QLEO quantum emulator gains GPU boost & full CUDA-Q support
Quobly and QPerfect have unveiled a major upgrade to their quantum emulator, QLEO, adding GPU acceleration and full compatibility with NVIDIA CUDA-Q. This update aims to provide researchers and developers with enhanced tools for simulating quantum algorithms at increased speeds and scales within current computing infrastructure.
Technical enhancements
The new version of QLEO integrates GPU acceleration using the NVIDIA cuQuantum SDK. This development means users can simulate quantum circuits much more efficiently, benefitting from hardware already widely deployed in data centres and on cloud platforms. The update also introduces native support for CUDA-Q. Developers can now write quantum circuits in CUDA-Q and run them on QLEO without needing to change any code.
The platform leverages QPerfect's MIMIQ engine for a robust simulation environment. This backend is designed to deliver over 100 times speed improvement when compared to CPU-only simulation. Users can deploy their workloads using either GPU or CPU, and the direct CUDA-Q integration allows circuits to be implemented within C++ or Python, streamlining the workflow for both industry and academia.
"With native CUDA-Q support, QLEO now provides a direct on-ramp to NVIDIA GPU infrastructure, empowering our community to accelerate their hybrid quantum-classical workflows", said Guido Masella, CTO & Co-Founder, QPerfect. "By enabling GPU acceleration, users can now run quantum simulations an order of magnitude faster on GPU than on CPU, all while leveraging the advanced features and optimisations of the MIMIQ backend."
Industry applications
The enhanced QLEO platform targets a broad spectrum of applications. Quantum circuit simulation is essential for developing algorithms in quantum chemistry, optimisation, materials science, quantitative finance, pharmaceuticals, and artificial intelligence. The significant performance improvements afforded by GPU acceleration are expected to widen the reach of quantum algorithm exploration.
The new release extends access to advanced quantum simulation by supporting existing NVIDIA GPU infrastructure, available on many enterprise and cloud platforms. This aims to ensure researchers and developers can scale their simulations according to the tools and hardware already at their disposal, whether working individually or across larger teams.
User experience
The new QLEO offering is designed to provide a seamless experience for developers and testers of quantum algorithms. By integrating with Python and C++ and supporting both CPU and GPU backends, the platform enables rapid prototyping and iteration. This also facilitates the design of hybrid quantum-classical applications, critical for current research and industry use cases.
"By integrating GPU acceleration and full CUDA-Q compatibility, we are empowering our community to explore quantum-classical workflows at unprecedented scale and speed, a key step toward bridging emulation, cloud access, and future hardware. This aligns perfectly with Quobly's industrialization strategy, shortening the path from emulation to a fully operational quantum machine", said Maud Vinet, CEO and Co-Founder, Quobly.
The QLEO platform is now available to users on the OVHcloud Quantum Platform and can also be installed locally on users' own machines. This approach is intended to lower adoption barriers, enabling engagement from a global user base with varying hardware capabilities.
