Business Wire

Toshiba Memory Corporation Develops High-Speed and High-Energy-Efficiency Algorithm and Hardware Architecture for Deep Learning Processor

Jaa

Toshiba Memory Corporation, the world leader in memory solutions, today announced the development of a high-speed and high-energy-efficiency algorithm and hardware architecture for deep learning processing with less degradations of recognition accuracy. The new processor for deep learning implemented on an FPGA [1] achieves 4 times energy efficiency compared to conventional ones. The advance was announced at IEEE Asian Solid-State Circuits Conference 2018 (A-SSCC 2018) in Taiwan on November 6.

Deep learning calculations generally require large amounts of multiply-accumulate (MAC) operations, and it has resulted in issues of long calculation time and large energy consumption. Although techniques reducing the number of bits to represent parameters (bit precision) have been proposed to reduce the total calculation amount, one of proposed algorithm reduces the bit precision down to one or two bit, those techniques cause degraded recognition accuracy. Toshiba Memory developed the new algorithm reducing MAC operations by optimizing the bit precision of MAC operations for individual filters [2] in each layer of a neural network. By using the new algorithm, the MAC operations can be reduced with less degradation of recognition accuracy.

Furthermore, Toshiba Memory developed a new hardware architecture, called bit-parallel method, which is suitable for MAC operations with different bit precision. This method divides each various bit precision into a bit one by one and can execute 1-bit operation in numerous MAC units in parallel. It significantly improves utilization efficiency of the MAC units in the processor compared to conventional MAC architectures that execute in series.

Toshiba Memory implemented ResNet50[3], a deep neural network, on an FPGA using the various bit precision and bit-parallel MAC architecture. In the case of image recognition for the image dataset of ImageNet[4], the above technique supports that both operation time and energy consumption for recognizing image data are reduced to 25 % with less recognition accuracy degradation, compared to conventional method.

Artificial intelligence (AI) is forecasted to be implemented in various devices. The developed high-speed and low-energy-consumption techniques for deep-learning processors are expected to be utilized for various edge devices like smartphones and HMDs[5] and datacenters which require low energy consumption. High-performance processors like GPU are important devices for high-speed operation of AI. Memories and storages are also one of the most important devices for AI which inevitably use big data. Toshiba Memory Corporation is continuously focusing on research and development of AI technologies as well as innovating memories and storages to lead data-oriented computing.

[1] FPGA: Field Programmable Gate Array, an integrated circuit designed to be configured by a customer or a designer after manufacturing.
[2] filter: Generally, there are many filters of up to several thousands in one layer of a neural network.
[3] ResNet50: One of deep neural network, generally used to benchmark deep-learning for image recognition.
[4] ImageNet: A large image database, generally used to benchmark image-recognition, the number of image data is over 14,000,000.
[5] HMD: Head Mounted Display

About Toshiba Memory Corporation

Toshiba Memory Corporation, a world leader in memory solutions, is dedicated to the development, production and sales of flash memory and SSDs. In June 2018, Toshiba Memory was acquired by an industry consortium led by Bain Capital. Toshiba Memory pioneers cutting-edge memory solutions and services that enrich people's lives and expand society's horizons. The company's innovative 3D flash memory technology, BiCS FLASH™, is shaping the future of storage in high density applications including advanced smartphones, PCs, SSDs, automotive and data centers. For more information on Toshiba Memory, please visit https://business.toshiba-memory.com/en-apac/top.html

Contact information

Toshiba Memory Corporation
Kota Yamaji, +81-3-3457-3473
Business Planning Division
semicon-NR-mailbox@ml.toshiba.co.jp

Tietoja julkaisijasta

For more than 50 years, Business Wire has been the global leader in press release distribution and regulatory disclosure.

Tilaa tiedotteet sähköpostiisi

Haluatko tietää asioista ensimmäisten joukossa? Kun tilaat mediatiedotteemme, saat ne sähköpostiisi välittömästi julkaisuhetkellä. Tilauksen voit halutessasi perua milloin tahansa.

Lue lisää julkaisijalta Business Wire

Warm welcome for ABB Formula E in Davos19.1.2019 12:00Tiedote

Just a few days before the start of the World Economic Forum 2019, Davos is celebrating a world premiere: The latest ABB FIA Formula E racing car - dubbed Gen2 – was transported on a train from the iconic Swiss Rhaetian Railway to Europe’s highest city, where it will be an exciting eyecatcher during the coming week at the Davos Platz railway station. This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20190119005003/en/ ABB FIA Formula E racing car in Davos (Photo: Business Wire) From this Saturday until next Sunday, January 27, the Gen2 will be on display in a special glass case. Interested passersby can learn about the ABB FIA Formula E racing series as well as about the activities of ABB in e-mobility via an information system. With over 8,500 fast charging stations installed in 70 countries around the world, ABB is the global market leader for charging infrastructures. To watch the video, click here ABB (ABBN: SIX Swiss Ex) is

PMI’s Mission Winnow Goes Full Throttle with Ducati Corse for 2019 MotoGP™18.1.2019 20:00Tiedote

Philip Morris International Inc. (PMI) (NYSE: PM) is pleased to announce that its Mission Winnow initiative is expanding in 2019 to include the Ducati Corse racing team. Mission Winnow is a PMI-led campaign to raise global awareness of our passion and determination to constantly improve and evolve, as well as highlight the power of science, technology and innovation to build a better future. PMI has partnered with Ducati Corse since 2002 and has extended the relationship for another three years until the end of 2021. As of the start of this year’s MotoGP season, the team will be officially known as Mission Winnow Ducati. Mission Winnow was first launched in October 2018 with Scuderia Ferrari Mission Winnow, which – like Ducati Corse – has a passion for innovation and a relentless drive to improve in the team’s pursuit of victory. Ducati Corse is one of the most inspiring and resilient teams in MotoGP, with a 70-year history in racing. Ducati fans form a passionate community who appreci

IFF’s Frutarom Division Completes Acquisition of 60% of Thailand-based Mighty18.1.2019 15:50Tiedote

Regulatory News: International Flavors & Fragrances Inc. (NYSE:IFF) (Euronext Paris:IFF) (TASE:IFF), announced that its Frutarom Division has completed the acquisition of 60% of the share capital of The Mighty CO. LTD. (“Mighty”), a leading savory solutions provider in Thailand, thus continuing its growth strategy in Southeast Asia. Amos Anatot, President of IFF's Frutarom Division, said, “The completion of this deal with Mighty underscores that the Frutarom division will continue on its growth strategy and pursue attractive companies that create new opportunities or build on current capabilities.” Mr. Anatot continued, “And in this case, we are helping to grow our capabilities in savory solutions – already an area of strength for legacy Frutarom, now IFF." Mighty, founded in 1989, develops, produces and markets reaction flavors, with particular expertise in savory solutions. The company’s portfolio includes flavors, seasoning blends, marinades, and specialty functional raw materials f

LTI Q3 FY19: Constant currency revenue growth up 6.1% QoQ and 20.6% YoY; Net Profit jumps 32.8% YoY18.1.2019 15:02Tiedote

Larsen & Toubro Infotech (BSE code: 540005, NSE: LTI), a global technology consulting and digital solutions company, announced its Q3 FY19 results today. In US Dollars: Revenue at USD 346.9 million; growth of 5.6% QoQ and 18.2% YoY Constant Currency Revenue growth of 6.1% QoQ and 20.6% YoY In Indian Rupees: Revenue at Rs 24,729 million; growth of 6.1% QoQ and 31.3% YoY Net Income at Rs 3,755 million; growth of (6.2%) QoQ and 32.8% YoY “We are pleased to deliver another strong quarter with 5.6% QoQ growth in USD revenues. Our broad-based revenue growth, superior margin delivery and steady cash generation in Q3 is a testimony of our focused execution and client centricity. We are also thrilled to welcome Ruletronics to LTI family. Ruletronics enables businesses to transform and evolve digitally by providing innovative BPM and CRM solutions leveraging Pega Platform.” - Sanjay Jalona, Chief Executive Officer & Managing Director, LTI Recent Deal Wins Nets, the leading payments company in th

Schlumberger Announces Full-Year and Fourth-Quarter 2018 Results18.1.2019 15:00Tiedote

Schlumberger Limited (NYSE: SLB) today reported results for full-year 2018 and the fourth quarter of 2018. (Stated in millions, except per share amounts) Full-Year Results Twelve Months Ended Change Dec. 31, 2018 Dec. 31, 2017 Year-on-year Revenue $32,815 $30,440 8% Pretax operating income $4,187 $3,921 7% Pretax operating margin 12.8% 12.9% -12 bps Net income (loss) - GAAP basis $2,138 $(1,505) n/m Net income, excluding charges & credits* $2,261 $2,085 8% Diluted EPS (loss per share) - GAAP basis $1.53 $(1.08) n/m Diluted EPS, excluding charges and credits* $1.62 $1.50 8% Full-Year Consolidated Revenue by Area North America $11,984 $9,487 26% Latin America 3,745 3,976 -6% Europe/CIS/Africa 7,158 7,072 1% Middle East & Asia 9,543 9,394 2% Other 385 511 n/m $32,815 $30,440 8% North America revenue $11,984 $9,487 26% International revenue $20,446 $20,442 - North America revenue, excluding Cameron $9,668 $7,518 29% International revenue, excluding Cameron $17,675 $17,423 1% *These are non

FLIR Systems Awarded $89 Million Contract from French Armed Forces to Deliver Black Hornet Personal Reconnaissance System18.1.2019 14:00Tiedote

FLIR Systems, Inc. (NASDAQ: FLIR) announced today it has been awarded a contract from the French Defense Procurement Agency (DGA) in support of the French Operational Pocket Drone (DrOP) program. The contract has a ceiling value of $89 million to provide the FLIR Black Hornet® 3 nano-unmanned aerial vehicle (UAV) and Personal Reconnaissance System (PRS) to support French Armed Forces operations. This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20190118005085/en/ The French Armed Forces awarded FLIR Systems a contract to deliver the Black Hornet Personal Reconnaissance System for military operations. (Photo: Business Wire) The Black Hornet PRS is the world’s smallest combat-proven nano-Unmanned Aerial System (UAS) and is currently deployed in more than 30 countries. The Black Hornet enables the warfighter to maintain situational awareness, threat detection, and surveillance no matter where the mission takes them. Equipped with el

Uutishuoneessa voit lukea tiedotteitamme ja muuta julkaisemaamme materiaalia. Löydät sieltä niin yhteyshenkilöidemme tiedot kuin vapaasti julkaistavissa olevia kuvia ja videoita. Uutishuoneessa voit nähdä myös sosiaalisen median sisältöjä. Kaikki STT Infossa julkaistu materiaali on vapaasti median käytettävissä.

Tutustu uutishuoneeseemme