BLACKFIRE, the industry-leading quantum edge-computing framework, provides a consistent, expressive, and extensible quantum development experience without compromise. Work in your choice of host programming language. With high-performance interactive quantum programming tools. Across quantum architectures of multiple vendors. From any quantum-cloud service. Running jobs at unprecedented scale. For any target device, application, or infrastructure. To do the job right that no one else can do. Quantum advantage. Powered by Black Brane Systems.
Community-driven marketplace, knowledge-sharing and learning platform for Blackfire quantum edge-computing products, services, resources, and components. Coming soon.
Managed, private, isolated, and self-hosted application instances of Blackfire Quantum-Core designed for the Enterprise and optimized for scalability. Coming soon to the Azure Marketplace.
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 four hundred native type-classes to handle the hairiest quantum programming domains.
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.
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.
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.
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.
Quantum Virtualization enables performance and scaling improvements over simulation and emulation techniques, by avoiding expensive operations whenever possible. VQMs are distinguished by the principle of least commitment applied to managing the trade-offs between time and space complexity for modelling quantum systems. This is achieved through use of hyperlattice codes (see below), for serialization of physical quantum states; lazy evaluation and dynamic programming techniques for maintaining a low footprint; higher-order type theory for strong semantics and proof of correctness; heuristics developed using quantum machine learning models; and a curated collection of significant design patterns from graph and flow-based network optimization, numerical algorithms, high-performance computing, structured parallel programming, artificial intelligence, machine learning, and compiler pipelining, to ensure best feasible use of available compute resources.
VQMs generate hyperlattice codes on-demand, using lazy evaluation for query resolution in their Quantum State-Space; the QSS is a "black box knowledge base", containing the eigenbasis of independent eigenvectors of the hyperlattice projection matrix operators, and the unique quantum signature of the system. A flow-based memory model is used to achieve self-optimizing predictive function caching as by dynamic programming.
Structured physical quantum-states and quantum state-spaces are harvested using Blackfire's own Universal Quantum Cloud API. These sampled quantum states are refined using proprietary variational methods for sampling, correlation, amplitude amplification, and eigendecomposition of the hyperlattice projection matrix operators.
The Quantum State-Space itself is a dynamic system under quantum evolution, subject to decoherence, and as such needs to be periodically refreshed from physical quantum machines.
Hyperlattice Coding is a maximally compressed, lossless, prefix-free, variable-width, executable binary type, optimized for higher-dimensional non-linear graphical models such as used in classical, quantum, hybrid, and quantum-enhanced machine learning applications. Hyperlattice Codes resemble those of other graphical models, such as Reduced Binary Decision Diagrams, Huffman and Arithmetic Coding, and real-time binary serialization protocols like SignalR, Cap'n Proto, Katai Structs, MSGPACK, etc.
Executability of Hyperlattice Codes enables them to be treated as first-class representations of quantum machines, quantum programs, and their components. A few noteworthy advantages of Hyperlattice Codes for serialization of machine, program, and component instances, include structural symmetry between machines and their programs, often enabling programs to be compressed into the machine instance that runs them.