Dell PowerEdge R770 Review: Modular, Puissant, and AI Supported

Dell’s PowerEdge R7x0 series servers have long been the cornerstone of data centers, renowned for their superior manufacturing quality, thoughtful design, performance, density, and reliability, as well as their versatile 2U form factor. These servers have been constantly evolving to meet ever-changing demands. Now, with the launch of Dell PowerEdge R770, the series has made a significant leap forward.

The TR770 is equipped with Intel’s new Xeon 6 processor series for the first time, including Xeon 6500 et 6700 P and E core processors. It marks Dell’s first comprehensive adoption of the OCP Data Center Modular Hardware Systems (DC MHS) standard in its mainstream server product line. These two changes together indicate a significant shift in functionality and design philosophy.

The launch of R770 comes at a time when data centers are facing increasing pressure. The workload is becoming increasingly diverse and demanding. The continuous growth of data has driven the demand for powerful analytics and databases. From training complex models to deploying real-time inference, artificial intelligence is no longer a niche application, but a core business driver that requires significant computing power and professional acceleration.

En même temps, people are also very concerned about energy efficiency and optimizing overall cost of ownership. En outre, the industry is increasingly looking to adopt open standards to promote innovation, enhance interoperability, and minimize supplier lock-in. The R770 features new processor options and adopts OCP DC MHS, designed to address these challenges head-on.

The R770 processor adopts Intel Xeon 6 series processors, including the 6700 et 6500 series, and integrates performance and efficiency cores built on the Socket E2 (LGA4710-2) platform. In this review, we pay special attention to the P series SKUs.
Intel uses a module based design to build these processors, combining I/O modules with one or two computing modules. This achieves scalability within the series, with a maximum configuration of 86 P-cores (XCC) when using two computing modules, and a minimum configuration of 48 P-cores (HCC) ou 16 P-cores (LCC) when using a single computing module.

A key difference between these processors and the previous generation Sapphire and Emerald Rapids processors is that all Xeon 6 processors generally come with built-in accelerators. This includes Intel QuickAssist technology for encryption and compression, Intel Data Stream Accelerator for data mobility, Intel Memory Analysis Accelerator for database and analytics acceleration, and Intel Dynamic Load balancers for network processing efficiency.

The memory and I/O bandwidth have also been significantly upgraded. The Xeon 6700/6500 P core series supports 8-channel DDR5 memory. They also pave the way for multiplex grade DIMMs (MRIMMs), providing speeds of up to 8800 MT/s. In terms of I/O, these processors support PCIe 5.0 and CXL 2.0. In a dual slot configuration, the platform can provide up to 88 PCIe channels per slot (a total of 176 channels).

Although there are differences between the P and E cores, the Xeon 6 series has an instruction set BIOS、 Consistent support for drivers, operating systems/applications, and RAS functionality simplifies integration and management of different deployment types. The P-core variant is designed for workloads that require critical performance, AI acceleration, high memory bandwidth, and a large amount of I/O per core; Think about demanding databases, HPC simulations, advanced analytics, and various AI applications.