Intel takes another vital step toward democratizing quantum computing

Zilber

New member
By launching a quantum compiler using the famous C++ language, Intel aims to strengthen the software ecosystem of this nascent computing universe and score points with “classic” programmers, limited by other kits based on less efficient languages such as Python.



Do you speak quantum? If you’re a programmer, you’ll soon be able to do this with future Intel machines, as the king of PC processors has just launched its first dedicated development kit (we’re talking SDKs). If it’s not the first in the segment, like the giant IBM that launched and improved its Qiskit SDK a long time ago, Intel’s has an argument that could hold weight: it works with the famous C++ language.

Anne Matsuur

Anne Matsuura is the Director of Quantum Applications and Architecture Division at Intel Labs.
© Walden Kirsch/Intel Corporation

For Anne Matsuura, head of the software part of Intel’s quantum project with whom we were able to speak,” the arrival of a C++ compiler based on a standard LLVM virtual machine is a way for us to bring something new and interesting. If IBM has done a good job with Qiskit, today we will bring a compiler that is capable of scaling as it can manage up to several hundred qubits and uses a language that is appreciated by developers ‘ says the scientist.

While players like IBM regularly make progress on their hardware roadmaps, the software part is less often it a Some news. ” With this C++ compiler we can start talking to engineers and developers “classicswho wonder how to work with a quantum computer. There is still a lot of work to be done to turn these machines into really usable computers. And for that, the software is essential “, continues the scientist.

Python for physicists and C++ for (real) developers?


01 – Intel’s Full Stack Approach to Quantum Computing

Intel’s compiler, which integrates a complete quantum simulator, is, as we have seen, aimed at “classical” programmers, not physicists more commonly encountered in the field. An audience that loves Python for its greater simplicity. On the other end of the spectrum, the hardcore programmers prefer compiled languages like C++. These are the talents that the quantum world radically needs. ” If I had to recruit a new person on my team, it would be a very good classic developer! explains Ms. Matsuura.

In order to attract the best names in programming, Intel has set up five scholarships, in addition to the C++ language itself, that it will award to partner universities. “ The arrival of C++ is a great asset for university professors. The feedback we’ve had since the beta launched last September 2022 has been excellent. », the scientist assures.


02 – Intel Quantum Software Development Kit

She adds that the compiler ” can simulate a quantum computer. Up to 32 qubits in one node and 40 qubits in multiple nodes “. And why not anymore?” Quantum computing does not work like classical. When you add a qubit, you have to double everything, like the memory size. It soon becomes impossible to simulate quantum machines that are too powerful “.
It is this theoretical power that should enable these machines to transform calculations that are untenable for conventional computers in certain specific areas such as fluid dynamics, astrophysics or even the design of new materials. However, this does not mean that this new kind of machine will replace our computers or even normal chips. ” To scale up, quantum computers need support in correcting errors. And for that they will have to work hand in hand with normal supercomputers ‘ says Mrs Matsuura.

Intel believes it can do anything

A 300 mm wafer of spin qubits on silicon.  ©Intel

A 300 mm wafer of spin qubits on silicon. ©Intel

If Intel invested late in the segment compared to its predecessor that is IBM, the processor giant wants to carve out a good place for itself in the segment. Indeed, through software, as we have just seen, but also through its mastery of hardware, Intel is designing not only quantum processors, but also control processors – such as the Horse Ridge II we told you about at the end of 2020. The Santa Clara giant also has its industrial capabilities as an asset here as far as its strategy in conventional chips is concerned. ” We are convinced that the advent of a truly useful quantum computer requires the design of a machine that collects millions of fault-resistant qubits. However, qubits are fragile. That is why we have chosen technologies that we master and that we know have already proven themselves. ‘ says Mrs Mastura. Referring here to the type of spin qubit that Intel uses, a method that depends on waffles of 300 mm silicon that Intel masters perfectly.

We have a 300mm production line completely devoted to the research and development of our quantum business “remembers Mrs. Matsuura. “We are good at producing high volume transistors and we have verification methods that allow us to get the waffles whole, no more chips in a row ». Although recently arrived in the field, Intel already believes itself capable of doing everything from hardware to software. ” Yes, we plan to provide our customers with complete quantum computers, with our hardware, our compiler and our software “, assures Anne Matsuura, who does not make false promises. ” Quantum is still an area of basic research. I hope that within ten years we will see the first commercial machines appear. But there are so many areas that are in the research phase that it’s impossible to predict what will or won’t work. “.

Source :

Intel
 
Top