20 September 2016
This is a bugfix release, parallelizing TX generation.
This release uses Arnold 126.96.36.199 and OpenVDB 3.1.0.
Binaries available for the following Houdini, Houdini FX, Houdini Indie and Houdini Education production builds:
and for the following platforms:
- Linux x86_64 (gcc4.8)
- Windows 7 x64 (vc11, vc14)
- Mac OS X 10.8+ (clang7.3)
Please note that Houdini Apprentice does not support third party renderers and thus cannot run HtoA.
- Parallel TX generation: The generation of TX textures is now run on a customizable number of worker processes in parallel. The number of workers can be set on the Arnold ROP for Auto-TX, or passed as a parameter to
htoa.material.updateAllTx(). For the default value of 0, the number of workers is the number of available logical cores, clamped to 10. This heuristic is what gave us the the best results on the various OSes and configuration we have tested so far, but you can set the default value with the
HTOA_MAKETX_WORKERSenvironment variable if needed. Compared to the old sequential maketx updating, we have measured speedups of 52x in the degenerate no update case to 7.3x when all textures are updated. More improvements to come. (htoa#815)
- Fixed SOHO initialization error in batch mode with no cameras (htoa#816)
- Fixed infinite IPR refresh loop with forced materials and object subnetworks (htoa#822)
- Fixed cannot use * as forced material pattern (htoa#823)
- Fixed exceptions when exporting relative camera paths (htoa#821)
- Fixed reflectance at normal going beyond 1 in the Standard shader UI (htoa#825)