NVIDIA Extends Slurm to Schedule Quantum Processors Across Supercomputing Centers

NVIDIA Blog · rss · 2026-11-17

NVIDIA announced at SC26 a Slurm expansion for quantum processors: QPU deployments at supercomputing centers have grown nearly 7x in five years, with 30 centers deploying 50+ QPUs. QPU vendors can now expose one interface across all centers via a simple file, building on QRMI/QDMI; IBM, Pasqal, IQM, Quantinuum, Alice & Bob and MQSC have adopted it. New CUDA-Q releases include a TCBP-JOBS error-correction decoder 20,000x faster than the industry standard, expanded fault-tolerant primitives (Fermilab cut a five-month study to three weeks), and a QunaSys run simulating 300M+ quantum operations on 1,024 Hopper GPUs at chemical accuracy. Slurm 26.11 ships later this month.

Original post →

More from Infra

Infra channel →