Opened 13 years ago

Closed 13 years ago

#245 closed defect (fixed)

Encoder results are compiler dependent

Reported by: davidf Owned by:
Priority: major Milestone: HM-5.0
Component: HM Version: HM-5.0rc1
Keywords: Cc: fbossen, ksuehring, davidf, jct-vc@…

Description

With high-efficiency configurations, the encoder produces two different bitstreams depending upon which compiler was used.

  • GCC-4.4 and MSVC produce identical results
  • GCC-4.5 and GCC-4.6 produce identical (but different) results.

This does not affect non-high-efficiency configurations.

Change History (4)

comment:1 Changed 13 years ago by DefaultCC Plugin

  • Cc fbossen ksuehring davidf added

comment:2 Changed 13 years ago by davidf

A bisection between 5.0rc1 to 4.1 has revealed:

6804b3c472d791ce315e72cad29a50f408a5eb2f is the first bad commit
commit 6804b3c472d791ce315e72cad29a50f408a5eb2f
Author: huawei <huawei>
Date:   Sun Dec 11 17:06:24 2011 +0000

    G521: Non-square hadamard transform for motion estimation and merge estimation
    
    git-svn-id: svn://hevc.kw.bbc.co.uk/svn/jctvc-tmuc/branches/HM-4.1-dev-motion@1600 ce97f6c2-8f86-0410-b44d-af84b13ec34f

Disabling NS_HAD on 5.0rc1 also resolves the issue.

comment:3 Changed 13 years ago by davidf

  • Resolution set to fixed
  • Status changed from new to closed

Resolved by r1674/jctvc-hm

comment:4 Changed 13 years ago by davidf

  • Cc jct-vc@… added
Note: See TracTickets for help on using tickets.

This list contains all users that will be notified about changes made to this ticket.

These roles will be notified: Reporter, Owner, Subscriber, Participant

  • David Flynn(Reporter, Subscriber, Participant)
  • Frank Bossen(Subscriber)
  • jct-vc@…(Subscriber)
  • karl.sharman@…(Always)
  • Karsten Suehring(Subscriber, Always)