It’s solely two years late, however the fourth technology of Xeon Scalable processors, aka Sapphire Rapids, is hitting the bottom operating, with each main OEM providing new servers that includes the chips.
The 4th Gen Xeon Scalable is notable as a result of it accommodates a lot of specialty computing engines along with its x86 cores, and it additionally has numerous cores as nicely; as much as 60. One of many particular engines is for AI acceleration, as Intel is decided to make the CPU viable as an AI processor as an alternative of GPUs. So not surprisingly, most of the new servers are constructed with AI processing in thoughts.
The entire servers help the newest applied sciences discovered within the 4th Gen Xeon Scalable: DDR5 reminiscence, PCI Categorical 5 interconnects, and CXL 1.1 reminiscence pooling.
This is the record.
Dell Applied sciences
Dell Applied sciences launched a complete of 13 PowerEdge servers, masking enterprise knowledge facilities, large-scale public clouds, and edge places in rack, tower, and multi-node designs. They arrive with Dell Good Movement, a brand new function within the Dell Good Cooling suite to extend airflow and cut back fan energy by as much as 52% in comparison with earlier technology servers.
Dell claims its PowerEdge R760 delivers as much as 2.9x larger AI inferencing with the brand new Xeon Scalable processors, whereas the PowerEdge R760 additionally gives as much as a 20% enhance in VDI customers and over 50% extra SAP Gross sales & Distribution customers on one server, in comparison with the earlier technology.
Along with updating current traces, Dell introduction the PowerEdge HS5610 and HS5620 servers delivers optimized options tailor-made for cloud service suppliers managing large-scale, multi-vendor knowledge facilities. They’re two-socket servers that are available each 1U and 2U kind components.
The entire servers include up to date Dell CloudIQ, which mixes proactive monitoring, machine studying and predictive analytics whereas providing a complete view of servers wherever they reside. Additionally they include Dell ProDeploy companies for fast deployment of the {hardware} and Dell iDRAC9 for simplified deployment and diagnostics.
HP Enterprise
HPE introduced a serious replace to its Alletra line of enterprise storage merchandise. Alletra is a rebranding of its Primera and Nimble storage arrays with a storage-as-a-service mannequin for its 9000 and 6000 sequence, respectively. The Alletra 4000 is a rebranding of its Apollo 4000 knowledge storage server line, with two new merchandise: the Alletra 4110 and 4120.
The Alletra 4110 is a 1U, twin CPU server with as much as 3TB of DDR5 reminiscence and as much as 307.2TB capability with 20x 15.36TB SSDs. The Alletra 4120 is a hybrid NVMe/HDD increased capability 2U chassis with as much as 6TB of DDR5 reminiscence. It helps a mixture of NVMe, SAS, or SATA drives.
Each methods are aimed toward giant enterprises with a spread of data-centric workload use circumstances from knowledge lakes and archives to high-throughput and in-place analytics and AI/ML, hyperconverged infrastructure (HCI), and cache-intensive workloads like databases.
Lenovo
One factor about Lenovo, they do issues in a giant manner. The corporate launched 25 new ThinkSystem, ThinkAgile servers and hyperconverged options operating the brand new Xeon Scalable processors. They’re offered underneath the banner of the Lenovo Infrastructure Options V3 portfolio and span makes use of like in-memory and huge transactional databases, real-time analytics, ERP, CRM, and virtualized and containerized workloads.
Some methods are updates to current traces, whereas others are new, such because the ThinkSystem SR850 V3, a four-socket 2U design for high-intensity workloads. The brand new Lenovo ThinkAgile V3 HX, MX and VX hyperconverged infrastructure merchandise include Microsoft, Nutanix, and VMware software program pre-installed for a HCI atmosphere.
Some methods are cooled by the fifth technology of Lenovo’s Neptune direct water-cooling expertise, which it claims can cut back energy consumption by as much as 40% vs. utilizing followers. Lenovo additionally stated choose fashions help liquid cooling for Excessive Bandwidth Reminiscence (HBM), a primary within the trade. HBM sits subsequent to the CPU or GPU, and air cooling them is a problem. However Lenovo says it should help Intel’s Xeon Max Collection CPUs and Intel’s GPU Max, which aren’t but delivery.
Supermicro
Supermicro launched greater than 50 servers and storage methods throughout 15 households collectively often called X13. They cowl the gamut of AI, HPC, cloud computing, media, enterprise, and 5G/telco/edge workloads. They vary from single-socket to eight socket designs and help GPU playing cards together with the CPU.
Among the many households:
- SuperBlade (high-performance, density-optimized blades),
- Hyper (high-performance rackmount servers),
- BigTwin (for cloud, storage, and media workloads),
- GrandTwin (an all-new structure for single-processor efficiency)
- SuperEdge (single-processor nodes in a 2U, short-depth kind issue for the sting)
- CloudDC (meant for cloud environments)
- Petascale Storage (all-flash NVMe storage methods)
Inspur Info
The Chinese language ODM Inspur Info, well-liked with hyperscale cloud suppliers, introduced that its G7 server platform would help Xeon Scalable processors with a complete of 16 several types of servers. Inspur stated efficiency is as much as 61% quicker, and energy consumption is as much as 30% much less in contrast with the earlier technology of Intel-based merchandise.
The servers are designed round quite a lot of use circumstances, resembling general-purpose computing, important computing, and AI. The G7 design is concentrated on inexperienced power, with each chilly plate and immersion cooling schemes. Inspur additionally claims its clever fault analysis has an accuracy charge as much as 95%.
The G7 server household ranges from the high-end TS860G7 with eight sockets in a 6U area and 128 DIMM slots of DDR5 reminiscence, to the NF5180G7 general-purpose 1U rack server, which helps immersion. Different servers are available two- and four-socket designs.
There’s additionally the NF5266G7 high-density storage server with a three-tier storage structure in a 2U area. Inspur says its computing efficiency is as much as 60% quicker than the earlier technology.
Giga Computing
Giga Computing, the enterprise spin-off of shopper element maker Gigabyte Computing, introduced servers and server motherboards based mostly across the 4th Gen Intel Xeon Scalable processor in addition to the Intel Xeon Max Collection.
Giga is focusing on AI, cloud computing, superior analytics, HPC, and networking, in addition to storage purposes. Giga Computing has 14 new sequence of servers with a complete of 78 configurations for patrons to select from.
The brand new traces embrace:
- Single and twin socket M-Collection motherboards
- R-series 1U and 2U rackmount servers
- G-series GPU servers with thermal and mechanical design for the warmer GPU and constructed for HPC, AI, and high-performance parallel processing.
- H-series high-density 2U 4-node servers
- E-series edge servers with brief kind components for small settings
Copyright © 2023 IDG Communications, Inc.
