Lenovo has introduced a family of servers tuned for AI and more.
The vendor announced three next-gen ThinkSystem V4 servers are built on P-core-based Xeon 6 processors, that are purpose-built to run a wide range of workloads, including the most compute intensive tasks such as algorithmic trading, astrophysics, and CRM.
Lenovo claims that with the new systems, customers can see up to 6.1x higher compute performance than previous generation of CPUs as well as twice the memory bandwidth when using the new MRDIMM technology.
The three new ThinkSystem V4 models include:
SR630 V4: A high density, 1U device designed to optimize performance in compact environments. The super dense 1U system is ideal for cloud service providers (CSPs), telcos and fintech operations, enabling them to manage real-time transactions requiring low latency and high throughout performance with limited floor space.
SR650 V4: Comes in a 2U design with increased GPU capacity for up to 2x computation performance in a compact form factor. It is targeted at engineering, modeling & simulation and AI with the promise of up to 50% cost savings over previous generation hardware for GPU-intensive workloads.
SR650a V4: a 2U2S design specifically meant for high-end use, such as AI and other GPU-intensive workloads like machine learning, virtual desktop infrastructure (VDI), and media analytics. The platform supports up to four double-wide GPUs with front GPU access. Of the three servers, this is the one most geared toward AI.
The ThinkSystem V4 solutions use Lenovo’s homegrown Neptune liquid cooling technology. A new feature called Neptune Core Compute Complex Module offers reduced fan speeds, quieter operation and lower power consumption. The module supports the cooling of CPUs and memory as well as cooling the four GPUs in the SR650a.
The ThinkSystem V4 lineup is complemented by XClarity One, a centralized management and deployment console capable of scaling from one system to 1,000. New integrated Enterprise Remote Control enables remote access and management of enterprise infrastructure no matter where it exists. A new AI-powered analytics tool offers Server SSD Predictive Failure Analysis (PFA) to catch potential storage drive failure before it happens.
The ThinkSystem V4 family ships March 17.
Source:: Network World