share_log

Marvell Introduces 1.6 Tbps LPO Chipset to Enable Optical Short-reach, Scale-up Compute Fabric Interconnects

Marvell Introduces 1.6 Tbps LPO Chipset to Enable Optical Short-reach, Scale-up Compute Fabric Interconnects

Marvell推出1.6 Tbps LPO芯片组,以支持光短距离、可扩展计算布线互连。
迈威尔科技 ·  2024/12/10 13:00

Addresses the Speed and Reach Limitations of Passive, Direct-attached Copper (DAC) Cables

解决了被动直连铜缆 (DAC) 电缆的速度和距离限制

SANTA CLARA, Calif., Dec. 10, 2024 /PRNewswire/ -- Marvell Technology, Inc. (NASDAQ: MRVL), a leader in data infrastructure semiconductor solutions, today announced the general availability of a 200G per lane optimized transimpedance amplifier (TIA) and laser driver chipset, enabling 800 Gbps and 1.6 Tbps linear-drive pluggable optics (LPO). Designed to address next-generation short-reach, scale-up compute fabric connectivity requirements, LPO modules enabled by the chipset overcome the reach limitations of passive, DAC cable interconnects. The LPO chipset expands the industry-leading interconnect portfolio from Marvell, which includes PAM4 optical DSPs, coherent DSPs, data center interconnects, Alaska A active electrical cable (AEC) DSPs and Alaska P PCIe retimers, delivering an optimized optical solution for short-reach compute fabric connections, delivering an optimized optical solution for short-reach compute fabric connections.

加利福尼亚州圣克拉拉,2024年12月10日 /PRNewswire/ — 数据基础设施半导体解决方案领域的领导者Marvell科技公司(纳斯达克股票代码:MRVL)今天宣布全面推出每通道200G的优化跨阻放大器(TIA)和激光驱动器芯片组,支持800 Gbps和1.6 Tbps的线性驱动可插拔光学器件(LPO)。该芯片组支持的 LPO 模块专为满足下一代短距离、纵向扩展的计算架构连接要求而设计,克服了被动 DAC 电缆互连的覆盖范围限制。LPO 芯片组扩展了 Marvell 业界领先的互连产品组合,包括 PAM4 光学 DSP、相干数字信号处理器、数据中心互连、阿拉斯加 A 有源电缆 (AEC) DSP 和 Alaska P PCIe 重定时器,为短距离计算架构连接提供了优化的光学解决方案,为短距离计算架构连接提供了优化的光学解决方案。

As artificial intelligence (AI) technologies advance, the demand for higher-bandwidth interconnects in data center networks is accelerating rapidly. This is particularly evident in compute fabric networks, which connect XPUs within and across racks. The next generation of XPU compute fabric networks will transition to data rates of 200 Gbps per lane, where passive DACs fall short of meeting speed and distance requirements.

随着人工智能 (AI) 技术的进步,数据中心网络中对更高带宽互连的需求正在迅速加速。这在计算架构网络中尤为明显,该网络将机架内和机架之间的 XPU 连接起来。下一代 XPU 计算架构网络将过渡到每通道 200 Gbps 的数据速率,而被动式 DAC 无法满足速度和距离要求。

To address this, cloud data centers will transition to a new type of interconnect that meets their specific requirements. Marvell introduced Alaska A for customers looking to extend copper capabilities using AECs, while others can leverage specialized LPO modules featuring the Marvell TIA and driver chipset. Designed for short and predictable host channels, these LPO modules enable longer reach, higher bandwidth and improved performance compared to copper interconnects.

为了解决这个问题,云数据中心将过渡到满足其特定要求的新型互连。Marvell为希望使用AEC扩展铜缆能力的客户推出了Alaska A,而其他客户则可以利用配备Marvell TIA和驱动器芯片组的专用 LPO 模块。这些 LPO 模块专为较短且可预测的主机通道而设计,与铜质互连相比,可实现更长的覆盖范围、更高的带宽和更高的性能。

"Marvell 1.6 Tbps LPO TIA and laser driver chipset is designed to address the growing demand for short-reach, high-bandwidth interconnect solutions, where passive copper cables are hitting a wall," said Xi Wang, vice president of product marketing for Optical Connectivity at Marvell. "As AI-driven data centers continue to scale, optimizing interconnect solutions across each layer of the network is becoming increasingly critical. The new LPO chipset complements and expands our industry-leading 1.6 Tbps connectivity portfolio, to address the growing spectrum of interconnects that cloud operators are seeking to optimize."

Marvell光学连接产品营销副总裁王希表示:“Marvell 1.6 Tbps LPO TIA和激光驱动器芯片组旨在满足对短距离、高带宽互连解决方案不断增长的需求,在这些解决方案中,无源铜缆会碰壁。“随着人工智能驱动的数据中心的持续扩展,优化网络各层的互连解决方案变得越来越重要。新的LPO芯片组补充并扩展了我们行业领先的1.6 Tbps连接产品组合,以满足云运营商寻求优化的不断增长的互连范围。”

"LPO has been a technology in search of the right solution. By optimizing chipsets for short, inside-the-rack connections, Marvell brings clarity and focus to LPO, delivering it in a more compelling and scalable manner," said Alan Weckel, co-founder of 650 Group. "Marvell's innovative approach to achieving performance gains helps drive better AI cluster TCO and highlights the industry's direction in optimizing networking links."

“LPO一直是寻找正确解决方案的技术。通过优化机架内短连接的芯片组,Marvell为LPO带来了清晰度和专注力,以更具吸引力和可扩展性的方式交付它。” 650集团联合创始人艾伦·韦克尔说。“Marvell实现性能提升的创新方法有助于提高人工智能集群的总体拥有成本,并突显了行业在优化网络链接方面的方向。”

The 1.6 Tbps LPO chipset, one of the latest additions to the Marvell interconnect portfolio, is optimized for specific use cases to help data centers maximize infrastructure utilization and performance while reducing overall cost and power per bit. This extensive portfolio spanning optical and copper interconnects includes Ara, the industry's first 3nm PAM4 interconnect platform; Aquila, the industry's first O-band-optimized coherent-lite DSP platform; Nova family of PAM4 DSPs featuring 200 Gbps electrical and optical interfaces; and Alaska A PAM4 DSP for active electrical cables.

1.6 Tbps LPO 芯片组是 Marvell 互连产品组合的最新产品之一,针对特定用例进行了优化,可帮助数据中心最大限度地提高基础设施利用率和性能,同时降低总体成本和每位功耗。涵盖光学和铜互连的广泛产品组合包括业界首款3纳米PAM4互连平台Ara;业界首款经过O波段优化的相干精简DSP平台Aquila;具有200 Gbps电气和光学接口的Nova系列PAM4 DSP;以及用于有源电缆的阿拉斯加A PAM4 DSP。

LPO Chipset Key Features

LPO 芯片组主要特性

  • TIA provides best-in-class linearity, power and BER for AI applications.
  • Laser driver to improve module performance margin while reducing overall transceiver module design complexity, power consumption, and TCO.
  • TIA and laser driver chipset provide adjustable equalization to compensate for channel loss.
  • TIA 为 AI 应用提供一流的线性度、功率和误码率。
  • 激光驱动器可提高模块性能裕度,同时降低收发器模块的整体设计复杂性、功耗和总拥有成本。
  • TIA 和激光驱动器芯片组提供可调均衡以补偿信道损耗。

About Marvell
To deliver the data infrastructure technology that connects the world, we're building solutions on the most powerful foundation: our partnerships with our customers. Trusted by the world's leading technology companies for over 25 years, we move, store, process and secure the world's data with semiconductor solutions designed for our customers' current needs and future ambitions. Through a process of deep collaboration and transparency, we're ultimately changing the way tomorrow's enterprise, cloud, automotive, and carrier architectures transform—for the better.

关于 Marvell
为了提供连接世界的数据基础设施技术,我们在最强大的基础上构建解决方案:我们与客户的合作伙伴关系。25 年来,我们深受全球领先科技公司的信赖,我们使用专为客户当前需求和未来抱负而设计的半导体解决方案来移动、存储、处理和保护世界数据。通过深度协作和透明的过程,我们最终将改变未来的企业、云、汽车和运营商架构的转型方式,使其变得更好。

Marvell and the M logo are trademarks of Marvell or its affiliates. Please visit for a complete list of Marvell trademarks. Other names and brands may be claimed as the property of others.

Marvell 和 m 徽标是 Marvell 或其附属公司的商标。请访问查看 Marvell 商标的完整列表。其他名称和品牌可能被视为他人的财产。

This press release contains forward-looking statements within the meaning of the federal securities laws that involve risks and uncertainties. Forward-looking statements include, without limitation, any statement that may predict, forecast, indicate or imply future events, results or achievements. Actual events, results or achievements may differ materially from those contemplated in this press release. Forward-looking statements are only predictions and are subject to risks, uncertainties and assumptions that are difficult to predict, including those described in the "Risk Factors" section of our Annual Reports on Form 10-K, Quarterly Reports on Form 10-Q and other documents filed by us from time to time with the SEC. Forward-looking statements speak only as of the date they are made. Readers are cautioned not to put undue reliance on forward-looking statements, and no person assumes any obligation to update or revise any such forward-looking statements, whether as a result of new information, future events or otherwise.

本新闻稿包含联邦证券法所指的前瞻性陈述,涉及风险和不确定性。前瞻性陈述包括但不限于任何可能预测、预测、表明或暗示未来事件、结果或成就的陈述。实际事件、结果或成就可能与本新闻稿中所考虑的事件、结果或成就存在重大差异。前瞻性陈述只是预测,受难以预测的风险、不确定性和假设的影响,包括我们在10-k表年度报告、10-Q表季度报告以及我们不时向美国证券交易委员会提交的其他文件的 “风险因素” 部分中描述的那些内容。前瞻性陈述仅代表其发表之日。提醒读者不要过分依赖前瞻性陈述,也没有人承担任何义务更新或修改任何此类前瞻性陈述,无论是由于新信息、未来事件还是其他原因。

For further information, contact:
Kim Markle
pr@marvell.com

欲了解更多信息,请联系:
金·马克尔
pr@marvell.com

big

SOURCE Marvell

来源 Marvell

声明:本内容仅用作提供资讯及教育之目的,不构成对任何特定投资或投资策略的推荐或认可。 更多信息
    抢沙发