「利用者:DingTo/GSoC 2014/Weekly Reports/Week11」の版間の差分

提供: wiki
移動先: 案内検索
 
(1版 をインポートしました)
 
(相違点なし)

2018年6月29日 (金) 06:09時点における最新版

Week 11

What I did this week

This week I worked mostly on Volumetric optimizations.

  • Avoid some calculations for Volume Lamp Sampling. Now we only calculate what is necessary for Equi-Angular, and then in case we have a valid scattering event, we do the remaining parts. This could be optimized further, by avoiding the pre calculations, in case we only use Distance Sampling. I also need to double check the correctness for Mesh Lights, it seems that there is a remaining issue. (9406c1989536, 776ce62bee4d).
  • Optimize heterogeneous volume shadow. This gives a few percent speedup, but needs to be tested with more files still. (1f432d2b06c4, 8f5eb7a5638d)
  • Cleanup: Avoid intermediate function when we sample all lamps. (773c58524d8e)
  • Bug tracker maintenance (check on reports, bisected one Cycles issue, close invalid/fixed ones...)

Next week

Finish the volume optimizations and merge them to master.

Questions

None.