Ieee arithmetic - Revision history

Manus at 22:25, 9 February 2010

2010-02-09T22:25:09Z

← Older revision		Revision as of 22:25, 9 February 2010
Line 5:		Line 5:
	To best show the trade-offs we will start by showing some benchmark results.		To best show the trade-offs we will start by showing some benchmark results.

	{{Note\|This is an EiffelStudio specific extension to potentially help the ECMA standard committee in making a reasonable decision on the semantic of '''NaN'''.}}		{{Note\|This is an EiffelStudio specific extension to potentially help the ECMA standard committee in making a reasonable decision on the semantic of '''NaN''', You can also read the following blog entry from Bertrand Meyer on the same subject [http://bertrandmeyer.com/2010/02/06/reflexivity-and-other-pillars-of-civilization/ Reflexivity, and other pillars of civilization].}}

	=Benchmarks=		=Benchmarks=

2010-02-09T22:22:16Z

← Older revision		Revision as of 22:22, 9 February 2010
Line 4:		Line 4:

	To best show the trade-offs we will start by showing some benchmark results.		To best show the trade-offs we will start by showing some benchmark results.

			{{Note\|This is an EiffelStudio specific extension to potentially help the ECMA standard committee in making a reasonable decision on the semantic of '''NaN'''.}}

	=Benchmarks=		=Benchmarks=

2010-02-09T22:20:23Z

← Older revision		Revision as of 22:20, 9 February 2010
Line 1:		Line 1:
	We will present some of the trade-offs for computation of IEEE arithmetic for REAL_64 and REAL_32 as implemented in EiffelStudio when enabling the '''Total Order on REALs''' where NaN is not an unordered value but a value less than all the other values (note that in some other frameworks, we have seen it defined as the largest value). In other words:		We will present some of the trade-offs for computation of IEEE arithmetic for REAL_64 and REAL_32 as implemented in EiffelStudio when enabling the '''Total Order on REALs''' option where NaN is not an unordered value but a value less than all the other values (note that in some other frameworks, we have seen it defined as the largest value). In other words:
	* NaN = NaN yields True		* NaN = NaN yields True
	* NaN < x for all x but NaN		* NaN < x for all x but NaN

2010-02-09T22:20:10Z

← Older revision		Revision as of 22:20, 9 February 2010
Line 1:		Line 1:
	We will present some of the trade-offs for computation of IEEE arithmetic for REAL_64 and REAL_32 as implemented in EiffelStudio where NaN is not an unordered value but a value less than all the other values (note that in some other frameworks, we have seen it defined as the largest value). In other words:		We will present some of the trade-offs for computation of IEEE arithmetic for REAL_64 and REAL_32 as implemented in EiffelStudio when enabling the '''Total Order on REALs''' where NaN is not an unordered value but a value less than all the other values (note that in some other frameworks, we have seen it defined as the largest value). In other words:
	* NaN = NaN yields True		* NaN = NaN yields True
	* NaN < x for all x but NaN		* NaN < x for all x but NaN

2010-02-05T19:20:16Z

Testing for 25% NaN values and 75% non-NaN values

2010-02-05T19:17:40Z

Testing for 50% NaN values and 50% non-NaN values

2010-02-05T19:14:38Z

Testing for 10% NaN values and 90% non-NaN values

2010-02-05T19:11:27Z

Testing for 100% NaN values

2010-02-05T19:07:53Z

Testing for non-NaN values which are always the same

2010-02-05T19:04:41Z

Testing for non-NaN values which are always different