来源： The Register
原英文标题： <Fujitsu's A64FX successor will be an Arm-based datacenter chip>
Fujitsu's Arm-based A64FX processor may have driven the most powerful supercomputer in the world, but it looks like its successor will be a more general-purpose chip that will focus on energy efficiency.
The Japanese IT giant built its own Arm processor for use in the Fugaku supercomputer, and commercialized it in the PRIMEHPC FX1000 and FX700 systems, as well as HPE using the chip in its Apollo 80 high-performance computing (HPC) platform.
'Overwhelming energy efficiency'
Fujitsu disclosed a successor chip during its recent ActivateNow: Technology Summit at the Computer History Museum in Mountain View, California. In a presentation given by Vivek Mahajan, there was mention of an "Arm-based CPU for Next-gen DC" slated for 2028.
We asked the company for further details on this processor, and it told us the chip is currently code-named MONAKA and expected to be released in 2027. It is planned to deliver higher performance with lower energy consumption, as well as contributing to "the realization of a carbon-neutral and sustainable society."
According to information seen by The Register, MONAKA is expected to not only boost traditional HPC workloads, but also provide high performance for AI and data analytical workloads.
However, at the same time the chip is forecast to provide "overwhelming energy efficiency" when compared with rival CPUs that will be available within the same timeframe, with Fujitsu indicating that MONAKA should have a 1.7x lead in application performance while offering 2x the performance per watt.
The development of MONAKA is part of a program overseen by Japan's New Energy and Industrial Technology Development Organization (NEDO) research agency, which aims to deliver energy savings of 40 percent or better in datacenters by 2030.
As well as contributing to this by developing an energy-efficient CPU, Fujitsu said it will collaborate on related technologies, such as low power consumption accelerators, photonics-based smart NICs, and disaggregation technology.
It thus seems that MONAKA will be less like the A64FX and perhaps more like the Arm-based server chips from Ampere, although these are aimed at cloud providers and hyperscalers and so are more focused on delivering full utilization rather than strictly saving energy.
"The next-generation DC CPU (MONAKA) that we're developing will have a wider range of features and will prove more energy efficient," a Fujitsu spokesperson told us.
"The range of potential applications is wider than that of the A64FX, which has special characteristics (e.g., interconnects) specific to Fugaku," the spokesperson added.
These special characteristics of the A64FX include the 28Gbps Tofu-D interconnect, high-speed HBM2 stacked memory, and the 512-bit Scalable Vector Extensions. It is unlikely that these will feature in MONAKA, although Fujitsu has few specific details to share on the architecture at this point.
富士通基于Arm的A64FX处理器可能已经驱动了世界上最强大的超级计算机，但看起来它的继任者将是一个更通用的芯片，将专注于能源效率。
这家日本IT巨头建立了自己的Arm处理器用于Fugaku超级计算机，并在PRIMEHPC FX1000和FX700系统中实现了商业化，HPE也在其Apollo 80高性能计算（HPC）平台中使用该芯片。
压倒性的能源效率
富士通在最近的ActivateNow.技术峰会上披露了一款后续芯片。技术峰会期间，富士通在加州山景城的计算机历史博物馆披露了后续芯片。在Vivek Mahajan的演讲中，提到了计划于2028年推出的 "基于Arm的下一代DC的CPU"。
我们向该公司询问了关于该处理器的进一步细节，它告诉我们该芯片目前的代号为MONAKA，预计将于2027年发布。它计划以更低的能耗提供更高的性能，并为 "实现碳中性和可持续发展的社会 "做出贡献。
根据The Register看到的信息，MONAKA预计不仅能提升传统的HPC工作负载，还能为AI和数据分析工作负载提供高性能。
然而，与此同时，该芯片预计将提供 "压倒性的能源效率"，与将在同一时间框架内推出的竞争对手的CPU相比，富士通表示，MONAKA在应用性能方面应该有1.7倍的领先优势，同时提供2倍的每瓦特性能。
MONAKA的开发是日本新能源和工业技术开发组织（NEDO）研究机构监督的项目的一部分，该项目旨在到2030年使数据中心的能源节约率达到40%或更高。
除了通过开发高能效的CPU做出贡献外，富士通说它将在相关技术方面进行合作，如低功耗加速器、基于光子技术的智能网卡和分解技术。
这样看来，MONAKA将不像A64FX，也许更像安培的基于Arm的服务器芯片，尽管这些芯片是针对云供应商和超大规模的，因此更注重于提供充分的利用率，而不是严格的节能。
"富士通的一位发言人告诉我们："我们正在开发的下一代DC CPU（MONAKA）将有更广泛的功能，并将证明是更节能的。
"潜在的应用范围比A64FX更广，因为A64FX具有富士通特有的特性（如互连），"该发言人补充说。
A64FX的这些特殊性包括28Gbps Tofu-D互连，高速HBM2堆叠内存，以及512位可扩展矢量扩展。尽管富士通目前几乎没有关于该架构的具体细节可以分享，但这些不太可能在MONAKA中出现。 |
310112100042806
发表于 2023-3-10 19:43
