Saturday, August 18, 2018
Search
  
Submit your own News for
inclusion in our Site.
Click here...
Breaking News
Amazon to Release Live TV Recorder
U.S. Wants to Wiretap Facebook Messenger, Report Says
More Affordable, 13-inch MacBook and $160 AirPower Expected at Apple's Event
Google to Launch Its Own Smart Display, Report Says
Nvidia Reports Record Revenue From Datacenter, Gaming, Professional Visualization, Automotive
Arm Client CPU Roadmap Includes Advanced Hercules and Deimos Chips
Google To Release Lightweight Version of Android Pie for Entry-level Smartphones
Japan Accuses Apple of Pressuring Game Rivals: Nikkei
Active Discussions
Which of these DVD media are the best, most durable?
How to back up a PS2 DL game
Copy a protected DVD?
roxio issues with xp pro
Help make DVDInfoPro better with dvdinfomantis!!!
menu making
Optiarc AD-7260S review
cdrw trouble
 Home > News > General Computing > Intel N...
Last 7 Days News : SU MO TU WE TH FR SA All News

Thursday, May 24, 2018
Intel Nervana NNP-L1000 Neural Network Processor Coming in 2019


Intel provided updates on its newest family of Intel Nervana Neural Network Processors (NNPs) at Intel AI DevCon, the company's inaugural AI developer conference.

The Intel Nervana NNP has an explicit design goal to achieve high compute utilization and support true model parallelism with multichip interconnects.

Intel is building toward the first commercial NNP product offering, the Intel Nervana NNP-L1000 (Spring Crest), in 2019. The company anticipates the Intel Nervana NNP-L1000 to achieve 3-4 times the training performance of its first-generation Lake Crest product. Intel will also support bfloat16, a numerical format being adopted industrywide for neural networks, in the Intel Nervana NNP-L1000. Over time, Intel will be extending bfloat16 support across its AI product lines, including Intel Xeon processors and Intel FPGAs.

The company showed initial performance benchmarks on Intel's NNP family. These results come from the prototype of the Intel Nervana NNP (Lake Crest) from which the company is gathering feedback from its early partners:

  • General Matrix to Matrix Multiplication (GEMM) operations using A(1536, 2048) and B(2048, 1536) matrix sizes have achieved more than 96.4 percent compute utilization on a single chip. This represents around 38 TOP/s of actual (not theoretical) performance on a single chip. Multichip distributed GEMM operations that support model parallel training are realizing nearly linear scaling and 96.2 percent scaling efficiency for A(6144, 2048) and B(2048, 1536) matrix sizes - enabling multiple NNPs to be connected together and freeing from memory constraints of other architectures.
  • Intel is measuring 89.4 percent of unidirectional chip-to-chip efficiency of theoretical bandwidth at less than 790ns (nanoseconds) of latency and are excited to apply this to the 2.4Tb/s (terabits per second) of high bandwidth, low-latency interconnects.

All of this is happening within a single chip total power envelope of under 210 watts.



Previous
Next
Qualcomm Snapdragon 710 Mobile Platform Brings Artificial Intelligence Features to a New Tier of Smartphones        All News        Uber Ends Arizona Self-driving Program
Uber's Self-driving Recognized Pedestrian Late and Failed to Brake: NTSB     General Computing News      Uber Ends Arizona Self-driving Program

Get RSS feed Easy Print E-Mail this Message

Related News
Intel Introduces New NUC Kits and NUC Mini PCs
Intel Discloses New Chip Security Flaws
Intel Sets the Stage for the Persistent Memory Revolution and FPGA Acceleration
Intel Promotes Optane And QLC Solutions, World's Densest 'Ruler' SSD
Intel Outlines 'Data-Centric' Strategy, Shows off Xeon Roadmap
Intel Intros 660p Series M.2 NVMe SSDs with QLC NAND Flash
Broadcom to Design 7-nm AI processor For Wave: report
Intel Claims Progress on 10nm Yields But Pushes back Release Schedule
DARPA to Work With Nvidia, Qualcomm, IBM and Intel on Post-Moore's-law Projects
Micron and Intel to End Their 3D XPoint Joint Development Partnership
Microsoft Calls for Public Regulation of AI Face Recognition Software
Intel to Buy eASIC to Expand Its Portfolio Beyond Processors

Most Popular News
 
Home | News | All News | Reviews | Articles | Guides | Download | Expert Area | Forum | Site Info
Site best viewed at 1024x768+ - CDRINFO.COM 1998-2018 - All rights reserved -
Privacy policy - Contact Us .