"Bit-pragmatic Deep Neural Network Computing" - Information and Links:

Bit-pragmatic Deep Neural Network Computing - Info and Reading Options


“Bit-pragmatic Deep Neural Network Computing” Metadata:

  • Title: ➤  Bit-pragmatic Deep Neural Network Computing
  • Authors:

“Bit-pragmatic Deep Neural Network Computing” Subjects and Themes:

Edition Identifiers:

  • Internet Archive ID: arxiv-1610.06920

AI-generated Review of “Bit-pragmatic Deep Neural Network Computing”:


"Bit-pragmatic Deep Neural Network Computing" Description:

The Internet Archive:

We quantify a source of ineffectual computations when processing the multiplications of the convolutional layers in Deep Neural Networks (DNNs) and propose Pragmatic (PRA), an architecture that exploits it improving performance and energy efficiency. The source of these ineffectual computations is best understood in the context of conventional multipliers which generate internally multiple terms, that is, products of the multiplicand and powers of two, which added together produce the final product [1]. At runtime, many of these terms are zero as they are generated when the multiplicand is combined with the zero-bits of the multiplicator. While conventional bit-parallel multipliers calculate all terms in parallel to reduce individual product latency, PRA calculates only the non-zero terms using a) on-the-fly conversion of the multiplicator representation into an explicit list of powers of two, and b) hybrid bit-parallel multplicand/bit-serial multiplicator processing units. PRA exploits two sources of ineffectual computations: 1) the aforementioned zero product terms which are the result of the lack of explicitness in the multiplicator representation, and 2) the excess in the representation precision used for both multiplicants and multiplicators, e.g., [2]. Measurements demonstrate that for the convolutional layers, a straightforward variant of PRA improves performance by 2.6x over the DaDiaNao (DaDN) accelerator [3] and by 1.4x over STR [4]. Similarly, PRA improves energy efficiency by 28% and 10% on average compared to DaDN and STR. An improved cross lane synchronication scheme boosts performance improvements to 3.1x over DaDN. Finally, Pragmatic benefits persist even with an 8-bit quantized representation [5].

Read “Bit-pragmatic Deep Neural Network Computing”:

Read “Bit-pragmatic Deep Neural Network Computing” by choosing from the options below.

Available Downloads for “Bit-pragmatic Deep Neural Network Computing”:

"Bit-pragmatic Deep Neural Network Computing" is available for download from The Internet Archive in "texts" format, the size of the file-s is: 0.67 Mbs, and the file-s went public at Fri Jun 29 2018.

Legal and Safety Notes

Copyright Disclaimer and Liability Limitation:

A. Automated Content Display
The creation of this page is fully automated. All data, including text, images, and links, is displayed exactly as received from its original source, without any modification, alteration, or verification. We do not claim ownership of, nor assume any responsibility for, the accuracy or legality of this content.

B. Liability Disclaimer for External Content
The files provided below are solely the responsibility of their respective originators. We disclaim any and all liability, whether direct or indirect, for the content, accuracy, legality, or any other aspect of these files. By using this website, you acknowledge that we have no control over, nor endorse, the content hosted by external sources.

C. Inquiries and Disputes
For any inquiries, concerns, or issues related to the content displayed, including potential copyright claims, please contact the original source or provider of the files directly. We are not responsible for resolving any content-related disputes or claims of intellectual property infringement.

D. No Copyright Ownership
We do not claim ownership of any intellectual property contained in the files or data displayed on this website. All copyrights, trademarks, and other intellectual property rights remain the sole property of their respective owners. If you believe that content displayed on this website infringes upon your intellectual property rights, please contact the original content provider directly.

E. Fair Use Notice
Some content displayed on this website may fall under the "fair use" provisions of copyright law for purposes such as commentary, criticism, news reporting, research, or educational purposes. If you believe any content violates fair use guidelines, please reach out directly to the original source of the content for resolution.

Virus Scanning for Your Peace of Mind:

The files provided below have already been scanned for viruses by their original source. However, if you’d like to double-check before downloading, you can easily scan them yourself using the following steps:

How to scan a direct download link for viruses:

  • 1- Copy the direct link to the file you want to download (don’t open it yet).
    • (a free online tool) and paste the direct link into the provided field to start the scan.
  • 2- Visit VirusTotal (a free online tool) and paste the direct link into the provided field to start the scan.
  • 3- VirusTotal will scan the file using multiple antivirus vendors to detect any potential threats.
  • 4- Once the scan confirms the file is safe, you can proceed to download it with confidence and enjoy your content.

Available Downloads

  • Source: Internet Archive
  • All Files are Available: Yes
  • Number of Files: 6
  • Number of Available Files: 6
  • Added Date: 2018-06-29 18:24:26
  • Scanner: Internet Archive Python library 1.7.0.dev2

Available Files:

1- Text PDF

  • File origin: original
  • File Format: Text PDF
  • File Size: 0.00 Mbs
  • File Name: 1610.06920.pdf
  • Direct Link: Click here

2- Metadata

  • File origin: original
  • File Format: Metadata
  • File Size: 0.00 Mbs
  • File Name: 1610.06920_metadata.xml
  • Direct Link: Click here

3- Metadata

  • File origin: original
  • File Format: Metadata
  • File Size: 0.00 Mbs
  • File Name: arxiv-1610.06920_files.xml
  • Direct Link: Click here

4- Metadata

  • File origin: original
  • File Format: Metadata
  • File Size: 0.00 Mbs
  • File Name: arxiv-1610.06920_meta.sqlite
  • Direct Link: Click here

5- Metadata

  • File origin: original
  • File Format: Metadata
  • File Size: 0.00 Mbs
  • File Name: arxiv-1610.06920_meta.xml
  • Direct Link: Click here

6- Archive BitTorrent

  • File origin: metadata
  • File Format: Archive BitTorrent
  • File Size: 0.00 Mbs
  • File Name: arxiv-1610.06920_archive.torrent
  • Direct Link: Click here

Search for “Bit-pragmatic Deep Neural Network Computing” downloads:

Visit our Downloads Search page to see if downloads are available.

Find “Bit-pragmatic Deep Neural Network Computing” in Libraries Near You:

Read or borrow “Bit-pragmatic Deep Neural Network Computing” from your local library.

Buy “Bit-pragmatic Deep Neural Network Computing” online:

Shop for “Bit-pragmatic Deep Neural Network Computing” on popular online marketplaces.