TDMMtgMinutesSep2512 < P10761

P10761 Web>TDMMtgMinutesSep2512 (2016-01-24, ErnstChristen) (raw view)EditAttach
-- Main.JoachimHaase - 2012-10-02

---++ IEEE P1076.1 Working Group - Table-Driven Modeling Project - Telco Meeting Minutes of September 25, 2012

*Status:* <em>draft<br /></em>

<br />%TOC%
---+++ Attendees
   * Ernst Christen
   * Joachim Haase (meeting convener, note taker)
---+++ Agenda
   * Call to order
   * Volunteer for minutes
   * Approval of agenda
   * Review and approve minutes from the last meetings
   * Examples and experiences using TDM functionality
   * Decision how to continue
   * Next meeting
   * AOB
   * Adjourn
---+++ Call to order on September 25, 2012 at 8:05 am PDT (17:05 CEST)

Joachim opens the meeting and started to present the meeting slides ([[%PUBURL%/%WEB%/ProjectTableDrivenModeling/1076.1_table_driven_modeling_2012_09_25.ppt][ppt]]).
---+++ Volunteer for minutes

Joachim takes the notes.
---+++ Approval of agenda

Agenda unanimously approved.
---+++ Review and approve minutes from last meeting

The minutes of the May 14 and July 9, 2012 meetings were approved.
---+++ Examples and experiences using TDM functionality

Examples can be found in the slides. The examples base on a VHDL 2002 compatible version of the TDM function. There was a discussion on the subsequent topics
   * There is no advantage of introducing a type REAL_MATRIX instead of a type REAL_VECTOR_ARRAY<br />for the description of data in the multidimensional case (concerns also code based on VHDL 2002).
   * It should pecified in more detail how to handle white-space characters in data files.
   * Character ' ' as SEPARATOR parameter should also bes used for other white space characters (HT, ...).
   * CSV handling should only base on given real values per data line. Function parameters point to columns. Thus, _SELECTROR in associated parameter identifier sholud be replaced by -COLUMN.
   * Only data point information should be extracted from IBIS files. The details must be better specified.
   * Data files might be very hudge (65000 data lines and more). This may lead to inefficiencies if only constrained arrays (as in VHDL 2002) can be used in record declarations.
---+++ Decision on how to continue

The subsequent problems remain
   * Hiding in the package header TABLE_INTERPOLATION_PKG the declaration of 
      * TABLE_DATA_REAL
      * TABLE_DATA_COMPLEX
   * Allow additional declarations only for the reference implementation, e.g.<br />type REAL_VECTOR_ARRAY is array (NATURAL range &lt;&gt;) of REAL_VECTOR;
   * Declaration of function interfaces to handle [TDM-R14} 
      * IBIS files
      * Excel CSV files
   * Final documents with the results of the activities of this subcommittee
---+++ Next meeting

Tuesday, November 11, 2012 at 8am PDT or 17:00 CEST. It must be checked whether this is a suitable date.
---+++ AOB

None.
---+++ Adjourn

The meeting was closed at 18:00 CEST. An invitation will be sent out for the next meeting.
---+++ Action items

Ernst:
   * Check on how to hide record declarations for TABLE_DATA_REAL and TABLE_DATA_COMPLEX in the package header.

Joachim:
   * Proposal for handling white space charcters in files.
   * Specification of PREPARE_DATA_FUNCTIONS for CSV and IBIS files.
Topic revision: r2 - 2016-01-24 - 02:42:56 - ErnstChristen
P10761