alexwbc wrote: ↑
Thu Jan 31, 2019 2:59 pm
The problem with nvidia is while they have the money power to hire pretty much the best minds around, they want to be sure to get their money back: it doesn't matter the revelance of a technology or advancement of the whole industry. They push forth or back being tightly tied on how much money they do straight away.
Principally business depending enterprises needs to follow market roules. Currently one could not speak of a free market for integrated circuits because there are only a few companies which can afford the enormous investment required for research and development and production of reliable clean and working prototypes in the low nanometer scale (less than 40 times of the wavelength of visible light).
Even if one is capable to simulates a 2019 at the edge CPU or GPU with all their circuits, before prototyping, with access to libraries for all kind of common base untis (e.g. memory, controllers, computing unitns, prediction units) and experiences to arrange and microcode them, the simulation has limitations especially for the next miniaturization step because of yet unknown interferences .
And if a physical simulation were possible, the simulatiom of billions of electronic units like transistors, capacitors and resistors with all their thermal and electromagnetic properties it required large scaled computing power.
Realizing a piece of hardware consisting of billions of single multilayered elements properly means x times of chemical treatment of perfectly cristallized silicon to get finally what endures thounds of billions charging and thermal cycless.
Therfore I cannot blame any of these companies for their strategy if they have to be profitable and cannot rely on support due to lack of strategic relevance.
160.8 | 42.8 (10.7) Gfp / Windows 10 Pro, intel i7 4770K@3.5, 32 GB | AMD R9 290x+R9 390x, 4 GB
17.3 | 19.0 ( 4.7) Gfp / macOS X 13.6, iMac 27'', 2010, intel i7 email@example.com, 24 GB | ATI Radeon HD 5750, 1 GB
#luxcorerender | Gfp = SFFT Gflops