IEEE 754 Revision Work

IEEE 754-1985 and 854-1987 are overdue for revision. Previous revisions have passed them unchanged, warts and all. The explosion of cross-platform reliability issues calls for a thorough re-examination.

The working group wants to emphasize reproducible results to ease implementation, testing, and porting. We have already agreed to the following:

Revision Scope and Purpose

Scope

This standard specifies formats and methods for floating-point arithmetic in computer programming environments: standard and extended functions with single, double, extended, and extendable precision, and recommends formats for data interchange. Exception conditions are defined and default handling of these conditions is specified.

Purpose

This standard provides a discipline for performing floating-point computation that yields results independent of whether the processing is done in hardware, software, or a combination of the two. For operations specified in this standard, numerical results and exceptions are uniquely determined by the values of the input data, sequence of operations, and destination formats, all under programmer control.

How to Participate

Participants are expected to have a solid knowledge of floating-point arithmetic. In particular, the group currently seeks feedback from hardware and language implementors. Meetings are open and participation is free of charge.

There is a mailing list associated with the revision: stds-754@ieee.org. To join the list, send the message text "subscribe stds-754" to listserv@listserv.ieee.org. Most work and debate occurs at physical meetings, but feedback and discussions on the mailing list provide crucial input. The list has a manually updated archive.

Outstanding Revision Issues

An incomplete list of issues tackled:

Resources

David Hough has previous minutes and some proposals available. There is also a page on Wikipedia discussing the revision effort.

There is a private working group area. To gain access, join the mailing list and actively participate.

Future Meeting Schedule

Conference calls may be available for remote participation. Please check with the mailing list.

Latest meeting style

Strict parliamentary procedure.

No more public meetings scheduled. Draft is in balloting.

Past Meetings

8 January, 2001
1-5 PM, Network Appliance, Craftsman conference room, Bldg 2
minutes
12 February, 2001
11 AM - 3 PM, Lana Networks
minutes
14 March, 2001
1-5 PM, Network Appliance, Santa Cruz conference room, Bldg 3
minutes
11 April, 2001
1-5 PM, Network Appliance, Craftsman conference room, Bldg 2
minutes
16 May, 2001
3-7 PM, UC Berkeley, Soda Hall
minutes
20 June, 2001
1-5 PM, Network Appliance, Craftsman conference room, Bldg 2
minutes
18 July, 2001
1-5 PM, Network Appliance, Santa Cruz conference room, Bldg 3
minutes
With a presentation by Jim Thomas on C99's support. The slides are available, and subsequent discussion is outlined in the minutes.
15 August, 2001
3-7 PM, 405 Soda Hall, UC Berkeley (map)
minutes
With a presentation on higher precision arithmetic by David Bailey. Slides are available in PostScript and PDF.
13 September, 2001
1-5 PM, Conference Room 101, Intel SC4 (map)
Cancelled
18 October, 2001
1-5 PM, Network Appliance, Craftsman conference room, Bldg 2
minutes
With a presentation by Joe Darcy on language issues. The slides are available, and subsequent discussion is outlined in the minutes.
15 November, 2001
1-5 PM, Network Appliance, Madagasgar conference room, Bldg 3
minutes
13 December, 2001
10AM-3PM PDT, Network Appliance, Craftsman conference room, Bldg 2.
minutes
17 January, 2002
1-5PM, Intel, Santa Clara 12, 538
minutes
21 February, 2002
1-5PM, Network Appliance, Craftsman conference room, Bldg 2
minutes
21 March, 2002
1-5 PM, Intel, Santa Clara 12, 520
minutes
Slides from Jim Thomas on compatibility issues are available. (Also in ppt.)
18 April, 2002
1PM-5PM Hewlett-Packard Cupertino minutes
Highlights: A presentation from Mike Cowlishaw of IBM on encodings for decimal arithmetic
23 May, 2002
1-5PM, Network Appliance, Craftsman conference room, Bldg 2 (map)
minutes
19 June, 2002
1-5 PM, Network Appliance, Santa Cruz conference room, Bldg 3
minutes
18 July, 2002
1-5 PM, Hewlett-Packard Cupertino
minutes
Presentation on formal verification by Marc Daumas
22 August, 2002
Hewlett-Packard, Palo Alto, Bldg. 3, Sigma Room
minutes
Presentation by David Bindel and Jason Riedy on reacting to exceptional situations.
19 September, 2002
1-5PM, Hewlett-Packard, Palo Alto Bldg. 3, Sigma Room
minutes
17 October, 2002
1-5PM, Sun, Menlo Park, Pasatiempo Converence Room
minutes
21 November, 2002
1PM-5PM Hewlett-Packard Cupertino
minutes
17 December, 2002
1-5PM, Sun
minutes
23 January, 2003
1-5PM, Sun Santa Clara
20 February, 2003
Intel, Santa Clara 12, 152
minutes
20 March, 2003
1-5PM, Hewlett Packard Palo Alto
minutes
17 April, 2003
1-5PM, Hewlett Packard Cupertino
29 and 30 May, 2003
1-5PM, 380 Soda Hall, UC Berkeley
Get ready for not one but two days of rowdy decimal fun!
First day's minutes.
Second day's minutes.
10 July, 2003
Sun, Menlo Park MPK14
13 and 14 August, 2003
HP, Cupertino
notes
17 and 18 September, 2003
Sun, Menlo Park, room 1431, MPK16.
15 and 16 October, 2003
HP, Palo Alto, Bldg. 3, Sigma Room
notes
11 and 12 November, 2003
HP, Cupertino
Ivan Godard: Concurrency Issues (pdf)
notes
10 and 11 December, 2003
Intel, Santa Clara 12 Room SC12-537
notes
14 and 15 January, 2004
HP Cupertino
Eric Schwarz: info on Quantize (pdf)
18 and 19 February, 2004
HP, Palo Alto
Notes
17 and 18 March, 2004
Sun, Santa Clara
Notes, more notes, draft minutes
14 and 15 April, 2004
Arithmatica
Notes, more notes
12 and 13 May, 2004
Intel, Santa Clara
16-17 June, 2004
1-5PM, HP Cupertino. Notes
21-22 July, 2004
1-5PM, Sun, Santa Clara. Draft review notes, meeting notes
18-19 August, 2004
1-5PM, Sun, Menlo Park. Draft review notes, meeting notes
15-16 September, 2004
12-5PM, 1-6PM HP, Palo Alto
13 and 14 October, 2004
1-5PM PDT, HP, Cupertino. Meeting notes.
17 and 18 November, 2004
1-5PM, Sun, Menlo Park. Meeting notes.
15 and 16 December, 2004
1-5PM, Intel, Santa Clara. Meeting notes.
19-20 January, 2005
1-5PM, Sun, Santa Clara. Meeting notes.
16-17 February, 2005
1-5PM, HP, Palo Alto
draft review notes, meeting notes
16-17 March, 2005
HP, Cupertino
meeting notes
20-21 April, 2005
Sun, Menlo Park, (attempt at an agenda, notes)
Eric Schwarz and Mike Cowlishaw's slides on decimal formats,
Peter Tang's slides on decimal conversion on the XScale,
Peter Tang and John Harrison's slides on table-free decimal conversion,
John Harrison and Peter Tang's slides on formats
11-12 May, 2005
Sun Microsystems, Menlo Park 1-5PM (directions). Meeting notes
Roger Golliver's slides on min and max
15-16 June, 2005
HP, Cupertino 1-5PM (directions). Meeting notes.
27-29 June, 2005
Some participants will be attending Arith 17.
13-14 July, 2005
HP, Palo Alto 1-5PM (directions)
(notes for the draft review)
17-18 August, 2005
1-5 PM, Sun, Menlo Park (directions, agenda, notes)
14-15 September, 2005
1-5PM, HP (directions)
19-20 October, 2005
1-5PM, HP
16-17 November, 2005
1-5PM, Sun
15 December, 2005
1-5PM, Sun (agenda, directions)
notes
18 January, 2006
1-5PM at HP Cupertino (directions, agenda)
notes
15 February, 2006
1-5PM at Sun Menlo Park (directions)
notes
15-16 March, 2006
1-5PM at Sun Santa Clara (directions)
minutes
19-20 April, 2006
1-5PM at Intel Santa Clara (agenda, directions)
minutes
17-18 May, 2006
1-5PM at Intel Santa Clara (directions)
minutes
21-22 June, 2006
1-5PM at HP Cupertino (directions)
minutes
19-20 July, 2006
1-5PM at Intel Santa Clara (directions)
minutes
9-10 August, 2006
1-5PM at Apple, Cupertino (directions)
minutes
7 September, 2006
1-5PM at HP Cupertino (directions
(unapproved) minues)
20 September, 2006
1-5PM at Sun, Menlo Park (directions, agenda)
2 November, 2006
1-5PM at Apple; special meeting, presentations from Jean-Michel Muller ("CR-LIBM, and Arenaire's results on function implementation") & Paul Zimmermann ("Worst cases for sin(BIG)"), (directions)

754 | revision | FAQ | references