Opened 3 years ago

Closed 3 years ago

#21417 closed defect (fixed)

numerical noise in gsl on 32-bit Linux

Reported by: dimpase Owned by:
Priority: major Milestone: sage-7.4
Component: packages: standard Keywords:
Cc: Merged in:
Authors: Dima Pasechnik Reviewers: Frédéric Chapoton
Report Upstream: N/A Work issues:
Branch: 8f40145 (Commits) Commit: 8f401455e147ea284e89da4fa2e96b8a9fb07fa7
Dependencies: Stopgaps:

Description

on 32-bil Lunux with gcc 6.1.1 one gets 

sage -t --long --warn-long 42.0 src/sage/gsl/probability_distribution.pyx
**********************************************************************
File "src/sage/gsl/probability_distribution.pyx", line 372, in sage.gsl.probability_distribution.RealDistribution
Failed example:
    T.cum_distribution_function_inv(.5)
Expected:
    3.532230067546424
Got:
    3.5322300675464247

Change History (6)

comment:1 Changed 3 years ago by dimpase

  • Branch set to u/dimpase/gslnoise
  • Commit set to 8f401455e147ea284e89da4fa2e96b8a9fb07fa7
  • Status changed from new to needs_review

not sure whether this is an upstream bug, or a compiler issue...

New commits:

​8f40145adding ... to supress the extra digit

comment:2 Changed 3 years ago by chapoton

  • Reviewers set to Frédéric Chapoton
  • Status changed from needs_review to positive_review

lgtm

comment:3 Changed 3 years ago by dimpase

  • Authors set to Dima Pasechnik

comment:4 follow-up: Changed 3 years ago by leif

For such things we have # tol .... (See for example #21411.)

comment:5 in reply to: ↑ 4 Changed 3 years ago by dimpase

Replying to leif:

For such things we have # tol .... (See for example #21411.)

yes, but this would mean a loss of precision compared to '...', as proper rounding would be done, right?

comment:6 Changed 3 years ago by vbraun

  • Branch changed from u/dimpase/gslnoise to 8f401455e147ea284e89da4fa2e96b8a9fb07fa7
  • Resolution set to fixed
  • Status changed from positive_review to closed
Note: See TracTickets for help on using tickets.