[{"data":1,"prerenderedAt":453},["ShallowReactive",2],{"content-query-LRtSwpcKHk":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"date":10,"cover":11,"type":12,"body":13,"_type":447,"_id":448,"_source":449,"_file":450,"_stem":451,"_extension":452},"/news/zh/3821","zh",false,"","MindSpore Quantum 0.11.0 正式发布：性能全面跃升，高效易用，加速前沿量子计算研究","发挥量子计算机的算力，除了量子硬件以外，量子软件和量子算法是必不可缺的组成部分。","2025-08-28","https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2025/08/29/e27d3d97b8304db58f6c1b20716a5247.png","news",{"type":14,"children":15,"toc":444},"root",[16,24,30,35,40,70,83,88,113,129,140,168,176,184,193,203,218,223,242,249,256,261,299,307,322,330,338,346,354,359,367,382,387,392,400,405,416,421,430,435],{"type":17,"tag":18,"props":19,"children":21},"element","h1",{"id":20},"mindspore-quantum-0110-正式发布性能全面跃升高效易用加速前沿量子计算研究",[22],{"type":23,"value":8},"text",{"type":17,"tag":25,"props":26,"children":27},"p",{},[28],{"type":23,"value":29},"量子计算已进入工程化阶段，发展一日千里，已经具备超越经典计算机的潜力，被认为是未来具有颠覆性影响的新型计算模式。要构建一套完整的量子计算机系统，发挥量子计算机的算力，除了量子硬件以外，量子软件和量子算法是必不可缺的组成部分。",{"type":17,"tag":25,"props":31,"children":32},{},[33],{"type":23,"value":34},"Mindspore Quantum是昇思MindSpore社区推出的全新量子经典混合计算框架，能够高效地支持主流的量子算法和经典量子混合算法，不但在量子机器学习、量子化学模拟和量子组合优化等变分量子算法上有着卓越的性能，而且在量子化学模拟规模上也处于领先地位，为量子计算的研究和发展提供了一个高效的开发平台。",{"type":17,"tag":25,"props":36,"children":37},{},[38],{"type":23,"value":39},"MindSpore Quantum的核心优势包括：",{"type":17,"tag":41,"props":42,"children":43},"ul",{},[44,50,55,60,65],{"type":17,"tag":45,"props":46,"children":47},"li",{},[48],{"type":23,"value":49},"高性能模拟：推出CPU和GPU架构优化的高性能量子模拟器，支持单/双精度模拟模式切换，满足不同精度需求。",{"type":17,"tag":45,"props":51,"children":52},{},[53],{"type":23,"value":54},"NISQ算法支持：高效支持量子机器学习、量子化学模拟和量子组合优化等主流变分量子算法。",{"type":17,"tag":45,"props":56,"children":57},{},[58],{"type":23,"value":59},"丰富的量子启发式算法 (QAIA)：内置业界领先的量子启发式优化算法库，包含模拟分岔机 (SB)、相干伊辛机 (SimCIM)、噪声平均场退火 (NMFA) 等多种高性能求解器。这些算法专为高效求解QUBO等大规模组合优化问题而设计，并针对不同硬件（CPU, GPU, NPU）进行了深度性能优化。",{"type":17,"tag":45,"props":61,"children":62},{},[63],{"type":23,"value":64},"简捷易用：提供丰富且高度集成的量子编程接口，降低开发难度，提升开发效率。",{"type":17,"tag":45,"props":66,"children":67},{},[68],{"type":23,"value":69},"深度学习集成：与MindSpore深度结合，让混合量子-经典算法开发更高效。",{"type":17,"tag":25,"props":71,"children":72},{},[73,75,81],{"type":23,"value":74},"经过开发者们几个月的不懈努力，MindSpore Quantum 0.11.0 版本正式上线！本次版本聚焦于核心组件的性能优化与硬件加速，推出了两大关键特性，旨在",{"type":17,"tag":76,"props":77,"children":78},"strong",{},[79],{"type":23,"value":80},"显著提升量子化学计算以及量子启发式算法的运行效率",{"type":23,"value":82},"，从而加速从理论到实践的研发进程。快来一起看看吧！",{"type":17,"tag":25,"props":84,"children":85},{},[86],{"type":23,"value":87},"本次版本更新的核心亮点包括：",{"type":17,"tag":41,"props":89,"children":90},{},[91,108],{"type":17,"tag":45,"props":92,"children":93},{},[94,106],{"type":17,"tag":76,"props":95,"children":96},{},[97,99,104],{"type":23,"value":98},"专用量子化学求解器：",{"type":17,"tag":76,"props":100,"children":101},{},[102],{"type":23,"value":103},"发布",{"type":23,"value":105},"mqchem",{"type":23,"value":107},"模块，在特定化学问题上获得百倍于主流框架的计算速度。",{"type":17,"tag":45,"props":109,"children":110},{},[111],{"type":23,"value":112},"**QAIA 硬件后端支持：**为量子启发式算法（QAIA）新增 GPU 与NPU 后端，大幅提升组合优化问题的求解效率。",{"type":17,"tag":18,"props":114,"children":116},{"id":115},"_01-mqchem面向化学模拟的专用求解器计算效率提升百倍",[117,122,124],{"type":17,"tag":76,"props":118,"children":119},{},[120],{"type":23,"value":121},"# 01",{"type":23,"value":123}," ",{"type":17,"tag":76,"props":125,"children":126},{},[127],{"type":23,"value":128},"mqchem：面向化学模拟的专用求解器，计算效率提升百倍",{"type":17,"tag":25,"props":130,"children":131},{},[132,134,138],{"type":23,"value":133},"量子化学是量子计算最具潜力的应用领域之一。为应对该领域对计算效率的严苛要求，MindSpore Quantum 0.11.0 推出了专为量子化学计算设计的求解器 ",{"type":17,"tag":76,"props":135,"children":136},{},[137],{"type":23,"value":105},{"type":23,"value":139},"。",{"type":17,"tag":25,"props":141,"children":142},{},[143,147,149,153,155,160,162,167],{"type":17,"tag":76,"props":144,"children":145},{},[146],{"type":23,"value":105},{"type":23,"value":148}," 模块通过在固定电子数的构型相互作用（CI）子空间内进行计算，并结合高效的算子处理技术，极大优化了UCCSD（幺正耦合簇）等变分量子算法的计算流程。在模拟 H6、H8、H10、H12 等典型氢链分子时，",{"type":17,"tag":76,"props":150,"children":151},{},[152],{"type":23,"value":105},{"type":23,"value":154}," 的计算速度相较于其他主流量子计算框架实现了",{"type":17,"tag":76,"props":156,"children":157},{},[158],{"type":23,"value":159},"百倍甚至千倍的提升",{"type":23,"value":161},"，并成功完成了在其他框架上难以",{"type":17,"tag":76,"props":163,"children":164},{},[165],{"type":23,"value":166},"在合理时间内完成的更大规模分子的模拟",{"type":23,"value":139},{"type":17,"tag":25,"props":169,"children":170},{},[171],{"type":17,"tag":172,"props":173,"children":175},"img",{"alt":7,"src":174},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2025/08/29/f9e66ecf251045358af5b4460f4bf6e2.png",[],{"type":17,"tag":25,"props":177,"children":178},{},[179],{"type":17,"tag":76,"props":180,"children":181},{},[182],{"type":23,"value":183},"快速上手",{"type":17,"tag":25,"props":185,"children":186},{},[187,191],{"type":17,"tag":76,"props":188,"children":189},{},[190],{"type":23,"value":105},{"type":23,"value":192}," 提供了高度集成的接口，可与 OpenFermion 无缝衔接，几行代码即可完成完整的 VQE 算法模拟。",{"type":17,"tag":194,"props":195,"children":197},"pre",{"code":196},"from openfermionpyscf import run_pyscf\nfrom openfermion import MolecularData\nfrom scipy.optimize import minimize\nfrom mindquantum.simulator import mqchem\n \n# 准备分子数据\nmolecule = MolecularData([(\"H\", (i, 0, 0)) for i in range(6)], 'sto-3g', 1, 0)\nmol = run_pyscf(molecule, run_ccsd=True)\n \n# 使用 mqchem 准备 VQE 任务\nhamiltonian, ansatz_circuit, initial_amplitudes = mqchem.prepare_uccsd_vqe(mol, 1e-3)\n \n# 创建 mqchem 模拟器并获取梯度算子\nsimulator = mqchem.MQChemSimulator(mol.n_qubits, mol.n_electrons)\ngrad_ops = simulator.get_expectation_with_grad(hamiltonian, ansatz_circuit)\n \n# 使用经典优化器求解\nresult = minimize(grad_ops, initial_amplitudes, method='L-BFGS-B', jac=True)\nprint(f\"VQE energy: {result.fun}\")\n",[198],{"type":17,"tag":199,"props":200,"children":201},"code",{"__ignoreMap":7},[202],{"type":23,"value":196},{"type":17,"tag":18,"props":204,"children":206},{"id":205},"_02-qaia新增-gpunpu-后端加速量子启发式算法探索",[207,212,213],{"type":17,"tag":76,"props":208,"children":209},{},[210],{"type":23,"value":211},"# 02",{"type":23,"value":123},{"type":17,"tag":76,"props":214,"children":215},{},[216],{"type":23,"value":217},"QAIA：新增 GPU/NPU 后端，加速量子启发式算法探索",{"type":17,"tag":25,"props":219,"children":220},{},[221],{"type":23,"value":222},"量子退火启发式算法（Quantum Annealing Inspired Algorithm）是解决组合优化问题的有力工具。MindSpore Quantum 的 QAIA 模块提供了一系列高性能的量子启发式算法实现。为了进一步提升求解大规模问题的能力，0.11.0 版本为 QAIA 模块新增了 GPU 和 NPU 后端支持。",{"type":17,"tag":25,"props":224,"children":225},{},[226,228,233,235,240],{"type":23,"value":227},"如下图所示，在处理一个包含 4096 个变量的大规模组合优化问题时，GPU 与 NPU 后端展现出强大的加速能力。相较于 CPU，",{"type":17,"tag":76,"props":229,"children":230},{},[231],{"type":23,"value":232},"GPU 后端可带来数十倍甚至上百倍的性能提升",{"type":23,"value":234},"，将原本需要十几秒的计算任务缩短至 0.2 秒左右。同样，",{"type":17,"tag":76,"props":236,"children":237},{},[238],{"type":23,"value":239},"NPU 后端也实现了显著的加速效果",{"type":23,"value":241},"。例如，对于 ASB 算法，CPU 需要 17.7 秒完成计算，而 GPU 和 NPU 分别仅需 0.25 秒和 0.48 秒。这种性能上的飞跃，使得利用 QAIA 求解更大规模、更具现实意义的组合优化问题成为可能。",{"type":17,"tag":25,"props":243,"children":244},{},[245],{"type":17,"tag":172,"props":246,"children":248},{"alt":7,"src":247},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2025/08/29/87b97bf069784c308cb9539896caf3f4.png",[],{"type":17,"tag":25,"props":250,"children":251},{},[252],{"type":17,"tag":76,"props":253,"children":254},{},[255],{"type":23,"value":183},{"type":17,"tag":25,"props":257,"children":258},{},[259],{"type":23,"value":260},"使用 GPU 后端需要安装 PyTorch 的 GPU 版本；使用 NPU 后端则需要安装 torch-npu。",{"type":17,"tag":25,"props":262,"children":263},{},[264,266,281,283,298],{"type":23,"value":265},"为 QAIA 算法启用硬件加速同样简单，只需在初始化求解器时通过 backend 参数指定硬件平台即可。比 backend=\"",{"type":17,"tag":76,"props":267,"children":268},{},[269,271,279],{"type":23,"value":270},"g",{"type":17,"tag":76,"props":272,"children":273},{},[274],{"type":17,"tag":76,"props":275,"children":276},{},[277],{"type":23,"value":278},"pu-float3",{"type":23,"value":280},"2\"",{"type":23,"value":282}," 或 ",{"type":17,"tag":76,"props":284,"children":285},{},[286,288,296],{"type":23,"value":287},"backen",{"type":17,"tag":76,"props":289,"children":290},{},[291],{"type":17,"tag":76,"props":292,"children":293},{},[294],{"type":23,"value":295},"d=\"n",{"type":23,"value":297},"pu-float****32\"",{"type":23,"value":139},{"type":17,"tag":194,"props":300,"children":302},{"code":301},"\nimport numpy as np\nfrom scipy.sparse import coo_matrix\nfrom mindquantum.algorithm.qaia import BSB\n \n# 适用于 ASB, DSB, LQA, CAC, CFC, SFC, NMFA, SimCIM 等多种算法\nJ = coo_matrix(np.array([[0, -1], [-1, 0]]))\n \n# 通过 backend 参数选择 'gpu-float32' 或 'npu-float32'\nsolver = BSB(J, n_iter=200, batch_size=10, backend=\"gpu-float32\")\nsolver.update()\ncut = solver.calc_cut()\nenergy = solver.calc_energy()\n",[303],{"type":17,"tag":199,"props":304,"children":305},{"__ignoreMap":7},[306],{"type":23,"value":301},{"type":17,"tag":18,"props":308,"children":310},{"id":309},"_03-快速安装与升级指南",[311,316,317],{"type":17,"tag":76,"props":312,"children":313},{},[314],{"type":23,"value":315},"# 03",{"type":23,"value":123},{"type":17,"tag":76,"props":318,"children":319},{},[320],{"type":23,"value":321},"快速安装与升级指南",{"type":17,"tag":25,"props":323,"children":324},{},[325],{"type":17,"tag":76,"props":326,"children":327},{},[328],{"type":23,"value":329},"首次安装 MindSpore Quantum:",{"type":17,"tag":194,"props":331,"children":333},{"code":332},"pip install mindquantum\n",[334],{"type":17,"tag":199,"props":335,"children":336},{"__ignoreMap":7},[337],{"type":23,"value":332},{"type":17,"tag":25,"props":339,"children":340},{},[341],{"type":17,"tag":76,"props":342,"children":343},{},[344],{"type":23,"value":345},"升级至 0.11.0 版本:",{"type":17,"tag":194,"props":347,"children":349},{"code":348},"pip install --upgrade mindquantum\n",[350],{"type":17,"tag":199,"props":351,"children":352},{"__ignoreMap":7},[353],{"type":23,"value":348},{"type":17,"tag":25,"props":355,"children":356},{},[357],{"type":23,"value":358},"注意：",{"type":17,"tag":41,"props":360,"children":361},{},[362],{"type":17,"tag":45,"props":363,"children":364},{},[365],{"type":23,"value":366},"使用 QAIA 的GPU后端需要安装 PyTorch 的 GPU 版本；使用 NPU 后端则需要安装 torch-npu。",{"type":17,"tag":18,"props":368,"children":370},{"id":369},"_04-结语",[371,376,377],{"type":17,"tag":76,"props":372,"children":373},{},[374],{"type":23,"value":375},"# 04",{"type":23,"value":123},{"type":17,"tag":76,"props":378,"children":379},{},[380],{"type":23,"value":381},"结语",{"type":17,"tag":25,"props":383,"children":384},{},[385],{"type":23,"value":386},"MindSpore Quantum 0.11.0 是一次以性能为核心的重大更新。我们相信，这些新特性将为量子计算领域的研究与开发工作提供强大的算力支持。",{"type":17,"tag":25,"props":388,"children":389},{},[390],{"type":23,"value":391},"我们对所有为 MindSpore Quantum 做出贡献的开发者表示诚挚的感谢。欢迎广大用户下载体验新版本，并随时通过 Gitee Issue 向我们反馈宝贵的建议。",{"type":17,"tag":25,"props":393,"children":394},{},[395],{"type":17,"tag":76,"props":396,"children":397},{},[398],{"type":23,"value":399},"参考链接：",{"type":17,"tag":25,"props":401,"children":402},{},[403],{"type":23,"value":404},"[1] MindSpore Quantum代码仓：",{"type":17,"tag":25,"props":406,"children":407},{},[408],{"type":17,"tag":409,"props":410,"children":414},"a",{"href":411,"rel":412},"https://gitee.com/mindspore/mindquantum",[413],"nofollow",[415],{"type":23,"value":411},{"type":17,"tag":25,"props":417,"children":418},{},[419],{"type":23,"value":420},"[2] MindSpore Quantum安装部署、教程、API等更多信息，请查看：",{"type":17,"tag":25,"props":422,"children":423},{},[424],{"type":17,"tag":409,"props":425,"children":428},{"href":426,"rel":427},"https://mindspore.cn/mindquantum/docs/zh-CN/master/index.html",[413],[429],{"type":23,"value":426},{"type":17,"tag":25,"props":431,"children":432},{},[433],{"type":23,"value":434},"[3] MindSpore Quantum白皮书:",{"type":17,"tag":25,"props":436,"children":437},{},[438],{"type":17,"tag":409,"props":439,"children":442},{"href":440,"rel":441},"https://arxiv.org/pdf/2406.17248",[413],[443],{"type":23,"value":440},{"title":7,"searchDepth":445,"depth":445,"links":446},4,[],"markdown","content:news:zh:3821.md","content","news/zh/3821.md","news/zh/3821","md",1776506090001]