Pull to refresh
Logo
Nvidia makes quantum processors a schedulable resource in Slurm

Nvidia makes quantum processors a schedulable resource in Slurm

New Capabilities

Slurm 26.11 adds a plugin interface that lets supercomputing centers queue a QPU next to CPUs and GPUs

Overview

Updated 40 minutes ago

A quantum processor at a supercomputing center has, until now, sat outside the job queue: researchers booked time through a vendor's portal and carried results back by hand. Slurm, the scheduler most large centers run, will let them request one the way they request a GPU.

The mechanism is a new plugin interface called Hierarchical Resources, shipping in Slurm 26.11. It gives a quantum processor a native entry in the scheduler's resource list, so a center can cap how many jobs use it, hold jobs while the device is down for calibration, and account for its time. Six quantum hardware and software companies say they support it.

Why it matters

Quantum hardware becomes ordinary cluster capacity, queueable and accountable, instead of a separate service with its own login and billing.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

6
Quantum vendors backing the release
Alice & Bob, IBM, IQM, Pasqal, Quantinuum and MQSC are named as adopters of the Slurm expansion.
~30
Supercomputing centers deploying quantum processors
NVIDIA counts nearly 30 centers in the process of installing quantum hardware, as of November 2026.
50+
Quantum processors being deployed at those centers
NVIDIA says more than 50 processors are headed into HPC sites, a roughly sevenfold rise in five years.
26.11
Slurm release carrying the feature
Scheduled for release later in November 2026, according to NVIDIA.

Voices

Curated perspectives — historical figures and your fellow readers.

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

2026 November 2026

4 events Latest: Today
Tap a bar to jump to that date
  1. NVIDIA adds quantum processors to Slurm's resource model

    Today Product announcement

    Slurm 26.11 adds a plugin interface for hierarchical resources, supporting the QRMI and QDMI interfaces. Alice & Bob, IBM, IQM, Pasqal, Quantinuum and MQSC are named as adopters.

  2. MQSC shows an IQM system scheduled through Slurm

    Today Demonstration

    The open-source MQSC Orchestrator combines Slurm's new resource interface with the device-management interface to schedule IQM machines at SC26.

  3. IBM, Fermilab and NVIDIA outline a modular quantum co-design platform

    Today Research partnership

    The three announced plans to co-design next-generation modular quantum computers, citing Slurm's ability to allocate and account for QPU time under existing HPC policies.

  4. CINECA runs production jobs on a quantum processor through Slurm

    Deployment

    The Italian center reached Pasqal's Sol processor through a plugin built on the Quantum Resource Management Interface, the route sites use ahead of the new release.

Scenarios

1

Centers start queueing QPUs like any other cluster resource

Likely Resolves by Q3 2027

Discussed by: NVIDIA, Pasqal and IQM, each of which describes centers moving to the new interface as they upgrade

Slurm 26.11 ships and centers that already host quantum hardware add a resource definition to their configuration. Jobs request the QPU in the same batch script as their GPUs, and administrators set sharing limits in one file. The signal to watch is documentation, not press releases: a center's user guide telling researchers to write a resource request instead of using a separate portal.

2

Quantum scheduling stays a vendor plugin, not a scheduler feature

Possible Resolves by Q3 2027

Discussed by: Pasqal's engineering post, which calls the older plugin route "the production route today"

The new interface is optional, and the existing plugin keeps working with the same programs. If centers are slow to upgrade Slurm, or if the new interface matures more slowly than expected, sites stay on the plugin and the change stays invisible to researchers. Pasqal itself describes moving from one to the other as a small step, which cuts both ways.

3

A published result comes out of a single Slurm allocation

Possible Resolves by End of 2027

Discussed by: NVIDIA and Pasqal, which both say CUDA-Q and Pulser programs run unchanged under the new path

The technical case only lands if a research group runs classical preparation, QPU execution and analysis inside one scheduled job and writes it up. That would show the interface survives contact with real science rather than a demonstration. Watch preprint servers and conference proceedings in computational chemistry and materials science.

Historical Context

2 moments from history that rhyme with this story — and how they unfolded.

2010s

GPUs become requestable resources in cluster schedulers (2010s)

As GPU accelerators spread through academic clusters, workload managers added generic resource tracking so a batch script could ask for one accelerator on one node. Before that, labs reserved whole GPU machines or managed access by hand.

Then

GPU jobs became a line in a submission script instead of a separate arrangement with whoever ran the hardware.

Now

Shared clusters could serve many labs running accelerator code, which is the arrangement most machine-learning research still depends on.

Why this matters now

The pattern repeats here: a new kind of processor gets a resource definition, and adoption follows the scheduler rather than the vendor.

2016

IBM puts quantum processors on the public cloud (2016)

IBM began letting outside researchers run experiments on its quantum processors over the web, with its own account system and queue. Other vendors followed with portals of their own.

Then

Thousands of people ran small quantum programs without a lab, but each vendor's access path was separate from campus computing.

Now

Cloud portals became the default way to reach quantum hardware, and that separation is what kept QPUs out of HPC queues for a decade.

Why this matters now

The Slurm change targets exactly that separation, replacing per-vendor portals with the queue a center already runs.

Sources

(7)