Changes between Initial Version and Version 1 of Ticket #2039


Ignore:
Timestamp:
Jul 16, 2013, 8:27:55 AM (11 years ago)
Author:
wraitii
Comment:

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #2039 – Description

    initial v1  
    1 After #1988 was committed, and because of other bits of code already using it (foam generation for the water, decal generation)
     1After #1988 was committed, and because of other bits of code already using it (foam generation for the water, decal generation)
    22
    3 Calculating terrain normals is fairly slow, uses a ton of temporary memory, and is generally not a good idea to do on the fly. #1988 means we'll be calculating normals all the time. We should cache the values of each vertex normal, which will considerably speed up such a task.
     3Calculating terrain normals is fairly slow, uses a ton of temporary memory, and is generally not a good idea to do on the fly. #1988 means we'll be calculating normals all the time. We should cache the values of each vertex normal, which will considerably speed up such a task.
     4
     5I'm fairly sure "CalcExactNormal" also relies on the vertex normals.
    46
    57A few concerns:[[BR]]-Atlas needs a way to not cache terrain or to invalidate the terrain cache (there is a message for that I believe)[[BR]]-A cache should be as quickly accessed as possible.[[BR]]-We need to sort out the fixed v not fixed issue with normals. Imo, we should use the non-fixed version as much as possible for anything graphical (ie not checked for OOS), because it's simply faster.