metal layer)1, a Ferromagnetic free layer (Ferromagnetic layer)2, a tunneling layer (oxide barrier)3, and a Ferromagnetic pinned layer (pinned layer)4, wherein the tunneling layer 3 and the Ferromagnetic pinned layer 4 above are divided into two parts with the same size to form two MTJ1 and MTJ2 for sensing the magnetoresistance change in the Ferromagnetic free layer. Since the skulls are generated in the ferromagnetic free layer and make random brownian motion, two different voltages Vleft and Vright are generated at upper MTJ1 and MTJ 2. Referring to fig. 6, by comparing the magnitudes of the two voltages, a random set of data stream bits can be generated, i.e., if Vleft > Vright, then 1 is output, otherwise 0 is output.
FIG. 7 illustrates a top view of the Squalmin random number generator of FIG. 4, with reference to its dimensional parameters, it is clear that the Squalmin random number generator employed in embodiments of the present invention is of a smaller size.
FIG. 8 is a block diagram illustrating a data storage module in a Scotch-based random data flow computing system provided by an embodiment of the present invention; as shown in fig. 8, the random number storage module 200 may include a skullet sub-racetrack memory, and the random data stream is written and read under the control of the control module.
The skyscraper racetrack memory is adopted, and the skyscraper is used for realizing data storage, so that the characteristics of small size, high stability and low power consumption of the skyscraper memory can be effectively utilized, and the power consumption and the area of a system are reduced.
In an alternative embodiment, the storyboard memory comprises: a plurality of tracks, each track comprising: the system comprises a Sgermin writing end, a Sgermin reading end, a nano track, a data input end and a data output end;
the skyrmion writing end generates skyrmions, the skyrmion reading end detects the skyrmions, the nano-track is used for completing distribution and movement of the multiple skyrmions, the data input end receives data to be stored, and the data output end reads the stored data. Referring to fig. 9, after the generation of the sigrons, whether to move or not is determined according to the value of the input random bit stream, and if the input is high, the motion is performed forward, and if the input is low, the motion is kept still. Thus, the input 0 and 1 can be stored by reading the relative position of the skullet on the nano-track, and the data can be read by reading the skullet.
FIG. 10 is a block diagram illustrating a random operation module in a random Stegming based data stream computing system according to an embodiment of the present invention; as shown in fig. 10, the random operation module may include a skmingming sub-logic operation unit, and under the control of the control module, the writing of the multiple random data streams (1-N) and the output of the calculation result are completed.
Specifically, the random operation module is a boolean logic operation unit, which includes logic gates, and may include: a combination of one or more of a sigmin and gate, a sigmin or gate, a sigmin nor gate.
It should be noted that the random operation module may include basic arithmetic operation units, such as a skming sub-adder and/or a skming sub-multiplier.
The skullet sub-logic operation unit is adopted, and skullet sub-logic operation is realized, so that the characteristics of small size, high stability and low power consumption of the skullet sub-logic operation unit can be effectively utilized, and the power consumption and the area of the system are reduced.
FIG. 11 shows a schematic diagram of a SgimenK AND gate in an embodiment of the present invention; as shown in FIG. 11, two random data streams S are input1And S2The multiplication operation is realized through an AND gate to obtain a product S3。
FIG. 12 illustrates a schematic diagram of a Schlemm OR gate in an embodiment of the present invention; as shown in FIG. 12, two random data streams S are input1And S2Through an OR gate, an OR operation is realized to obtain S3。
Fig. 13 (a) shows a schematic diagram of a logical and, logical or operation device based on the skynman; fig. 13 (b) shows a timing chart of a logical and, logical or operation device based on the sgrminal; referring to fig. 13 (a) and 13 (b), an embodiment of a device for implementing the sggming sub-logic and or operation is shown, wherein the device has a three-layer structure, which includes a heavy metal layer, a ferromagnetic free layer and an electrode (V) from bottom to topg1And Vg2). Wherein, the heavy metal layer (not shown in the figure) is used for driving the flow of current, the ferromagnetic free layer is used for the motion of the siganus, and the electrode is used for regulating and controlling the motion trail of the siganus. Two input terminals (In) on the left side of the device1And In2) When the siganus are input, the two output ends on the right side are both output by the siganus through current driving and electrode voltage regulation. If there is only one sgemin input, the sgemin is finally output from the Or logic terminal (Or _ out) And no output is output from the And logic terminal (And _ out) due to the squeeze of the device edge. If there is no Sgimenk input, there is no output at both output ends.In conclusion, the device is capable of performing an AND or OR operation.
The embodiment of the invention also provides a random data stream calculation control method based on the skullet, which is used for controlling the random data stream calculation based on the skullet and can be executed by a control module in a random data stream calculation system based on the skullet. Referring to fig. 14, the random data stream calculation control method based on the skamming may include the steps of:
step S100: transmitting the input data to an input cache unit to finish data caching;
step S200: transmitting the input data buffered in the input buffer unit to a comparator, and controlling a Skeleton random number generator to generate a random number and transmitting the random number to the comparator so that the comparator compares the input data with the random number to generate a true random data stream;
step S300: transmitting the true random data stream to a data storage module for data storage;
step S400: transmitting the data in the data storage module to a random number operation module for logic operation;
step S500: and sending the operation result output by the random number operation module to the data storage module and/or outputting the operation result to an external microprocessor.
By adopting the random data flow calculation control method based on the skullet, the random data flow calculation system based on the skullet provided by the embodiment of the invention can be effectively controlled, the skullet is effectively utilized to realize random data flow calculation, the skullet has higher stability and lower system power consumption and area due to topological discontinuity on energy, and the skullet is easy to move and is not influenced by lattice pinning, so that the size of random data flow calculation hardware based on the skullet is smaller, the system power consumption and area are reduced, and new hardware support is provided for the deployment of random calculation on end-side equipment.
The principle and the implementation mode of the invention are explained by applying specific embodiments in the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
Although the present invention has been described with reference to the preferred embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but may be embodied or carried out by various modifications, equivalents and changes without departing from the spirit and scope of the invention.
Claims (9)
1. A sigecum-based random data stream computing system, comprising: the random number generation module, the data storage module and the random operation module are sequentially connected; the random number generation module generates a true random data stream according to input data; the data storage module is used for storing the true random data stream and/or storing the operation result output by the random operation module; the random operation module performs logic operation on the data in the data storage module and sends an operation result to the data storage module and/or outputs the operation result;
wherein at least one of the random number generation module, the data storage module and the random operation module comprises a skyrmion.
2. The skyrmion-based random data stream computing system of claim 1, wherein the random number generation module comprises: the system comprises a Sgimenk random number generator, a comparator and an input cache unit;
the input buffer unit is used for buffering the input data, the Sjgren sub random number generator is used for generating random numbers, and the comparator compares the input data with the random numbers to generate the true random data stream.
3. A siganus-based random data stream computing system according to claim 2, wherein the siganus random number generator comprises: heavy metal layer, ferromagnetic free layer, tunneling layer and the ferromagnetic fixed layer that from the bottom up set up according to this, the tunneling layer and the ferromagnetic fixed layer forms two magnetic tunnel junctions including two parts that the size is the same for the magnetic resistance in the perception ferromagnetic free layer changes.
4. The skynet-based random data stream computing system of claim 1, wherein the data storage module comprises: a siglec sub-track memory.
5. The skyburn-based random data stream computing system of claim 4, wherein the skyburn racetrack memory comprises: a plurality of tracks, each track comprising: the system comprises a Sgermin writing end, a Sgermin reading end, a nano track, a data input end and a data output end;
the skyrmion writing end generates skyrmions, the skyrmion reading end detects the skyrmions, the nano-track is used for completing distribution and movement of the multiple skyrmions, the data input end receives data to be stored, and the data output end reads the stored data.
6. The skynet-based random data stream computing system of claim 1, wherein the random operation module is a boolean logic operation unit comprising: a combination of one or more of a sigmin and gate, a sigmin or gate, and a sigmin not gate.
7. The skybook-based random data stream computing system of claim 6, wherein the random operation module comprises: a sigmin adder and/or a sigmin multiplier.
8. The skybook-based random data stream computing system of claim 1, further comprising: and the control module is connected with the random number generation module, the data storage module and the random operation module.
9. A random data flow calculation control method based on a Sgimenk is characterized by comprising the following steps:
transmitting the input data to an input cache unit to finish data caching;
transmitting the input data buffered in the input buffer unit to a comparator, and controlling a Skeleton random number generator to generate a random number and transmitting the random number to the comparator so that the comparator compares the input data with the random number to generate a true random data stream;
transmitting the true random data stream to a data storage module for data storage;
transmitting the data in the data storage module to a random operation module for logic operation;
and sending the operation result output by the random operation module to the data storage module and/or outputting the operation result to an external microprocessor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010200108.0A CN111427539B (en) | 2020-03-20 | 2020-03-20 | Random data stream computing system and computing control method based on siganmin |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010200108.0A CN111427539B (en) | 2020-03-20 | 2020-03-20 | Random data stream computing system and computing control method based on siganmin |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111427539A true CN111427539A (en) | 2020-07-17 |
| CN111427539B CN111427539B (en) | 2022-12-23 |
Family
ID=71549663
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010200108.0A Active CN111427539B (en) | 2020-03-20 | 2020-03-20 | Random data stream computing system and computing control method based on siganmin |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111427539B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112002687A (en) * | 2020-08-03 | 2020-11-27 | 中国计量大学 | Device and method for continuously controlling movement and pinning of siganus |
| CN112835556A (en) * | 2021-01-28 | 2021-05-25 | 广东省大湾区集成电路与系统应用研究院 | MTJ-based true random number generator |
| WO2023178596A1 (en) * | 2022-03-21 | 2023-09-28 | 中电海康集团有限公司 | Storage and computing device, counter, shift accumulator, and in-memory multiply-accumulate structure |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106877858A (en) * | 2016-12-27 | 2017-06-20 | 南京大学 | A kind of logic gates based on magnetic Skyrmion |
| CN108492845A (en) * | 2018-04-03 | 2018-09-04 | 电子科技大学 | A kind of racing track memory based on magnetic Skyrmion |
-
2020
- 2020-03-20 CN CN202010200108.0A patent/CN111427539B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106877858A (en) * | 2016-12-27 | 2017-06-20 | 南京大学 | A kind of logic gates based on magnetic Skyrmion |
| CN108492845A (en) * | 2018-04-03 | 2018-09-04 | 电子科技大学 | A kind of racing track memory based on magnetic Skyrmion |
Non-Patent Citations (2)
| Title |
|---|
| NIKHIL RANGARAJAN 等: "A spin-based true random number generator exploiting the stochastic precessional switching of nanomagnets", 《JOURNAL OF APPLIED PHYSICS》 * |
| WILLIAM A. BORDERS 等: "Integer factorization using stochastic magnetic tunnel junctions", 《NATURE》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112002687A (en) * | 2020-08-03 | 2020-11-27 | 中国计量大学 | Device and method for continuously controlling movement and pinning of siganus |
| CN112002687B (en) * | 2020-08-03 | 2024-04-26 | 中国计量大学 | Device and method for continuously controlling movement and pinning of stigmata |
| CN112835556A (en) * | 2021-01-28 | 2021-05-25 | 广东省大湾区集成电路与系统应用研究院 | MTJ-based true random number generator |
| CN112835556B (en) * | 2021-01-28 | 2024-01-16 | 广东省大湾区集成电路与系统应用研究院 | True random number generator based on MTJ |
| WO2023178596A1 (en) * | 2022-03-21 | 2023-09-28 | 中电海康集团有限公司 | Storage and computing device, counter, shift accumulator, and in-memory multiply-accumulate structure |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111427539B (en) | 2022-12-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111427539B (en) | Random data stream computing system and computing control method based on siganmin | |
| CN108171317B (en) | Data multiplexing convolution neural network accelerator based on SOC | |
| Walus et al. | RAM design using quantum-dot cellular automata | |
| Eshraghian et al. | Memristive device fundamentals and modeling: Applications to circuits and systems simulation | |
| JP2005235307A (en) | Logic-in-memory circuit using magneto-resistance effect element | |
| Sengupta et al. | Spin-transfer torque magnetic neuron for low power neuromorphic computing | |
| Patil et al. | Spintronic logic gates for spintronic data using magnetic tunnel junctions | |
| Xiao et al. | Energy and performance benchmarking of a domain wall-magnetic tunnel junction multibit adder | |
| Natsui et al. | Design of MTJ-based nonvolatile logic gates for quantized neural networks | |
| CN111768806B (en) | Magnetic skynet based memory device and method for storing information by using same | |
| CN111406326B (en) | Magnetic memory structure and device | |
| Nisar et al. | Hybrid Multilevel STT/DSHE Memory for Efficient CNN Training | |
| JP6817922B2 (en) | Arithmetic logic unit | |
| CN117651953A (en) | Hybrid machine learning architecture with neural processing unit and in-memory computing processing | |
| KR20240036594A (en) | Subsum management and reconfigurable systolic flow architectures for in-memory computation | |
| CN117751346A (en) | Folded column adder architecture for digital computation in memory | |
| CN113451505B (en) | Magnetic random access memory cell, memory and device | |
| KR20180089701A (en) | Magnetic tunneling junction memory device with in-memory processing structure | |
| CN112508174A (en) | Pre-calculation column-by-column convolution calculation unit for weight binary neural network | |
| Gulafshan et al. | Power Efficient Image Processing with TMR Tunable Hybrid Approximate Adders | |
| Yadav et al. | Impact of edge defects on the synaptic characteristic of a ferromagnetic domain-wall device and on on-chip learning | |
| Patil et al. | Error-resilient spintronics via the Shannon-inspired model of computation | |
| Hoe | Bayesian inference using spintronic technology: A proposal for an MRAM-based stochastic logic gate | |
| Suzuki et al. | Design framework for an energy-efficient binary convolutional neural network accelerator based on nonvolatile logic | |
| Kim et al. | Low energy domain wall memory based convolution neural network design with optimizing MAC architecture |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |