FPGA布局问题中的量子退火方法
FPGA-Placement via Quantum Annealing
AI总结:
本文提出将FPGA布局这一NP难问题转化为一系列QUBO问题,利用绝热量子计算求解,并在D-Wave Advantage 5.4上验证了其有效性。
摘要:
现场可编程门阵列(FPGA)通过提供灵活、可重构的硬件平台,已成为当代计算中的重要资产。基于FPGA的加速器为实时数据处理、机器学习或密码学等领域(仅举几例)的突破性进展孕育了机会。布局过程——确定FPGA上功能块的最优空间排列以最小化通信延迟并提升性能——是一个NP难问题,尤其需要复杂的算法才能高效求解。显然,改进布局可以在实现阶段降低资源利用率。绝热量子计算(AQC)凭借其遍历广阔解空间的能力,有潜力解决此类组合问题。在本文中,我们将布局问题重新表述为一系列所谓的二次无约束二元优化(QUBO)问题,随后通过AQC求解。我们的新颖表述便于直接集成设计约束。此外,子问题的规模可以方便地根据可用硬件能力进行调整。除了提出一种新方法外,我们还探究当代量子硬件是否足够稳健,能够为实际规模的FPGA找到布局。在D-Wave Advantage 5.4量子退火器上的数值评估表明答案是肯定的。
英文摘要:
Field-Programmable Gate Arrays (FPGAs) have asserted themselves as vital assets in contemporary computing by offering adaptable, reconfigurable hardware platforms. FPGA-based accelerators incubate opportunities for breakthroughs in areas, such as real-time data processing, machine learning or cryptography -- to mention just a few. The procedure of placement -- determining the optimal spatial arrangement of functional blocks on an FPGA to minimize communication delays and enhance performance -- is an NP-hard problem, notably requiring sophisticated algorithms for proficient solutions. Clearly, improving the placement leads to a decreased resource utilization during the implementation phase. Adiabatic quantum computing (AQC), with its capability to traverse expansive solution spaces, has potential for addressing such combinatorial problems. In this paper, we re-formulate the placement problem as a series of so called quadratic unconstrained binary optimization (QUBO) problems which are subsequently solved via AQC. Our novel formulation facilitates a straight-forward integration of design constraints. Moreover, the size of the sub-problems can be conveniently adapted to the available hardware capabilities. Beside the sole proposal of a novel method, we ask whether contemporary quantum hardware is resilient enough to find placements for real-world-sized FPGAs. A numerical evaluation on a D-Wave Advantage 5.4 quantum annealer suggests that the answer is in the affirmative.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。