<p>/PRNewswire/ -- MicroCloud Hologram Inc. (NASDAQ: HOLO)（&#34;HOLO&#34;或&#34;公司&#34;）是一家技术服务提供商，提出了一种创新的硬件加速技术，将量子张量网络算法转换为可以在现场可编程门阵列（FPGA）上运行的并行计算电路，实现了在经典硬件上高效的量子自旋模型模拟。这一成就为量子物理研究、量子算法验证以及未来量子设备的数字双胞胎模拟提供了一条全新的工程路径。</p>
<p>在量子多体系统的研究中，张量网络（TN）算法是一种极为高效的数值工具。它通过将高维量子态分解为多个低维张量的网络结构，在一定程度上克服了指数状态空间扩展的问题。典型的张量网络模型包括矩阵乘积态（MPS）、投影纠缠对态（PEPS）和多尺度纠缠重整化（MERA）等。这些算法在凝聚态物理、量子相变、量子自旋模型模拟等方面发挥着基础性作用。</p>
<p>然而，当我们希望提高系统表征的精度并引入更高的纠缠自由度时，张量的维度和连接性急剧增长，导致计算复杂性迅速从多项式级别跨越到指数级别。以二维自旋系统为例，当纠缠秩从χ=8 扩展到χ=32 时，每次迭代的浮点运算次数几乎增加了百倍，存储带宽和内存访问延迟成为瓶颈。这种指数爆炸特性使得即使是高端 CPU 和 GPU 平台也难以在合理的时间内完成模拟任务。</p>
<p>为此，HOLO 尝试突破传统处理器架构的限制，探索在硬件层面上算法重构和逻辑映射的可行路径。现场可编程门阵列（FPGA）凭借其可重构性、并行性和低延迟特性，为张量网络计算提供了新的可能性。通过将核心计算模块（如张量收缩、张量展开、矩阵乘法 - 加法操作等）直接映射到硬件电路的逻辑层面，可以大幅减少内存访问消耗和控制开销，实现深度流水线的高密度并行计算。</p>
<p>HOLO 技术的核心在于算法 - 硬件协同设计，将张量网络算法从软件逻辑层面剖析为可以直接硬件化的计算单元，并以 FPGA 为载体构建高密度并行可扩展架构。</p>
<p>首先，对量子自旋模型的张量网络结构进行了系统分析。以海森堡自旋链和二维伊辛模型为代表的典型系统，其哈密顿量可分解为局部相互作用项，通过张量网络编码为多个局部张量。每个张量节点的收缩本质上对应于张量乘积、矩阵乘法和求和操作。传统的 CPU 计算依赖于通用指令集的顺序执行，而 GPU 虽然具有大规模并行性，但受限于内存访问延迟和内核调度，难以实现针对性的优化。FPGA 架构允许在硬件层面直接定义这些计算逻辑，消除冗余调度环节，使数据能够在片上高速缓存中以流水线方式连续流动。</p>
<p>在实现过程中，HOLO 构建了一个分层张量收缩流水线。该流水线包括三个主要层级：</p>
<p>输入和调度层：负责将高维张量分解为多个可管理的块结构，并进行数据流调度和依赖分析。</p>
<p>核心计算层：由多个 MAC 阵列组成，支持任意维度的张量收缩操作。每个计算单元采用定制逻辑，实现浮点加法和乘法的流水线级并行。</p>
<p>输出和归约层：执行张量结果的合并、归一化和中间状态缓存，为后续迭代提供输入。</p>
<p>在硬件逻辑设计中，通过 Verilog 和高级综合（HLS）工具的结合方法，自动生成张量操作电路，并针对不同的张量连接图采用多分区策略。通过静态调度和数据重用机制，计算单元在片上形成高密度并行阵列，在有限的逻辑资源下实现最大计算吞吐率。</p>
<p>该技术以 FPGA 为核心硬件平台，提出并实现了一种加速量子张量网络计算的并行化硬件架构。通过算法结构重构、逻辑电路映射、流水线设计和混合精度优化，HOLO 成功将复杂的张量网络计算任务转化为高效的 FPGA 逻辑操作，实现了比 CPU 快 1.7 倍的性能突破，能效提高超过 2 倍。这项技术不仅展示了 FPGA 在量子模拟中的潜力，还为量子算法硬件实现和可重构量子加速器设计提供了实际基础。</p>
<p>未来，HOLO 将继续沿着从算法到电路的设计理念，推动更多量子计算核心模块的硬件实现，包括量子变分算法（VQE）、量子线性系统求解器（QLSA）以及量子机器学习模型的 FPGA 化，以构建完整的量子算法加速生态系统。相信通过在这一方向的持续研究，FPGA 将成为量子计算与经典计算之间的重要桥梁，为量子技术的产业化发展提供坚实的技术支持。</p>
<p>关于 MicroCloud Hologram Inc.</p>
<p>MicroCloud Hologram Inc.（纳斯达克：HOLO）致力于全息技术的研发和应用。其全息技术服务包括基于全息技术的全息光检测与测距（LiDAR）解决方案、全息 LiDAR 点云算法架构设计、技术全息成像解决方案、全息 LiDAR 传感器芯片设计以及全息车辆智能视觉技术，为提供全息高级驾驶辅助系统（ADAS）的客户提供服务。MicroCloud Hologram Inc.为全球客户提供全息技术服务。MicroCloud Hologram Inc.还提供全息数字双胞胎技术服务，并拥有专有的全息数字双胞胎技术资源库。其全息数字双胞胎技术资源库结合了全息数字双胞胎软件、数字内容、空间数据驱动的数据科学、全息数字云算法和全息 3D 捕捉技术，以 3D 全息形式捕捉形状和物体。MicroCloud Hologram Inc.专注于量子计算和量子全息等领域的发展，现金储备超过 30 亿元人民币，并计划从现金储备中投资超过 4 亿美元，参与区块链开发、量子计算技术开发、量子全息技术开发以及人工智能 AR 等前沿技术领域的衍生品和技术开发。MicroCloud Hologram Inc.的目标是成为全球领先的量子全息和量子计算技术公司。</p>
<p>安全港声明</p>
<p>本新闻稿包含根据 1995 年《私人证券诉讼改革法案》定义的前瞻性陈述。前瞻性陈述包括有关计划、目标、策略、未来事件或表现及其基本假设的陈述，以及其他非历史事实的陈述。当公司使用诸如 “可能”、“将”、“打算”、“应该”、“相信”、“期望”、“预期”、“项目”、“估计” 或类似表达时，这些表达并不单纯与历史事项相关，而是在进行前瞻性陈述。前瞻性陈述并不保证未来的表现，并涉及可能导致实际结果与公司在前瞻性陈述中讨论的预期存在重大差异的风险和不确定性。这些陈述受到不确定性和风险的影响，包括但不限于以下内容：公司的目标和策略；公司的未来业务发展；产品和服务的需求和接受度；技术变化；经济状况；声誉和品牌；竞争和定价的影响；政府法规；一般经济波动；财务状况和经营结果；全息行业的预期增长以及公司计划服务的中国和国际市场的商业条件，以及与上述任何内容相关的假设和公司向证券交易委员会（“SEC”）提交的报告中包含的其他风险，包括公司最近提交的 10-K 表格年度报告和 6-K 表格当前报告及其后续提交的文件。因此，投资者被提醒不要对本新闻稿中的任何前瞻性陈述过度依赖。其他因素在公司向 SEC 提交的文件中进行了讨论，这些文件可在 www.sec.gov 上查阅。公司没有义务公开修订这些前瞻性陈述，以反映本日期后发生的事件或情况。</p>
<p>来源：MicroCloud Hologram Inc.</p>

HOLO

<p>MicroCloud Hologram Inc.（纳斯达克代码：HOLO）推出了一种利用现场可编程门阵列（FPGA）的硬件加速技术，以增强量子张量网络计算。这项创新使得在经典硬件上高效模拟量子自旋模型成为可能，解决了量子多体系统中计算复杂性的问题。通过将核心计算任务直接映射到硬件电路中，HOLO 的技术在性能上比 CPU 提高了 1.7 倍，并且将能效翻倍。该公司旨在进一步开发量子计算模块，包括量子变分算法和量子机器学习模型</p>

<p>, /PRNewswire/ -- MicroCloud Hologram Inc. (NASDAQ: HOLO), (&#34;HOLO&#34; or the &#34;Company&#34;), a technology service provider, proposed an innovative hardware acceleration technology that converts the quantum tensor network algorithm into parallel computing circuits that can run on field programmable gate arrays (FPGA), achieving efficient quantum spin model simulation on classical hardware. This achievement provides a brand-new engineered path for quantum physics research, quantum algorithm verification, and digital twin simulation of future quantum devices.</p>
<div class="lb-trans"><p>/PRNewswire/ -- MicroCloud Hologram Inc. (NASDAQ: HOLO)（&#34;HOLO&#34;或&#34;公司&#34;）是一家技术服务提供商，提出了一种创新的硬件加速技术，将量子张量网络算法转换为可以在现场可编程门阵列（FPGA）上运行的并行计算电路，实现了在经典硬件上高效的量子自旋模型模拟。这一成就为量子物理研究、量子算法验证以及未来量子设备的数字双胞胎模拟提供了一条全新的工程路径。</p>
</div><p>In the research of quantum many-body systems, the tensor network (TN) algorithm is an extremely efficient numerical tool. It overcomes the problem of exponential state space expansion to a certain extent by decomposing high-dimensional quantum states into a network structure of multiple low-dimensional tensors. Typical tensor network models include matrix product states (MPS), projected entangled pair states (PEPS), and multi-scale entanglement renormalization (MERA), etc. These algorithms play a foundational role in condensed matter physics, quantum phase transitions, quantum spin model simulations, and other aspects.</p>
<div class="lb-trans"><p>在量子多体系统的研究中，张量网络（TN）算法是一种极为高效的数值工具。它通过将高维量子态分解为多个低维张量的网络结构，在一定程度上克服了指数状态空间扩展的问题。典型的张量网络模型包括矩阵乘积态（MPS）、投影纠缠对态（PEPS）和多尺度纠缠重整化（MERA）等。这些算法在凝聚态物理、量子相变、量子自旋模型模拟等方面发挥着基础性作用。</p>
</div><p>However, when we hope to improve the precision of system characterization and introduce higher entanglement degrees of freedom, the dimensions and connectivity of tensors grow sharply, causing the computational complexity to rapidly cross from polynomial to exponential levels. Taking a two-dimensional spin system as an example, when the entanglement rank expands from χ=8 to χ=32, the floating-point operations per iteration increase nearly a hundredfold, and storage bandwidth and memory access latency become bottlenecks. This exponential explosion characteristic makes it difficult even for high-end CPU and GPU platforms to complete simulation tasks within a reasonable time.</p>
<div class="lb-trans"><p>然而，当我们希望提高系统表征的精度并引入更高的纠缠自由度时，张量的维度和连接性急剧增长，导致计算复杂性迅速从多项式级别跨越到指数级别。以二维自旋系统为例，当纠缠秩从χ=8 扩展到χ=32 时，每次迭代的浮点运算次数几乎增加了百倍，存储带宽和内存访问延迟成为瓶颈。这种指数爆炸特性使得即使是高端 CPU 和 GPU 平台也难以在合理的时间内完成模拟任务。</p>
</div><p>To this end, HOLO attempts to break out of the limitations of traditional processor architectures and explore feasible paths for algorithm reconstruction and logic mapping at the hardware level. Field programmable gate arrays (FPGA), with their reconfigurability, parallelism, and low-latency characteristics, provide new possibilities for tensor network computations. By directly mapping core computational modules (such as tensor contraction, tensor unfolding, matrix multiplication-addition operations, etc.) into hardware circuits at the logic level, it can greatly reduce memory access consumption and control overhead, achieving deep pipelined high-density parallel computing.</p>
<div class="lb-trans"><p>为此，HOLO 尝试突破传统处理器架构的限制，探索在硬件层面上算法重构和逻辑映射的可行路径。现场可编程门阵列（FPGA）凭借其可重构性、并行性和低延迟特性，为张量网络计算提供了新的可能性。通过将核心计算模块（如张量收缩、张量展开、矩阵乘法 - 加法操作等）直接映射到硬件电路的逻辑层面，可以大幅减少内存访问消耗和控制开销，实现深度流水线的高密度并行计算。</p>
</div><p>The core of HOLO's technology lies in algorithm-hardware co-design, which dissects the tensor network algorithm from the software logic level into computational units that can be directly hardware-ized, and builds a high-density parallel scalable architecture with FPGA as the carrier.</p>
<div class="lb-trans"><p>HOLO 技术的核心在于算法 - 硬件协同设计，将张量网络算法从软件逻辑层面剖析为可以直接硬件化的计算单元，并以 FPGA 为载体构建高密度并行可扩展架构。</p>
</div><p>First, a systematic analysis was performed on the tensor network structure of quantum spin models. Typical systems represented by the Heisenberg spin chain and the two-dimensional Ising model have their Hamiltonians decomposable into local interaction terms, encoded into several local tensors through tensor networks. The contraction of each tensor node essentially corresponds to tensor product, matrix multiplication, and summation operations. Traditional CPU computations rely on sequential execution of general instruction sets, while GPUs, although possessing large-scale parallelism, are limited by memory access latency and kernel scheduling, making it difficult to achieve targeted optimization. The FPGA architecture allows direct definition of these computational logics at the hardware level, eliminating redundant scheduling links, enabling data to flow continuously in a pipeline manner in the on-chip high-speed cache.</p>
<div class="lb-trans"><p>首先，对量子自旋模型的张量网络结构进行了系统分析。以海森堡自旋链和二维伊辛模型为代表的典型系统，其哈密顿量可分解为局部相互作用项，通过张量网络编码为多个局部张量。每个张量节点的收缩本质上对应于张量乘积、矩阵乘法和求和操作。传统的 CPU 计算依赖于通用指令集的顺序执行，而 GPU 虽然具有大规模并行性，但受限于内存访问延迟和内核调度，难以实现针对性的优化。FPGA 架构允许在硬件层面直接定义这些计算逻辑，消除冗余调度环节，使数据能够在片上高速缓存中以流水线方式连续流动。</p>
</div><p>In the implementation, HOLO constructed a Hierarchical Tensor Contraction Pipeline. This pipeline includes three main levels:</p>
<div class="lb-trans"><p>在实现过程中，HOLO 构建了一个分层张量收缩流水线。该流水线包括三个主要层级：</p>
</div><p>Input and scheduling layer: responsible for decomposing high-dimensional tensors into several manageable block structures, and performing data flow scheduling and dependency analysis.</p>
<div class="lb-trans"><p>输入和调度层：负责将高维张量分解为多个可管理的块结构，并进行数据流调度和依赖分析。</p>
</div><p>Core computing layer: composed of multiple MAC Array, supporting tensor contraction operations of arbitrary dimensions. Each computing unit adopts customized logic to achieve pipeline-level parallelism for floating-point addition and multiplication.</p>
<div class="lb-trans"><p>核心计算层：由多个 MAC 阵列组成，支持任意维度的张量收缩操作。每个计算单元采用定制逻辑，实现浮点加法和乘法的流水线级并行。</p>
</div><p>Output and reduction layer: executes the merging, normalization, and intermediate state caching of tensor results, providing input for subsequent iterations.</p>
<div class="lb-trans"><p>输出和归约层：执行张量结果的合并、归一化和中间状态缓存，为后续迭代提供输入。</p>
</div><p>In the hardware logic design, through the combined method of Verilog and high-level synthesis (HLS) tools, tensor operation circuits are automatically generated, and multi-partition strategies are adopted for different tensor connectivity graphs. Through static scheduling and data reuse mechanisms, the computing units form a high-density parallel array on-chip, achieving the maximum computational throughput rate under limited logic resources.</p>
<div class="lb-trans"><p>在硬件逻辑设计中，通过 Verilog 和高级综合（HLS）工具的结合方法，自动生成张量操作电路，并针对不同的张量连接图采用多分区策略。通过静态调度和数据重用机制，计算单元在片上形成高密度并行阵列，在有限的逻辑资源下实现最大计算吞吐率。</p>
</div><p>This technology takes FPGA as the core hardware platform, proposing and implementing a parallelized hardware architecture for accelerating quantum tensor network computations. Through algorithm structure reconstruction, logic circuit mapping, pipelined design, and mixed-precision optimization, HOLO successfully transforms complex tensor network computational tasks into efficient FPGA logic operations, achieving a performance breakthrough that is 1.7 times faster than CPU and with energy efficiency improved by more than 2 times. This technology not only demonstrates the potential of FPGA in quantum simulation but also provides practical basis for quantum algorithm hardware implementation and reconfigurable quantum accelerator design.</p>
<div class="lb-trans"><p>该技术以 FPGA 为核心硬件平台，提出并实现了一种加速量子张量网络计算的并行化硬件架构。通过算法结构重构、逻辑电路映射、流水线设计和混合精度优化，HOLO 成功将复杂的张量网络计算任务转化为高效的 FPGA 逻辑操作，实现了比 CPU 快 1.7 倍的性能突破，能效提高超过 2 倍。这项技术不仅展示了 FPGA 在量子模拟中的潜力，还为量子算法硬件实现和可重构量子加速器设计提供了实际基础。</p>
</div><p>In the future, HOLO will continue along the design philosophy from algorithm to circuit, promoting the hardware implementation of more quantum computing core modules, including quantum variational algorithms (VQE), quantum linear system solvers (QLSA), and FPGA-ization of quantum machine learning models, to build a complete quantum algorithm acceleration ecosystem. It is believed that, through continuous research in this direction, FPGA will become an important bridge between quantum computing and classical computing, providing solid technical support for the industrialization development of quantum technology.</p>
<div class="lb-trans"><p>未来，HOLO 将继续沿着从算法到电路的设计理念，推动更多量子计算核心模块的硬件实现，包括量子变分算法（VQE）、量子线性系统求解器（QLSA）以及量子机器学习模型的 FPGA 化，以构建完整的量子算法加速生态系统。相信通过在这一方向的持续研究，FPGA 将成为量子计算与经典计算之间的重要桥梁，为量子技术的产业化发展提供坚实的技术支持。</p>
</div><p>About MicroCloud Hologram Inc.</p>
<div class="lb-trans"><p>关于 MicroCloud Hologram Inc.</p>
</div><p>MicroCloud Hologram Inc. (NASDAQ: HOLO) is committed to the research and development and application of holographic technology. Its holographic technology services include holographic light detection and ranging (LiDAR) solutions based on holographic technology, holographic LiDAR point cloud algorithm architecture design, technical holographic imaging solutions, holographic LiDAR sensor chip design, and holographic vehicle intelligent vision technology, providing services to customers offering holographic advanced driving assistance systems (ADAS). MicroCloud Hologram Inc. provides holographic technology services to global customers. MicroCloud Hologram Inc. also provides holographic digital twin technology services and owns proprietary holographic digital twin technology resource libraries. Its holographic digital twin technology resource library utilizes a combination of holographic digital twin software, digital content, space data-driven data science, holographic digital cloud algorithms, and holographic 3D capture technology to capture shapes and objects in 3D holographic form. MicroCloud Hologram Inc. focuses on developments such as quantum computing and quantum holography, with cash reserves exceeding 3 billion RMB, and plans to invest more than 400 million in USD from the cash reserves to engage in blockchain development, quantum computing technology development, quantum holography technology development, and derivatives and technology development in frontier technology fields such as artificial intelligence AR. MicroCloud Hologram Inc.'s goal is to become a global leading quantum holography and quantum computing technology company.</p>
<div class="lb-trans"><p>MicroCloud Hologram Inc.（纳斯达克：HOLO）致力于全息技术的研发和应用。其全息技术服务包括基于全息技术的全息光检测与测距（LiDAR）解决方案、全息 LiDAR 点云算法架构设计、技术全息成像解决方案、全息 LiDAR 传感器芯片设计以及全息车辆智能视觉技术，为提供全息高级驾驶辅助系统（ADAS）的客户提供服务。MicroCloud Hologram Inc.为全球客户提供全息技术服务。MicroCloud Hologram Inc.还提供全息数字双胞胎技术服务，并拥有专有的全息数字双胞胎技术资源库。其全息数字双胞胎技术资源库结合了全息数字双胞胎软件、数字内容、空间数据驱动的数据科学、全息数字云算法和全息 3D 捕捉技术，以 3D 全息形式捕捉形状和物体。MicroCloud Hologram Inc.专注于量子计算和量子全息等领域的发展，现金储备超过 30 亿元人民币，并计划从现金储备中投资超过 4 亿美元，参与区块链开发、量子计算技术开发、量子全息技术开发以及人工智能 AR 等前沿技术领域的衍生品和技术开发。MicroCloud Hologram Inc.的目标是成为全球领先的量子全息和量子计算技术公司。</p>
</div><p>Safe Harbor Statement</p>
<div class="lb-trans"><p>安全港声明</p>
</div><p>This press release contains forward-looking statements as defined by the Private Securities Litigation Reform Act of 1995. Forward-looking statements include statements concerning plans, objectives, goals, strategies, future events or performance, and underlying assumptions and other statements that are other than statements of historical facts. When the Company uses words such as &#34;may,&#34; &#34;will,&#34; &#34;intend,&#34; &#34;should,&#34; &#34;believe,&#34; &#34;expect,&#34; &#34;anticipate,&#34; &#34;project,&#34; &#34;estimate,&#34; or similar expressions that do not relate solely to historical matters, it is making forward-looking statements. Forward-looking statements are not guarantees of future performance and involve risks and uncertainties that may cause the actual results to differ materially from the Company's expectations discussed in the forward-looking statements. These statements are subject to uncertainties and risks including, but not limited to, the following: the Company's goals and strategies; the Company's future business development; product and service demand and acceptance; changes in technology; economic conditions; reputation and brand; the impact of competition and pricing; government regulations; fluctuations in general economic; financial condition and results of operations; the expected growth of the holographic industry and business conditions in China and the international markets the Company plans to serve and assumptions underlying or related to any of the foregoing and other risks contained in reports filed by the Company with the Securities and Exchange Commission (&#34;SEC&#34;), including the Company's most recently filed Annual Report on Form 10-K and current report on Form 6-K and its subsequent filings. For these reasons, among others, investors are cautioned not to place undue reliance upon any forward-looking statements in this press release. Additional factors are discussed in the Company's filings with the SEC, which are available for review at www.sec.gov. The Company undertakes no obligation to publicly revise these forward-looking statements to reflect events or circumstances that arise after the date hereof.</p>
<div class="lb-trans"><p>本新闻稿包含根据 1995 年《私人证券诉讼改革法案》定义的前瞻性陈述。前瞻性陈述包括有关计划、目标、策略、未来事件或表现及其基本假设的陈述，以及其他非历史事实的陈述。当公司使用诸如 “可能”、“将”、“打算”、“应该”、“相信”、“期望”、“预期”、“项目”、“估计” 或类似表达时，这些表达并不单纯与历史事项相关，而是在进行前瞻性陈述。前瞻性陈述并不保证未来的表现，并涉及可能导致实际结果与公司在前瞻性陈述中讨论的预期存在重大差异的风险和不确定性。这些陈述受到不确定性和风险的影响，包括但不限于以下内容：公司的目标和策略；公司的未来业务发展；产品和服务的需求和接受度；技术变化；经济状况；声誉和品牌；竞争和定价的影响；政府法规；一般经济波动；财务状况和经营结果；全息行业的预期增长以及公司计划服务的中国和国际市场的商业条件，以及与上述任何内容相关的假设和公司向证券交易委员会（“SEC”）提交的报告中包含的其他风险，包括公司最近提交的 10-K 表格年度报告和 6-K 表格当前报告及其后续提交的文件。因此，投资者被提醒不要对本新闻稿中的任何前瞻性陈述过度依赖。其他因素在公司向 SEC 提交的文件中进行了讨论，这些文件可在 www.sec.gov 上查阅。公司没有义务公开修订这些前瞻性陈述，以反映本日期后发生的事件或情况。</p>
</div><p>SOURCE MicroCloud Hologram Inc.</p>
<div class="lb-trans"><p>来源：MicroCloud Hologram Inc.</p>
</div>

MicroCloud Hologram Inc. 利用 FPGA 加速张量网络计算以实现量子自旋模型 | HOLO 股票新闻