I started my technology career in 1990 selling NCR’s Intel UNIX and X86 servers to major financial institutions. At the time, minicomputers were all the rage, and Intel-based servers weren’t even considered powerful enough for the data center, but good enough for print servers. The way things have changed – X86 based servers now dominate the data center and minicomputers are almost dead.
Eight Failed Attempts on Arm Servers
Five years ago, IT also didn’t see Arm-based servers as industrial strength, which felt like a life sentence for edge computing, and who could blame people back then? By my count, there have been eight failed attempts at general-purpose Arm-based data center processors: Marvell V1, Marvell V2, Calxeda, Samsung, AMD, Qualcomm, Broadcom, and APM. Literally, companies have invested billions in cash with no return on investment. I know the reason for the disappearance of each of these chips, whether it is a lack of software ecosystem, insufficient performance, lack of system ecosystem, bad timing, etc.
AWS Provides Arm Data Center Server Credibility
But then, in 2019, things changed. AWS introduced Graviton 1 based on its Nitro edge compute, and then what seemed like overnight, the Peanut Gallery said Arm-based general-purpose servers arrived. I was part of that peanut gallery and through a combination of Arm and AWS investment, a marketplace was created. This market may have been for AWS users, but how about an Arm-based server cloud instance for everyone else? Enter Ampere Computing.
Ampère enters the scene
In just a few short years, Ampere has racked up an impressive number of design contracts with world-class cloud providers and a particularly prominent on-premises cloud service provider, HPE. Today, end customers can purchase Ampere-based compute instances from sixteen different companies around the world.
Ampere calculation
I have never seen such rapid adoption of a new server chip vendor in my career and led AMD’s corporate marketing group during Opteron. While I’d love to spend more time on where Ampere is headed, I’ll spend some time explaining how Ampere has achieved this level of success so quickly with cloud providers.
CSPs are different
First, it’s important to understand that public cloud service software demands are different from legacy on-premises software such as SAP, VMware, and Windows Server. Early on, Ampere optimized its processor and platforms for cloud workloads that promoted high single-threaded integer performance with consistent, scalable performance at the lowest power consumption. Ampere managed to cram 3,328 of those cores into a standard 12kW, 3-4x AMD and Intel rack. Ampere’s ultra-dense design has generated high performance and performance per watt for web services like NGINX, MySQL databases, in-memory caching like Redis, and H.264 media transcoding.
More than an IP
Some are confused and think that a vendor like Ampere just licenses technology from Arm, easily integrates it, and goes to market with chips like Ampere Altra and Ampere Altra Max. Arm deserves a lot of credit for its investment and its IPs, but there’s a lot of work to be done to then turn that IP into a high-performance, low-power, and reliable server solution. For Altra and Altra Max, Ampere allows Arm N1 core and other Arm IP like MMU. Ampere also licenses key Cadence and Synopsis IPs. Ampere is also creating many of its own IPs for its current product line related to mesh optimization, power delivery, and I/O scaling for higher core counts/ sockets, and integrates it with Arm, Cadence and Synopsis IPs to create the SoC.
You would think we would be done at this point, but you would be wrong. Ampere needs to create hardware and software platforms to make the SOC useful. Scalable hardware platforms should be standards compliant and include BMCs and all peripherals such as memory and storage. Ampere must then build UEFI-compliant firmware and, with Arm and other IP vendors, integrate hardware enabling software. At this point, Ampere integrates the SOC, hardware, and software platforms, validates the combination, and then you have an Ampere-compatible platform ready for an ODM or OEM. And you thought the job ended at IP. 😊 For what it’s worth, I know I’m grossly oversimplifying the effort required above. It’s a lot more complex than that, but what I’m trying to make is that the work doesn’t end with the creation of the IP.
Ampère now has an architecture license, like Apple
But now I’d like to speak from the other side of my mouth and say how excited I am about Ampere’s next-gen SoC, AmpereOne, which packs more Ampere IP. Like Apple, Ampere now has an architectural license from Arm that gives it the right to create its own custom processor core from scratch. And what have we seen Apple do with its architectural license? It has created SOCs that have twice the CPU performance of competitors in the smartphone space and twice the performance per watt in the laptop space.
While Ampere is already sampling AmpereOne with its custom cores and IPs, it doesn’t provide much detail beyond the fact that it will be made in TSMC 5nm, support DDR5 and PCIe Gen5, and slot into its socket. current. What we don’t know yet are performance, power or release dates, obviously three key variables for evaluating the chip. Given what Arm is planning for its new Neoverse N2 processor, it would make sense to me for AmpereOne to deliver at least 25% more power or performance or why invest all that investment in a custom core? Seeing what Alibaba has done with the Yitian 710 and the prowess of Ampere’s engineering team, 25% doesn’t seem like a stretch to me.
Like Apple, I’d expect Ampere to get more years of experience with their custom kernel, they’ll be looking to get even fewer licenses and build their own IPs around memory and I/O where he could strive to be the first, for example, with a type of memory. I could see Ampere creating its own mesh to connect all of its broadband to gain a competitive advantage in the market. Just like Apple.
In conclusion
Arm-based cloud data center servers had been in development for over a decade. Like X86 in the late 80s and early 90s, Arm was considered unworthy in the data center. Now times have changed, and we have companies like Ampere Computing offering their cloud-optimized instances through sixteen major public cloud providers and soon HPE for on-premises cloud. Like Apple, Ampere has taken an Arm architectural license for its AmpereOne SoC which could outperform anything in the Arm SoC market on a performance-per-watt basis. I can’t wait to see how AmpereOne performs, and we should know soon as it’s been sampling since May. It’s an exciting time to be a compute consumer with the increased competitiveness brought by companies like Ampere.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.
Moor Insights & Strategy, like all research and technology industry analytics companies, provides or has provided paid services to technology companies. These services include research, analysis, consulting, consulting, benchmarking, acquisition matching and conference sponsorship. Company has had or currently has paid business relationships with 8×8, Accenture, A10 Networks, Advanced Micro Devices, Amazon, Amazon Web Services, Ambient Scientific, Anuta Networks, Applied Brain Research, Applied Micro, Apstra, Arm, Aruba Networks (now HPE), Atom Computing, AT&T, Aura, Automation Anywhere, AWS, A-10 Strategies, Bitfusion, Blaize, Box, Broadcom, C3.AI, Calix, Campfire, Cisco Systems, Clear Software, Cloudera, Clumio, Cognitive Systems , CompuCom, Cradlepoint, CyberArk, Dell, Dell EMC, Dell Technologies, Diablo Technologies, Dialogue Group, Digital Optics, Dreamium Labs, D-Wave, Echelon, Ericsson, Extreme Networks, Five9, Flex, Foundries.io, Foxconn, Frame ( now VMware), Fujitsu, Gen Z Consortium, Glue Networks, GlobalFoundries, Revolve (now Google), Google Cloud, Graphcore, Groq, Hiregenics, Hotwire Global, HP Inc., Hewlett Packard Enterprise, Honeywell, Huawei Technologies, IBM, Infinidat, Infosys, Inseego, IonQ, IonVR, Inseego, Info sys, Infiot, Intel, Interdigit al, Jabil Circuit, Keysight, Konica Minolta, Lattice Semiconductor, Lenovo, Linux Foundation, Lightbits Labs, LogicMonitor, Luminar, MapBox, Marvell Technology, Mavenir, Marseille Inc, Mayfair Equity, Meraki (Cisco), Merck KGaA, Mesophere, Micron Technology, Microsoft, MiTEL, Mojo Networks, MongoDB, MulteFire Alliance, National Instruments, Neat, NetApp, Nightwatch, NOKIA (Alcatel-Lucent), Nortek, Novumind, NVIDIA, Nutanix, Nuvia (now Qualcomm), onsemi, UNOG, OpenStack Foundation, Oracle, Palo Alto Networks, Panasas, Peraso, Pexip, Pixelworks, Plume Design, PlusAI, Poly (formerly Plantronics), Portworx, Pure Storage, Qualcomm, Quantinuum, Rackspace, Rambus, Rayvolt E-Bikes, Red Hat, Renesas, Residio, Samsung Electronics, Samsung Semi, SAP, SAS, Scale Computing, Schneider Electric, SiFive, Silver Peak (now Aruba-HPE), SkyWorks, SONY Optical Storage, Splunk, Springpath (now Cisco), Spirent, Splunk, Sprint (now T-Mobile), St ratus Technologies, Symantec, Synaptics, Syniverse, Synopsys, Tan ium, Telesign, TE Connectivity, TensTorrent, Tobii Technology, Teradata, T-Mobile, Treasure Data, Twitter, Unity Technologies, UiPath, Verizon Communications, VAST Data, Ventana Micro Systems, Vidyo, VMware, Wave Computing, Wellsmith, Xilinx, Zayo, Zebra, Zededa, Zendesk, Zoho, Zoom and Zscaler. Patrick Moorhead, Founder, CEO and Chief Analyst of Moor Insights & Strategy, is an investor in dMY Technology Group Inc. VI, Dreamium Labs, Groq, Luminar Technologies, MemryX and Movandi.
Patrick Moorhead, Founder, CEO and Chief Analyst of Moor Insights & Strategy, is an investor in dMY Technology Group Inc. VI, Dreamium Labs, Groq, Luminar Technologies, MemryX and Movand
#Ampere #Computing #building #cloud #data #center #Apple #Computing