1 / 4

Comments for Standard for Proposed Power Modeling to Enable System Level Analysis PAR

Comments for Standard for Proposed Power Modeling to Enable System Level Analysis PAR. Comments from the Yes Votes (1). Section 5.2 may need to more explicitly state software compiler developers as a stakeholder .

woody
Download Presentation

Comments for Standard for Proposed Power Modeling to Enable System Level Analysis PAR

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Comments for Standard for Proposed Power Modeling to Enable System Level Analysis PAR

  2. Comments from the Yes Votes (1) Section 5.2 may need to more explicitly state software compiler developers as a stakeholder. My vote is “For” with the additional comment that this working group has to be coordinated with the other two related system-level power activities to ensure interoperability. So, now that Nagu has made it clear that this is to be a Meta Standard focussed on parameterization/abstraction and requirements for the extension of other related standards (such as 1801) then I am happy to vote yes. However, should we wind up with 3 separate work groups all working various aspects of System Level Power then we all must continue to work hard to minimise any potential for overlap/divergence. I vote "Yes" with the additional comment:  I feel that the working group for this PAR should not be separate from the working group for the previous PAR (“LPSG should recommend Unified Hardware Abstraction and Layer PAR proposal for Design Automation Standards Committee (DASC) sponsorship”).  My apologies for missing the vote on that PAR, but I intended to vote "Yes" on that one as well, with a similar comment about this PAR.

  3. Comments from the Yes Votes (2) 5. The current draft of the PAR mentions that “This standard would also define the requirements for necessary extensions, as needed, within other related power, workload and functional modeling standards,…”.  After further consideration, I would suggest that this should be simplified slightly.  As currently written, the above statement implies that the working group for the proposed standard would need to determine how each of the other related standards does or does not fulfill the requirements for development of paramaterized, accurate, efficient, and complete power models, in order to identify where extensions are required in each standard.  This would require the working group for the proposed standard to deeply understand all of the other related standards and track their on-going development.  This would seem to be onerous and inefficient.  I believe it would be sufficient for the working group for the proposed standard to focus on defining requirements for the information content of paramaterized, accurate, efficient, and complete power models, without trying to determine what extensions are required in each related standard to represent such information.  Determining the degree to which any given related standard fulfills those requirements, and the definition of any extensions to address unfulfilled requirements, can best be done by the working group responsible for each related standard. Given that, I would suggest that the above-mentioned text be amended to read as follows:  “This standard would also define the requirements for necessary extensions, as needed, the information content of parameterized, accurate, efficient, and complete power models, to help guide development and usage of within other related power, workload and functional modeling standards,…”.  I make the above suggestion as a friendly amendment.  My “Yes” vote is not contingent upon its acceptance; the suggestion is intended only to simplify the task of the working group for this proposed effort and better structure its interaction with other working groups.

  4. Comments from the No Votes I vote “NO” on this PAR. The system level aspects of this are already being addressed well in 1801 and extensions will be covered in the UHA PAR2 which is approved. I vote “no”, as the modeling in this PAR overlaps with 1801, SLP, and the UHA PAR just approved. The proposed PAR does not reflect the initial presentations, research contributions and discussions during the multi-month LPSG activities. It especially lost its initial valid direction during the writing of the PAR text as it faced the initial comments.  In its current form this PAR is substantially overlapping with the ongoing UPF System Level Power (SLP)  effort and the new PAR for the Unified Hardware Abstraction and Layer. The concerning overlaps are on the system and software power abstractions and in its most recent version on the modeling of the power management. As this PAR now seems to be freely evolving we may experience additional overlaps with established IEEE standards, like those in the ESL or IP proliferation. This may lead to unnecessary confusion among the key industry contributors, future users and IEEE standardization decision makers and as such could cause a long term stall of all IEEE efforts aimed at power and energy. My suggestion is to first review this PAR from the UPF SLP perspective to avoid overlaps and confusion. Once this is cleared and these two efforts are aligned, most of the potential overlaps with the new PAR Unified Hardware Abstraction and Layer we proposed shall be automatically removed and the remaining ones simply resolved.  I personally highly value the technical contribution from Nagu and his team during the multi-month LPSG activity, but due to the above concerns my vote is no.

More Related