Pulse-Level Languages, Interfaces and Intermediate Representations

Title

Compiling quantum circuits to control electronics - Abstraction levels, optimizations, pitfalls

Abstract

The coding of experiments has become increasingly complex. Different coding platforms/languages make up an indigestible forest of options to choose from. With increasing qubit numbers and experiment complexity, there is an ever growing need for faster execution of experiments. At the same time, control stack electronics have become more powerful and support more complex instruction sets than traditional AWGs.

In order to both minimize the time required for coding and maximize the speed of execution it is crucial that the software stack offers flexibility with many abstraction levels, but also has an internal structure that is rich enough that the compiler can leverage hardware capabilities.

We show how these challenges can be overcome using Quantify, an open source platform for quantum experiments.

We will first introduce the full software stack for a quantum computer. We then investigate how mixing abstraction levels can simplify the interface for the user. Finally, we show how compiler optimizations improve hardware utilization with minimal overhead for the user.

Title

Enabling experiments at scale using automation in Quantum EDGE 

Abstract

The drive for larger and more performant quantum computing systems has increased the size and complexity of experiments. Such experiments are performed for multiple reasons ranging from QPU calibration, QPU characterisation to pure research. Regardless of these objectives the requirements remain the same: Measurements must be performed, data must be analyzed, decisions must be made and information must be managed.

In this talk we present automation as a solution to this challenge which we achieve using Quantum EDGE. We talk about some of the difficulties of achieving this such as interfacing with control electronics, describing signals and organizing data. We also provide some real world demonstrations of tuning up a 5-qubit chip with transmon qubits from no knowledge to more advanced two qubit gate calibrations.

Title

Design of Pulse-Level Representations for Quantum Systems

Abstract

Quantum computers are typically programmed using the circuit model in a high-level domain-specific language. However, these instructions need to be converted into the language of the quantum hardware - pulses.  Pulse-level representations are used to express these pulses either internally, while being compiled down from gates, or as a domain-specific language for users interfacing with the hardware at the pulse-level. In this talk, I begin by motivating the need for well-designed pulse-level representations. I then discuss important considerations for designing a pulse-level representations, this primarily includes information retention and sufficient semantics for expressing arbitrary pulses.

Motivated by these considerations, I describe pulselib - a graph-based device-agnostic pulse-level presentation developed by us at Duke University. Finally, I showcase how pulselib lends itself to be lowered to a representation compatible with any pulse generations hardware, thereby highlighting its potential as a pulse-level intermediate representation.

Title

The evolution of low-level control for large-scale quantum systems

Abstract

From early NMR platforms to the development of state-of-the-art quantum computers based on superconducting qubits, ion traps, and neutral atoms there has been a rich history of exploring the trade-offs in pulse and circuit-level control. At IBM Quantum we have seen our own evolution of pulse-level control for our systems - from exposing the first pulse-level quantum computer on the cloud to sunsetting these capabilities for our users. In this talk we explore the evolution of pulse-level control, the challenges we have encountered, and the opportunities we see on the error correction horizon.

Title

Crafting a Specialized Pulse Level Language for Quantum Control

Abstract

When controlling qubits, a high level of control and the necessary abstraction are crucial to enable quantum computing research. In this talk we will share our development journey behind the creation of a specialized pulse level language for quantum control. We will trace the origin of this innovation, exploring the challenges in software programming and engineering faced during development.
Key milestones and collaborative efforts that significantly advanced the language will be highlighted and offer the opportunity to understand the background to architectural decisions though this comprehensive overview. Attendees will gain valuable insights into lessons learned, such as precision timing, system level approach and real-time vs near-time distinction.

Title

Pulse Level Control for Automated System Calibration – Learnings from developing a control stack agnostic solution.

Abstract

Automated calibration and characterisation is crucial both for production quantum devices in the cloud or in HPC centres as well as for R&D labs developing next generation quantum hardware. For the former, the goal is usually to consistently maintain gates at high enough fidelity levels with minimal system downtime while for the latter the focus is usually deeper system characterisation to gain insights for design iterations. At Qruise, we develop software tools to address both of these needs and over the last years we have expanded support to use our products with a variety of control electronics stacks and their respective APIs for pulse level control. This talk discusses the various challenges in developing a unified solution that covers the somewhat conflicting requirements of fast calibration, in-depth characterisation, and robust automation; and our learnings from scaling this to multiple control stacks.

Title

A prescriptive view into the future of quantum software stacks

Abstract

The evolution of quantum software stacks seems to be dictated by the interplay between the constant expansion of hardware capabilities and the mounting complexity brought forth simultaneously by quantum control, error correction, and pulse-level optimization. As a result, the boundaries between layers within that stack remain blurry with three distinct consequences for quantum programs: permeation of hardware-level concerns into algorithms, limited opportunities for separation of architectural concerns, and the dependency of algorithms on specific hardware platform implementations. In the long run, all these point to challenges of quantum architecture and software sustainability, as well as a high barrier of entry for those interested in developing quantum applications or investigating new quantum algorithms. In short, successful quantum software stacks need to be characterized by their ability to enable opportunistic refinement across multiple layers of abstraction: current quantum stack implementations fall short in this dimension.
 
This talk takes a bird’s eye view of quantum software stacks from the perspective that it is possible to find general transferable lessons from the evolution of classical computing successfully, lessons that speak about abstract processes and human practices. We will review core existing and emerging principles that suggest how the quantum computing community can re-focus on the larger picture provided by a prescriptive view of quantum software stacks. By prescriptive, we mean dictated by top-level, intended, and specific structural, functional and performance goals and constraints from the start instead of assuming that these will emerge bottom-up as a natural consequence of solving challenges incrementally. As a byproduct, prescriptive views are conducive to self-refining processes for separation of concerns across hardware-software stacks, the identification of information that helps estimate how changes in a layer propagate across those above and next to it, as well as anticipating and avoiding costly dead-ends due to future overcommitment to specific details in a quantum hardware landscape which remains very much in flux. Finally, we will discuss implications for the discovery of new algorithms, the execution of performant applications, and the rational design that should characterize dependable classical-quantum computer systems engineering.

Title

 How might quantum compilers better support pulse level quantum computing?

Abstract

Quantum compilation faces many challenges that traditional compilers do not, equally there's a lot we can learn from and directly leverage from existing traditional compilers. In particular, quantum computation has a richer underlying physics that we are able to exploit for computation beyond a purely logical conceptual model, when compared to classical computing. A growing challenge within the quantum computing community is how we best support the full power of various different quantum computing modalities in a cross-platform way, in order to enable open exploration of quantum algorithms and runtimes with minimal overhead to the user. In this workshop we'll take a look through quantum compilation, how we currently support these challenges and how we might be able to work better collectively towards these aims.