The Framework Magazine

What is Carbon Free Compute?

Carbon Free Compute means much more than limiting the energy sources that powers personal, bussnesss servers, AI, cloud computing and the internet backbone communication to renewable energy.

Even if all that computation would have run on renewable energy, producing the materials and equipments that generate that renewable energy consume vast amount of carbon-based fuels and cause enormous pollution.

Achieving carbon-free computing require optimal computing systems. The vast overhead operations that is inherent to current computing technology must be effectivly eliminated to reducing it to the absolute physically-possible minimal level. That goal is achievable with the technology we have developed. This website is dedicates to a complete new model of computation as well as new semiconductor architecture that can achieve just that. The entire concept of computation must be transformed with new theories of information and knowledge. The technologies we aim to bring to market are based on new theories of control systems, parallelism, energetic optimality and a new linguistic-based architecture of knowledge and computing.

The new technology replaces the CPU (The Turing Machine and Von-Newman based concepts) as well as Neural Networks based AI concepts with a unified linguistic silicon-engine architecture that effectively eliminates the software concept itself, replacing it with a computational language that is literally understood by the silicon engine itself.

Eliminating the CPU centrality in computers with the need to copy CPU instructions from memory to a processing unit is enabled by the replacement of the current arithmetic Turing Machine computational-function-based model of computation with a new linguistic-based model of computation and a semiconductor architecture that is inherently parallel and energetically optimal.

The similarities between the new technologies to current technologies are few but consequential. The functionality and general Programmability of CPU-based systems is preserved albeit with a speed increase by a factor of hundreds if not thousands. (Big-O of 1 is at the engine core) and a comparable level of energy consumption reduction (eliminating for example the need for Dark-Silicon and making it possible to use existing Dark-Silicon for parallelism).

General Programability is provided by more sophisticated and capable scripting languages that can empower programmers by removing the limits and complexities imposed by current operating systems and compilers. Most important, existing semiconductor design and manufacturing methods (with current PDK's) can be used to implement the new devices making the enormous investment that is being made in the semiconductor industry more consequential and sensible in relation to climate change and the growing (perhaps exponential) demand for computational power and capacity.

The Framework Magazine is the crowd funding and publishing core of the effort to announce and disseminate new knowledge theory concepts, as well as bring to market computing and semiconductor technologies that reset Moore's Law to where it was at its beginning and shift control over information and knowledge from the computing industry to the users hands.

Framework and Views "Nonware" are based on new theories of information and knoladge, and the technologies we aim to bring to market are based on new theories of control systems, parallelism, energetic optimality and language as an architecture of knoladge.

The new technology replaces the CPU and Neural Networks with a unified linguistic architecture silicon engine that effectively eliminates software by replacing it with a language that is literally understood by the silicon engine itself.

Eliminating the CPU centrality, the Neural Networking intensive statistical dependency and the need to copy CPU instructions from memory to a processing unit is enabled by the replacement of the current arithmetic Turing Machine computational-function-based model of computation with a new linguistic-based model of computation and a semiconductor architecture that is inherently parallel and energetically optimal.

The similarities between the new technologies to current technologies are few but consequential. The functionality and general Programmability of CPU based systems is preserved albeit with a speed increase by a factor of hundreds if not thousands. (Big-O of 1 is at the engine core) and a comparable level of energy consumption reduction (eliminating for example the need for Dark-Silicon and making it possible to use existing Dark-Silicon for parallelism).

General Programability is provided by more sophisticated and capable scripting languages that can empower programmers by removing the limits and complexities imposed by current operating systems and compilers. Most important, existing semiconductor design and manufacturing methods (with current PDK's) can be used to implement the new devices making the enormous investment that is being made in the semiconductor industry more consequential and sensible in relation to climate change and the growing (perhaps exponential) demand for computational power and capacity.