BLACKFIRE is the industry-leading quantum edge-computing framework by Black Brane Systems, providing a consistent development experience across every quantum architecture without compromise, streamlined access to every available quantum computing architecture and quantum cloud service, and breakthrough proprietary quantum application development and embedding tools for domains including: quantum algorithms, quantum virtualization, generalized quantum automata, quantum metaprogramming, quantum language design, quantum compilation, quantum chemistry, quantum machine learning, and quantum cognitive programming.
Coming soon to the Azure Marketplace.
Using the included quantum virtualization tools, Blackfire can be used to implement Quantum Turing Complete virtual quantum machines with up to eight 128-qubit, fully entangled quantum registers partially coupled over an 8-qureg quantum bus.
Access all commercially-available quantum computing services through a single, consistent API without compromising on platform-specific features.
Protocol "Phi", included with BLACKFIRE, is a robust, extensible, strongly typed, higher-order quantum object system providing over 400 native type-classess to handle the hairiest quantum programming domains.
BLACKFIRE is available as a runtime library for software applications, and for device embedding; we currently support Windows, Linux, macOS, iOS, and Android operating systems on a variety of host architectures, including x86-64, ARM, RISC-V, and NVIDIA GPUs.
The BLACKFIRE Quantum SDK was built to support every programming language from the beginning. Use what you know, integrate components of any language, get up and running fast. No other solution on the market can provide this kind of coherence.
Instances of Blackfire adapt to the customer's needs, taking advantage of Quantum Virtualization to optimize cache-layout and speculative execution strategies for the common use-cases, in complete isolation from each other, as well as from us and our service providers.
Simulated Quantum Machine (SQM)
The evolution of a quantum state-space is explicitly represented and computed by a classical host; this requires the complete state-vector for every quantum state to be represented in memory, and application of quantum operators (equivalent to matrix multiplications over state-vectors), correctly ordered, limiting the advantage of real-time scaling through parallelization and pipelining based optimization techniques.
Emulated Quantum Machine (EQM)
The evolution of a quantum state-space is made more efficient than possible with SQMs, by relying on classical equivalent operations and approximation wherever possible. For example, the Hadamard Gate can be emulated most efficiently by using either non-deterministic choice available in many logical, functional, and multi-paradigm programming languages, or phi-functions from the Static Single-Assignment (SSA) representation model used within compiler frameworks.
Virtual Quantum Machine (VQM)
Quantum Objects within virtualized quantum state-spaces are represented internally as hyperlattice codes, a maximally-compressed, lossless, prefix-free, variable-width binary data-structure optimized for higher-dimensional non-linear graphical models such as used in classical, quantum, hybrid, and quantum-enhanced machine learning applications. Hyperlattice Coding resembles many other graphical model encodings, such as Reduced Binary Decision Diagrams, Huffman and Arithmetic Coding, and real-time binary serialization protocols like SignalR, Cap'n Proto, Katai Structs, Messagepack, etc. Instead of needing to re-compute the solution to the time-dependent Hamiltonian of the system, which has at minimum exponential space complexity for SQMs and EQMs, VQMs generate lattice codes on-demand (lazy evaluation) from a basis model encoding the unique quantum signature of the system, and use a flow-based memory model to achieve self-optimizing function caching (dynamic programming). which has been harvested and needs to be periodically refreshed from a coherent physical quantum system.