Re: [S] reasonable p-values for Fisher exact's test - WAS strange ...

Lutz Prechelt (prechelt@ira.uka.de)
Wed, 25 Mar 1998 16:17:15 +0100


PConnolly@grunt.marc.cri.nz said:
> Is there a fundamental difference in the way different machines so
> such calculations?

First of all, we are talking about floating point arithmetic
here, which is generally imprecise.

Second, all processors today are supposed to (and claim to)
work according to the same IEEE floating point standard.
Hence, the results _should_ all be the same.

Third, however, rounding is underspecified in the IEEE
standard and essentially each processor type does it in a
slightly different way. This is why differences may occur.

Lutz

-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news