View Issue Details

Jump to NotesJump to History
IDProjectCategoryView StatusDate SubmittedLast Update
0002653DarkRadiantRendererpublic24.02.2011 12:0724.09.2011 08:20
Reportergreebo Assigned Togreebo  
PrioritynormalSeveritynormalReproducibilityhave not tried
Status closedResolutionfixed 
Product Version1.5.0 
Target Version1.6.0Fixed in Version1.6.0 
Summary0002653: Convert math classes from double back to float
DescriptionThere is a 10-20% performance penalty when using double precision floats over single-precision ones (and a few dozen MB larger memory footprint compared to single-precision). As the original issue 0000266 was solved by setting the precision() hint on std::ofstreams during map export there seems to be no downside when using floats. Tests showed that moving maps to the corner of the mappable area (near the 65k coordinates) don't exhibit any precision loss between repeat map load and save operations.
TagsNo tags attached.

Relationships

related to 0000266 closedgreebo Loss of significant information when using float 
child of 0000035 closed Renderer performance is slow 

Activities

There are no notes attached to this issue.

Issue History

Date Modified Username Field Change
24.02.2011 12:07 greebo New Issue
24.02.2011 12:07 greebo Status new => assigned
24.02.2011 12:07 greebo Assigned To => greebo
24.02.2011 12:07 greebo Relationship added related to 0000266
24.02.2011 12:09 greebo Status assigned => resolved
24.02.2011 12:09 greebo Fixed in Version => 1.6.0
24.02.2011 12:09 greebo Resolution open => fixed
24.02.2011 12:10 greebo Relationship added child of 0000035
24.09.2011 08:20 greebo Status resolved => closed