Compression in machine learning and deep learning processing

US10546393B2 · kind B2 · utility

Assignee

• Intel Corporation · US

Inventors

• Joydeep Ray · Folsom, US
• Ben J. Ashbaugh · Folsom, US
• Prasoonkumar Surti · Folsom, US
• Pradeep Ramani · Boise, US
• Rama Harihara · Santa Clara, US
• Jerin C. Justin · San Jose, US
• Jing Huang · Chandler, US
• Xiaoming Cui · Santa Clara, US
• Timothy B. Costa · Beaverton, US
• Ting Gong · Beijing, CN
• Elmoustapha Ould-Ahmed-Vall · Chandler, US
• Kumar Balasubramanian · Chandler, US
• Anil Thomas · Redmond, US
• Oguz H. Elibol · Sunnyvale, US
• Jayaram Bobba · Portland, US
• Guozhong Zhuang · Hillsboro, US
• Bhavani Subramanian · Hillsboro, US
• Gokce Keskin · Mountain View, US
• Chandrasekaran Sakthivel · Cupertino, US
• Rajesh Poornachandran · Portland, US

Key dates

• Filing date: Dec 30, 2017
• Grant date: Jan 28, 2020
• Expiry date: Jan 15, 2038

Classification

• Technology area (CPC G): Physics
• CPC primary: G06F2212/401
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

Embodiments are generally directed to compression in machine learning and deep learning processing. An embodiment of an apparatus for compression of untyped data includes a graphical processing unit (GPU) including a data compression pipeline, the data compression pipeline including a data port coupled with one or more shader cores, wherein the data port is to allow transfer of untyped data without format conversion, and a 3D compression/decompression unit to provide for compression of untyped data to be stored to a memory subsystem and decompression of untyped data from the memory subsystem.