git.ucc.asn.au / ipdf / code.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: b8c16a2) | patch
Tester from some lecture notes by Kahan
authorSam Moore <[email protected]>
Thu, 1 May 2014 10:55:28 +0000 (18:55 +0800)
committerSam Moore <[email protected]>
Thu, 1 May 2014 10:55:28 +0000 (18:55 +0800)
commite6e138e8baad2dda39cecbb35557710f2a550398
treea749d7cd6f0bca0c1a799b5d26af7ff8a8cab390tree | snapshot
parentb8c16a21a8f59ff8cde6bb546cc44a6fc422c2b7commit | diff
Tester from some lecture notes by Kahan

Supposedly this is a benchmark for determining the worst accuracy of a CPU using roots to a quadratic
I think something wierd is going on; a result of infinity means there is no rounding errors but I'm not sure what results of -nan mean
and that's what I get.

The article is referenced as kahan1996ieee754 in the literature notes
src/tests/qtest.cpp [new file with mode: 0644] blob
Infinite Precision Document Formats - Code base

UCC git Repository :: git.ucc.asn.au