Peer to Patent, Community Patent Review

APPLICATION LIST|ARCHIVED APPLICATIONS|US PATENT CLASSIFICATIONS

Home > Archived Applications > 20070118696 > Discussion

Pre-Grant Publication Number: 20070118696

Collaborate on the process of community review for this application. Posting will not be forwarded to the USPTO. Flagging a post as an ACTION ITEM signals further research. Flagging SPAM and ABUSE helps to manage discussion. Placing double brackets around a reference to a claim or prior art will create a hyperlink to the original ex. [[claim 1]] and [[prior art 2]].

Please review the Community Code of Conduct prior to posting

Discussion (15)

View: Date|Claims|Prior Art|Action Item|

Facilitator's Comment

Action Item

Show without Noise

Juliana Agon (about 1 year ago)

Thanks for the 2 pieces of prior art already loaded. Keep it coming. Also keep in mind that these discussions (including patent numbers mentioned and synonyms) do not get forwarded to the USPTO, only the actual prior art links and the annotations on that art. So please convert discussion into the submission of prior art with claims entered as links to the prior art. To enter a claim as a link, enclose it in brackets like this: Claim 1.
This will also make it easier for the reader to understand the discussion of that claim and to rate it.

Matthew Zehrer (about 1 year ago)

I wanted to let the community know that “backward slices” are at least related if not a synonym to “pre-computational slices”. It looks like in one of the first patents utilizing the term “p-slice” (US 6928645) it was taught that, “a suitable example for constructing pre-computation slices is described in C. Zilles and G. Sohi. Understanding the Backward Slices of Performance Degrading Instructions, In Proc. 27th International Symposium on Computer Architecture, pages 172 – 181, June 2000.” I plan on posting this seminal article as prior art.

Juliana Agon (about 1 year ago)

Thanks for the slicing information and synonyms, including prior patent numbers, and for your future posting of the seminal article. Once the article is posted teaching the general idea of slices, then more details of claim 1 (and so on) of the "Register tracking for speculative prefecting" patent application can be compared to the article and all patent numbers already listed to check if all the details are already known. Otherwise, other prior art or synonyms of related areas can be suggested for the different details.

Juliana Agon (about 1 year ago)

Finding an actual copy of an article, publication or other pieces of concrete information in these various simmilar areas of work is the hardest to do. So please upload any such material if possible.

Keith O'Neil (about 1 year ago)

Please advise about copyright issues if material is uploaded. Would not want to be in the position of violating copyright laws, so the URL given points directly to the document.

Brian Pyne (about 1 year ago)

Regarding the material you discussed and linked to in your prior art submission, the posting of any materials freely available from uspto.gov will not trigger any copyright or intellectual property issues. As for posting materials that come from other sources, it is safer to describe and link to the source of materials you are unsure about instead of including excerpts in your prior art submissions.

That having been said, if you think it would be particularly helpful to include a portion of content that you think may be protected, the Digital Millennium Copyright Act (DMCA) requires that the contents owner contact Peer to Patent and give us the opportunity to remove any content on the site that infringes on their rights as owner of the content.

For more information, see the relevant sections of the DMCA and the discussion of intellectual property in Peer to Patent's Terms of Use.

http://dotank.nyls.edu/communitypatent/DMCA.html http://dotank.nyls.edu/communitypatent/Terms_of_Use.html

Andrew Oram (about 1 year ago)

Vector processors, which go back over two decades, may be a good source of prior art for address pre-fetching.

Brian Pyne (about 1 year ago)

Can you find an example of vector processors which can be submitted as prior art for this patent?

Juliana Agon (about 1 year ago)

Good. The inventor's own company's closest art (patents with the same company or same inventor, and research, application or other articles) is a good place to find closest prior art. Other software engineers' contribution in similar work could also yield prior art.

Keith O'Neil (about 1 year ago)

It appear that much work have been done in this area. In my view the following paper is worth reading
because it discusses some aspect of tracking that then allows speculative precomputation. See it at:
http://www.intel.com/research/mrl/library/148_collins_j.pdf

Stacy Olivas (about 1 year ago)

Here are some patents that I found which sound like they have some things in common. I have not had the time to completely review this application, so I am posting them here in hopes that someone else might be able to do more research:

Interpreter for stack-based languages
United States Patent 6205578
Issued March 20, 2001
Inventor Daniel D. Grove
http://www.patentstorm.us/patents/6205578.html

Method and apparatus employing lookahead to reduce memory bank contention for decoupled operand references
United States Patent 5323489
Issued March 20, 2001
Inventor Daniel D. Grove
http://www.patentstorm.us/patents/5323489.html

Anyways, just my ideas on a place to start.

Brian Pyne (about 1 year ago)

Go ahead and submit these two items as prior art so that other members of the community can comment and vote on them. Keep in mind that discussion posts do not get forwarded to the USPTO and only prior art submissions can help the USPTO decide whether an application will be granted a patent.

Juliana Agon (about 1 year ago)

This is great feedback. Yes, try to think of similar concepts that are basically the same idea but different words or names. Prefecting similar to buffering similar to slices, etc. etc. to come up with the closest concept to the claims.

C Tips (about 1 year ago)

There are several ideas that are "well-known"
- using a second program (helper threads) in a multi-threaded processor to warm-up caches
- having hardware that runs ahead of the main thread to get data (decoupled access execute)
- using stride based prefetching
- automatically loading data when an address is stored into a register (CDC6600!)
I would focus on the how the ART is used to generate a p-slice; the section from "figure 2 illustrates..." to "figure 3 illustrate

Rahan Uddin (about 1 year ago)

Hello All - here's a brief description of this invention, share your thoughts, and MOST IMPORTANTLY, upload references relevant to the claims (prior art/submit prior art link):

technique described in this patent application, enables numerous types of prefetch access patterns used by a processor to prefetch data and/or instructions, so as to help increase the number of prefetched data and/or instructions that are actually used by the processor during the course of program execution, while redu

0 days left