The New QuantumFlow AI Architecture Boosts Deep Learning Speed By 10x-15x Faster

AI and Quantum Computing is a deadly combination. The basic of AI involves building & training of the model, whereas Quatam Computer offers a whole new style of computation that is way faster than a traditional computer, recently Google also claims that there Quantum Computer can do 10k-year calculation in 200sec.

In Consumer Electronics Show (CES 2020) PQ Labs Inc introduces a QuantaFlow AI architecture which is first of its kind in the industry and it could change the future of AI and Deep Learning inference solutions.



The new QuantaFlow AI architecture includes a classical RISC-V processor, a QuantaFlow Generator and a QF Evolution Space. The QuantaFlow AI SoC architecture is designed to simulate massive parallel transformation/evolution that is very similar to Quantum Computation.

The QuantaFlow simulates a virtual transformation/evolution space for qf-bit registers.




A classical single-core RISC-V processor is implemented to provide logical control, results in observation retrieval, etc. The QuantaFlow Generator converts input data from low dimensional space to high dimensional space and then starts continuous transformation/evolution.

The process is of minimum granularity, highly parallel in nature and asynchronous. By the end of the process, information needs to be extracted from the evolution space by the Bit Observer unit.




In addition, Hot-Patching can be used to change the evolution path of qf-bits dynamically.

When a more significant deformation for the evolution space is needed, the RISC-V processor will issue a warm-“reboot” to the evolution space. All these operations can be executed in a blink of time.

With dynamic operations, QuantaFlow is possible to run all kinds of neural network models e.g. ResNet-50 (2015), MobileNet (2017), EfficientNet (2019), etc.) without speed degradation or hitting the “memory wall.”

By comparison, GPUs and ASIC AI accelerators degrade performance in newer models (MobileNet, EfficientNet), because these new models are all memory-bound.

With all the above efforts, QuantaFlow can achieve 10X speedup in ResNet-50 (batch=1, accuracy=93%, INT8) compared to Nvidia V100 in the same network configuration.

For newer network models, there will be significantly higher speedups to be announced.

QuantaFlow architecture is just a one-step further to explorer superior performance in AI deep learning inference. There are many devils in the details and innovation areas.



The QuantaFlow architecture design flow is accelerated by high-level languages (instead of using Verilog) and implementation is optimized by in-house algorithms to extract maximum horsepower from silicon.

The classical computer function in a binary fashion: they carry out tasks using tiny fragments of data known as bits that are only ever either 1 or 0. But fragments of data on a quantum computer, known as qubits, can be both 1 and 0 at the same time.

Quantum Computing is based on the continuous unitary transformation of qu-bits. A qu-bit (quantum bit) can represent different possible states (e.g. a cat being mathematically both live and dead at the same time).

Like the classical computing model, quantum computing also has 3 major procedures: Input, Process, and Output. But unlike the classical model, the process part of quantum computing is done by continuous transformation / Evolution other than step by step read-control-writeback of a classical Turing machine.

There is no observation (read/writeback) until the final stage.

Each qubit is made from a tiny, plus sign-shaped loop of superconducting wire. This property, known as superposition, means a quantum computer, made up of several qubits, can crunch an enormous number of potential outcomes simultaneously, which ultimately leads to faster execution.

In the face of such tech difficulties, other approaches of quantum-like computation are being explored.

A Canadian tech startup D-Wave announced its 5,000 qubits quantum annealer computer in 2019. However, scientists argue that it’s is not “true quantum,” but a quantum algorithm and quantum simulation.

Despite the academic debate, D-Wave is the world’s first “quantum computer” that solves actual problems such as drug molecule classification, ads optimization, etc. However, the computer is huge and costs about $15 Million, thus a lack of mass adoption.

There are other approaches trying to bridge the gaps between Quantum-style algorithm implementations into real silicon of Artificial Intelligence. QuantaFlow is only a step toward that bright future.

More in AI

Facebook’s AI Mathematician Can Solve University Calculus Problems

AWS launches SageMaker Studio, a web-based IDE for Machine Learning

No Degree In AI Or Data Science Needed To work In Tesla, Elon Musk Reveals

Yann LeCun and Yoshua Bengio Elected as AAAI-20 Fellows



Leave a Reply

Your email address will not be published.