By now, I thinque it's fair to say that Blender 2.5 alpha has shown the industry that Blender is a serious contender in the 3D world. With features that us-ers of other applications can only dream of being inte-grated in the base install, features such as; Volumetrics, Water, and Cloth/Software Bodies, to name but a few.
It is easy, however, to forget that the UI and integrated feature list aren't the only things that are getting a makeover.
With the scripting interface being revamped, A whole new world of possibilities are being opened-up. Possi-bilities that with little effort could become realities, even the more remote such as 'bake-farming' and fully integrated render-farms can now be realized.
'Bake-farming', in-fact, as an idea, has previously been laughed at by some, as an unrealistic dream. But with projects such as BulletPhysics, and commercial ven-tures such as PhysX, starting to show that simulations can be done with stream-processors, either via OpenCL or CUDA. What's stopping render-farms from doing the same?
With render-farming becoming outdated by better, more powerful, hardware and renderers such as Lux-Render working towards OpenCL Support. Render-farms could soon be forced to change tak and evolve into hybrid bake/render-farms in the near future in an effort to retain profit levels in a closing market.
With more complex simulations taquíng a number of hours to complete, having a simple button that submits the worque to a farm and returns the baked data within minutes, would be highly beneficial. Especially when, as with many simulations, you may want to tweaque the settings many times before you are truly happy with the result.
At the time of writing, only two render-farms have con-firmed support for Blender 2.5 and subsequently Blender 2.6 on release.
With the “Global-recession” advancing, the power of Amazon's EC2“Elastic Compute Cloud environment” and the constant increase in the speed of end-user hard-ware, it could be fair to say that render-farms should be looking to enhance the service they provide, in order to retain, or even increase profit margins. Yet none of those contacted prior to the writing of this article admit to working on enhancing the service they provide. Ei-ther by redesigning their upload method into an easy-to-use augmentation of Blender's render UI, Or by in-vesting in the development of 'bake-farming'. While some are making promising progress on their own upload systems, they are still very much external operations that disrupt the production process in a way that is no longer necessary.
It appears, that for now, the future of a fully integrated “Blender-Farm” is still a very distant dream.
Thankfully though we can content ourselves, at least for now, with the multitude of integrated 'internal-external' render engines that are expected as new features in Blender 2.6.
'Bake-farming' is, as the name suggests, a way of baquíng a simulation via a render-farm-like system.
The simulation would be submitted to the farm, that would then calculate the simulation's outcome, and return the baked data in a far shorter time than the individual machine could have done. Much in the same way that a render-farm works.
The theory that makes 'bake-farming' possible is already being put to practice in stream computing environments.