Quantum Machines (QM) has launched the NVIDIA DGX Quantum Early Customer Program, involving six research groups and quantum computer builders, including Diraq and the Israeli Quantum Computing Center (IQCC). The program features NVIDIA DGX Quantum, a reference architecture developed by NVIDIA and QM, designed to integrate quantum and classical computing systems. This solution supports quantum error correction (QEC) and parameter drift compensation, addressing the increasing demand for classical resources as quantum systems scale.
The NVIDIA DGX Quantum platform combines QM’s OPX1000, a modular high-density hybrid control system, with NVIDIA’s GH200 Grace Hopper Superchips. This integration achieves ultra-low round-trip latency of less than 4 µs, enabling real-time QEC decoding, AI-driven quantum processor calibration, and hybrid quantum-classical applications. The solution is scalable, allowing for additional OPX1000 controllers and Grace Hopper servers as needed.
Participants in the program, such as Professor Benjamin Huard of ENS de Lyon and Professor Andre Saraiva of Diraq, emphasize the platform’s potential to accelerate hybrid quantum algorithms and improve quantum system calibration. The IQCC has demonstrated record calibration speeds for single- and two-qubit gates using reinforcement learning agents on Grace Hopper Superchips. QM will showcase NVIDIA DGX Quantum at the APS March Meeting 2025 and present its capabilities at NVIDIA GTC 2025.
For more details, visit the original announcement here.
March 18, 2025
Leave A Comment