Dirac-3

Hard problems require unconventional solutions

diract3

Introducing Dirac-3, our highest performance quantum machine, purpose built for complex optimization problems.

Trial cloud access

Get access to entire line of Dirac device, including Dirac-3, with enough time allocation to test basic capabilities.

This includes:

  • 10 minutes of time allocation

Free


Hourly cloud access

Purchase access to the entire line of Dirac devices, including Dirac-3, on an hourly basis.

This includes:

  • 60 minutes of time allocation

$1,000/hour

$4,000/5 hours


Cloud + concierge

Work directly with our team of applications scientists to formulate, optimize, and execute your problems.

This includes:

  • 60 minutes of time allocation

  • 1 hour of scheduled, dedicated access to one of our Dirac machines

  • 1 hour session with a QCi Application Scientist for technical support and application coaching

$2,000/hour

$10,000/6 hours


On prem

Install Dirac-3 within your local premises.

  • Eliminate queueing delays from shared usage

  • Achieve fastest loading speeds with EQC-Direct

  • Server room temperature operation

  • Superior SWaP-C characteristics

Starting at $300,000


Being different makes all the difference

We are making steps with an unconventional approach to computing with nanophotonics and quantum optics.
And we have the papers to prove it.

Interaction- and measurement-free quantum Zeno gates for universal computation...

Experimental demonstration of interaction-free all-optical switching via the quantum Zeno effect

Antibunched emission of photon pairs via quantum zeno blockade

A SWaP-C superiority in datacenter infrastructure

swap-c

Size

5U rack-mounted

Weight

Under 30kg

Power

Under 100 W total

Cost

On-prem: starting at $300,000,

Cloud access: starting at $1,000/hour

The computationally hard made easy

Dirac-3 is purpose-built to solve computationally hard optimization problems and provide a polynomial advantage for approximate solutions. To learn how to use Dirac-3, visit our Introduction to Dirac-3 learning module.

Discrete solver

Non-convex Optimization

Two body interactions

Combinatorial optimization

All to all connectivity

Low power consumption

In a league of its own

Runtime by schedule
Power consumption

949

Variables

Qudit

Based architecture

Global

Connectivity

100W

Computational Power Draw

Quantum machines for a brighter future

Dirac-3 Image 1
Diract-3 Image 2

Let's get a little technical

A brief description

Dive a little deeper into how Dirac systems work. To see our research and publications, click here.

In quantum information processing, loss and noise are usually detrimental and must be minimized. This is why quantum systems using atomic and alike qubits must be hosted in cryogenic vacuum chambers, and why photon loss is the roadblock to quantum communications and computing. This requirement translates to exceeding challenges in quantum system manufacture and operations, and has been the bottleneck preventing the scaling up of the qubit number and connectivity. 

With entropy quantum computing, we flip the coin around. Instead of trying to avoid loss and noise, we harness them to build quantum machines whose capacity and speed outmatch existing computing modalities. 

This fundamentally new quantum computing approach is called Entropy Quantum Computing (EQC). It roots deeply in the intriguing principles of quantum mechanics. First, loss or decoherence of a quantum state occurs through its coupling to an entropy source with many degrees of freedom. The apparent diminishing of quantum characteristics as a result is just a statistically averaged manifestation of many possible outcomes of such coupling. Second, vacuum is never quiet, although it does not appear to contain any energy or particle. There are, in fact, enormous amounts of random fluctuations occurring at all times in each of the vacuum mode.

EQC is conceived and developed with those intriguing quantum principles. Rather than trying to create and manipulate pristine qubits isolated from the environment, EQC utilizes loss and decoherence, and turns entropy into super-power fuels of its computing engine. In sharp contrast to any existing quantum platforms, there is no need for cryogenic or isolated housing, and the implementation can use integrated photonics, leading to SWAP-C friendly devices, just like regular PC’s.

How it works

Our quantum analog computers take an ising hamiltonian as an input and find the lowest energy state, solving highly interconnected variable optimization problems, very quickly.

These problems correspond to minimizing or maximizing the expected return of the objective function:

dirac-equation

under the constraint of a fixed resource R=i=1NViR=\sum_{i=1}^NV_i where ViVi is the value of each variable, CiCi is the linear coefficient of each variable, which is a real number that can be positive, negative, or zero, JijJij is the coupling coefficient of two variables, which can be any real number.

how it works

Specifications

Solver Type

Constrained Discrete Number Optimization

Hardware Type

Hybrid Analog Machine with Quantum Optics and Digital Electronics

Maximum Number of Variables

949

Order of Correlation

Any types of first- through
fifth-order correlations,
where the interaction amongst variables
can be repulsive (positive correlation)
or attractive (negative correlation).

Connectivity

All-to-all

System Power Consumption

under 100 Watts

Storage Temperature

-25◦C to 85◦C

Operating Temperature

20◦C to 27◦C

Maximum Rate of Change

2◦C per hour

Software Requirement

On-prem: eqc-direct software package, Python 3.10.6 (recommended)

Cloud: qci-client software package

OS requirement

On-prem: Linux (recommended)

Dimension

5U rack-mounted

Frequently asked questions

Dirac-3 is available on prem or via cloud access. For cloud access, you can sign up for an account that includes 10 minutes of free runtime. It is then $1,000/hr after that. For on prem, a device can be purchased for $300,000.

Since our Dirac systems run at room temperature and use industry standard I/O, they can be integrated into your data center without having to make any changes to your existing IT infrastructure footprint or facility. More details on the integration can be found in our user guide here.

A quick summary of entropy quantum computing can be found here and a list of our publications on the subject can be found here.

We offer a comprehensive, flexible suite of quantum professional services including problem identification, formulation, execution, monitoring, results analysis, and training. These services are provided by highly experienced quantum scientists, with extensive backgrounds in mathematics, quantum physics,  AI/ML, or optimization.

You can purchase time allocation in blocks. As a user submits problems to the queue, the run time of a problem is deducted from the user’s available time allocation balance. The run time deducted for a particular problem represents only the time when the problem is running on a Dirac machine – you are not “charged” for time spent in the queue. Blocks of time allocation can be used for up to 12 months and can be used on any of Dirac-1, Dirac-2, or Dirac-3 that are available on QCi’s cloud. Once your available time allocation is consumed, the user will be unable to submit problems to the queue until the allocation balance is increased.

If you run a problem that consumes more Time Allocation than you have in your balance, the submitted problem will still be queued and processed. Your Time Allocation will then reflect a negative balance for any overage consumed. You must purchase sufficient additional Time Allocation to obtain a positive balance before new problems can be submitted to the queue. For example, if you have a negative balance of 4 minutes and purchase a 60 minutes of Time Allocation, your net balance will then reflect 56 minutes of Time Allocation and then you will be permitted to submit problems.

The time it takes to run a problem can vary greatly, depending on the problem you run. In general, the time required scales linearly to the size of the problem, but the selected schedule can have an effect as well. Small test problems may only take a few seconds. Larger problems could take several minutes. This does not include the time in queue, which you will not be charged for. More info can be found here.

QCi offers Dedicated Access to our Dirac-1, Dirac-2 and Dirac-3 machines, so that users may access these machines without any interruption from other problems in the queue. Dedicated Access is provided on an hourly reservation basis.

Additionally, during those Dedicated Access sessions, a QCi Application Scientist is available to provide the user with over-the-shoulder technical support and application coaching.