Time-Domain Multiply-Accumulator using Digital-to-Time Multiplier for CNN Processors in 28-nm CMOS
| العنوان: | Time-Domain Multiply-Accumulator using Digital-to-Time Multiplier for CNN Processors in 28-nm CMOS |
|---|---|
| المؤلفون: | Teerachot Siriburanon, Xutong Wu, Robert Bogdan Staszewski |
| المصدر: | 2020 31st Irish Signals and Systems Conference (ISSC). |
| بيانات النشر: | IEEE, 2020. |
| سنة النشر: | 2020 |
| مصطلحات موضوعية: | 010302 applied physics, Computer science, business.industry, Computation, 020208 electrical & electronic engineering, Sign bit, Schematic, 02 engineering and technology, Computational resource, 01 natural sciences, Convolutional neural network, CMOS, 0103 physical sciences, 0202 electrical engineering, electronic engineering, information engineering, Multiplier (economics), Time domain, business, Computer hardware |
| الوصف: | This paper proposes a time-domain multiply-accumulator (MAC) circuit exploiting the use of a digital-to-time multiplier (DTM) unit for performing convolutional operations and designed in 28nm CMOS. As the foundational calculation of state-of-the-art convolutional neural network (CNN) models, convolutional operation could normally be executed millions of times in one CNN task. The proposed circuit is designed to support this large computational resource requirement during the CNN computations. It is running in time-domain with 6-bit resolution (1 sign bit) and performs calculations based on corresponding time delays. Compared with other analog-domain propositions, time-domain designs perform better, with higher operating frequencies up to 50 MHz. In schematic simulations, the proposed DTM unit, operating with 5 bits, achieves 0.1 GOPS ideal throughput and consumes 74.79 µW at 1.0V supply. |
| DOI: | 10.1109/issc49989.2020.9180202 |
| URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_________::6cebdd2bcae13e0fa9c477e3093b7de3 https://doi.org/10.1109/issc49989.2020.9180202 |
| Rights: | CLOSED |
| رقم الانضمام: | edsair.doi...........6cebdd2bcae13e0fa9c477e3093b7de3 |
| قاعدة البيانات: | OpenAIRE |
| ResultId |
1 |
|---|---|
| Header |
edsair OpenAIRE edsair.doi...........6cebdd2bcae13e0fa9c477e3093b7de3 845 3 unknown 844.682861328125 |
| PLink |
https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&scope=site&db=edsair&AN=edsair.doi...........6cebdd2bcae13e0fa9c477e3093b7de3&custid=s6537998&authtype=sso |
| FullText |
Array
(
[Availability] => 0
)
Array ( [0] => Array ( [Url] => https://explore.openaire.eu/search/publication?articleId=doi_________::6cebdd2bcae13e0fa9c477e3093b7de3# [Name] => EDS - OpenAIRE [Category] => fullText [Text] => View record in OpenAIRE [MouseOverText] => View record in OpenAIRE ) ) |
| Items |
Array
(
[Name] => Title
[Label] => Title
[Group] => Ti
[Data] => Time-Domain Multiply-Accumulator using Digital-to-Time Multiplier for CNN Processors in 28-nm CMOS
)
Array ( [Name] => Author [Label] => Authors [Group] => Au [Data] => <searchLink fieldCode="AR" term="%22Teerachot+Siriburanon%22">Teerachot Siriburanon</searchLink><br /><searchLink fieldCode="AR" term="%22Xutong+Wu%22">Xutong Wu</searchLink><br /><searchLink fieldCode="AR" term="%22Robert+Bogdan+Staszewski%22">Robert Bogdan Staszewski</searchLink> ) Array ( [Name] => TitleSource [Label] => Source [Group] => Src [Data] => <i>2020 31st Irish Signals and Systems Conference (ISSC)</i>. ) Array ( [Name] => Publisher [Label] => Publisher Information [Group] => PubInfo [Data] => IEEE, 2020. ) Array ( [Name] => DatePubCY [Label] => Publication Year [Group] => Date [Data] => 2020 ) Array ( [Name] => Subject [Label] => Subject Terms [Group] => Su [Data] => <searchLink fieldCode="DE" term="%22010302+applied+physics%22">010302 applied physics</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+science%22">Computer science</searchLink><br /><searchLink fieldCode="DE" term="%22business%2Eindustry%22">business.industry</searchLink><br /><searchLink fieldCode="DE" term="%22Computation%22">Computation</searchLink><br /><searchLink fieldCode="DE" term="%22020208+electrical+%26+electronic+engineering%22">020208 electrical & electronic engineering</searchLink><br /><searchLink fieldCode="DE" term="%22Sign+bit%22">Sign bit</searchLink><br /><searchLink fieldCode="DE" term="%22Schematic%22">Schematic</searchLink><br /><searchLink fieldCode="DE" term="%2202+engineering+and+technology%22">02 engineering and technology</searchLink><br /><searchLink fieldCode="DE" term="%22Computational+resource%22">Computational resource</searchLink><br /><searchLink fieldCode="DE" term="%2201+natural+sciences%22">01 natural sciences</searchLink><br /><searchLink fieldCode="DE" term="%22Convolutional+neural+network%22">Convolutional neural network</searchLink><br /><searchLink fieldCode="DE" term="%22CMOS%22">CMOS</searchLink><br /><searchLink fieldCode="DE" term="%220103+physical+sciences%22">0103 physical sciences</searchLink><br /><searchLink fieldCode="DE" term="%220202+electrical+engineering%2C+electronic+engineering%2C+information+engineering%22">0202 electrical engineering, electronic engineering, information engineering</searchLink><br /><searchLink fieldCode="DE" term="%22Multiplier+%28economics%29%22">Multiplier (economics)</searchLink><br /><searchLink fieldCode="DE" term="%22Time+domain%22">Time domain</searchLink><br /><searchLink fieldCode="DE" term="%22business%22">business</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+hardware%22">Computer hardware</searchLink> ) Array ( [Name] => Abstract [Label] => Description [Group] => Ab [Data] => This paper proposes a time-domain multiply-accumulator (MAC) circuit exploiting the use of a digital-to-time multiplier (DTM) unit for performing convolutional operations and designed in 28nm CMOS. As the foundational calculation of state-of-the-art convolutional neural network (CNN) models, convolutional operation could normally be executed millions of times in one CNN task. The proposed circuit is designed to support this large computational resource requirement during the CNN computations. It is running in time-domain with 6-bit resolution (1 sign bit) and performs calculations based on corresponding time delays. Compared with other analog-domain propositions, time-domain designs perform better, with higher operating frequencies up to 50 MHz. In schematic simulations, the proposed DTM unit, operating with 5 bits, achieves 0.1 GOPS ideal throughput and consumes 74.79 µW at 1.0V supply. ) Array ( [Name] => DOI [Label] => DOI [Group] => ID [Data] => 10.1109/issc49989.2020.9180202 ) Array ( [Name] => URL [Label] => Access URL [Group] => URL [Data] => <link linkTarget="URL" linkTerm="https://explore.openaire.eu/search/publication?articleId=doi_________::6cebdd2bcae13e0fa9c477e3093b7de3" linkWindow="_blank">https://explore.openaire.eu/search/publication?articleId=doi_________::6cebdd2bcae13e0fa9c477e3093b7de3</link><br /><link linkTarget="URL" linkTerm="https://doi.org/10.1109/issc49989.2020.9180202" linkWindow="_blank">https://doi.org/10.1109/issc49989.2020.9180202</link> ) Array ( [Name] => Copyright [Label] => Rights [Group] => Cpyrght [Data] => CLOSED ) Array ( [Name] => AN [Label] => Accession Number [Group] => ID [Data] => edsair.doi...........6cebdd2bcae13e0fa9c477e3093b7de3 ) |
| RecordInfo |
Array
(
[BibEntity] => Array
(
[Identifiers] => Array
(
[0] => Array
(
[Type] => doi
[Value] => 10.1109/issc49989.2020.9180202
)
)
[Languages] => Array
(
[0] => Array
(
[Text] => Undetermined
)
)
[Subjects] => Array
(
[0] => Array
(
[SubjectFull] => 010302 applied physics
[Type] => general
)
[1] => Array
(
[SubjectFull] => Computer science
[Type] => general
)
[2] => Array
(
[SubjectFull] => business.industry
[Type] => general
)
[3] => Array
(
[SubjectFull] => Computation
[Type] => general
)
[4] => Array
(
[SubjectFull] => 020208 electrical & electronic engineering
[Type] => general
)
[5] => Array
(
[SubjectFull] => Sign bit
[Type] => general
)
[6] => Array
(
[SubjectFull] => Schematic
[Type] => general
)
[7] => Array
(
[SubjectFull] => 02 engineering and technology
[Type] => general
)
[8] => Array
(
[SubjectFull] => Computational resource
[Type] => general
)
[9] => Array
(
[SubjectFull] => 01 natural sciences
[Type] => general
)
[10] => Array
(
[SubjectFull] => Convolutional neural network
[Type] => general
)
[11] => Array
(
[SubjectFull] => CMOS
[Type] => general
)
[12] => Array
(
[SubjectFull] => 0103 physical sciences
[Type] => general
)
[13] => Array
(
[SubjectFull] => 0202 electrical engineering, electronic engineering, information engineering
[Type] => general
)
[14] => Array
(
[SubjectFull] => Multiplier (economics)
[Type] => general
)
[15] => Array
(
[SubjectFull] => Time domain
[Type] => general
)
[16] => Array
(
[SubjectFull] => business
[Type] => general
)
[17] => Array
(
[SubjectFull] => Computer hardware
[Type] => general
)
)
[Titles] => Array
(
[0] => Array
(
[TitleFull] => Time-Domain Multiply-Accumulator using Digital-to-Time Multiplier for CNN Processors in 28-nm CMOS
[Type] => main
)
)
)
[BibRelationships] => Array
(
[HasContributorRelationships] => Array
(
[0] => Array
(
[PersonEntity] => Array
(
[Name] => Array
(
[NameFull] => Teerachot Siriburanon
)
)
)
[1] => Array
(
[PersonEntity] => Array
(
[Name] => Array
(
[NameFull] => Xutong Wu
)
)
)
[2] => Array
(
[PersonEntity] => Array
(
[Name] => Array
(
[NameFull] => Robert Bogdan Staszewski
)
)
)
)
[IsPartOfRelationships] => Array
(
[0] => Array
(
[BibEntity] => Array
(
[Dates] => Array
(
[0] => Array
(
[D] => 01
[M] => 06
[Type] => published
[Y] => 2020
)
)
[Identifiers] => Array
(
[0] => Array
(
[Type] => issn-locals
[Value] => edsair
)
)
[Titles] => Array
(
[0] => Array
(
[TitleFull] => 2020 31st Irish Signals and Systems Conference (ISSC)
[Type] => main
)
)
)
)
)
)
)
|
| IllustrationInfo |