Lenovo 4XG7A72958 ST650 V2 Platinum 8358 32C 250W 2.6GHZ

Lenovo 4XG7A72958 ThinkSystem ST650 V2 Xeon Platinum 8358 32-Core Processor Option Kit

Overview

The Lenovo 4XG7A72958 is a factory-configured processor option kit pairing the Intel Xeon Platinum 8358 with the ThinkSystem ST650 V2 tower server. Designed for organizations scaling compute density in a two-socket tower chassis, this kit adds a second 32-core, 64-thread processor operating at a 2.6 GHz base clock with a 3.4 GHz Turbo Boost peak — delivering the headroom needed for parallel workloads like virtualization, database serving, and security analytics without a rack migration. At 250W TDP on a 10 nm process node, it represents a deliberate tradeoff: meaningful core count and frequency in a form factor that fits a server closet or branch office without raised-floor infrastructure.

Sourced through channel-direct distribution with no grey-market or parallel imports, the 4XG7A72958 is a genuine factory option — not a field-harvested pull — which matters for platform firmware compatibility and manufacturer warranty continuity on the ST650 V2.

Key Features

• 32 Cores / 64 Threads (Intel Xeon Platinum 8358): With 64 logical processors available to the hypervisor, the ST650 V2 can host substantially more VMs per chassis than a 16- or 24-core configuration — critical if you're consolidating branch-office workloads or running concurrent VMS analytics engines on a single physical node.
• 2.6 GHz Base / 3.4 GHz Turbo Boost: Single-threaded tasks — license checks, database index operations, real-time event correlation — accelerate to 3.4 GHz on demand. The base clock holds at 2.6 GHz across all 32 cores under sustained load, so throughput doesn't collapse under parallel batch jobs.
• Intelligent Core Tiering (12 High-Priority + 20 Efficiency Cores): The 8358 exposes 12 high-priority cores running at 2.8 GHz and 20 standard cores at 2.4 GHz. Schedulers and hypervisors that recognize this topology can pin latency-sensitive threads — RDBMS foreground queries, VMS live decoding — to the faster cores while batch processes consume the efficiency pool. This reduces contention without adding sockets.
• 48 MB L3 Cache: 48 MB of last-level cache reduces main-memory round trips on large working sets. For in-memory databases and analytics workloads operating on datasets that exceed L2 capacity, this cache depth directly translates to lower average latency per query.
• 250W TDP on 10 nm Lithography: Intel's 10 nm Ice Lake-SP fabrication process delivers the 32-core count within a 250W thermal envelope. Verify your ST650 V2 power supply and cooling configuration supports 250W per socket before adding the second CPU — the tower's thermal headroom is finite and the ST650 V2 cooling subsystem must be validated for dual high-TDP processors.
• 64-bit / 11.2 GT/s UPI Interconnect: The 11.2 GT/s Ultra Path Interconnect bus rate between sockets keeps NUMA latency in check on the dual-socket ST650 V2 platform. Memory-intensive workloads that cross socket boundaries — common in NUMA-unaware applications — benefit from the faster interconnect compared to prior-generation 10.4 GT/s implementations.
• LGA 4189 Socket (Ice Lake-SP Platform): The LGA 4189 socket is the 3rd Generation Intel Xeon Scalable platform socket. Confirm your ST650 V2 motherboard revision supports 3rd Gen Xeon Scalable before ordering — earlier ST650 V2 board revisions may require a firmware update or may not support 250W-class processors.

Integration and Compatibility

The 4XG7A72958 is validated for the ThinkSystem ST650 V2 server platform. The Intel Xeon Platinum 8358 sits within the Intel Xeon Platinum 8000 Series and is compatible with 3rd Generation Intel Xeon Scalable infrastructure — including DDR4 memory DIMMs and PCIe Gen 4 I/O already present in the ST650 V2 platform. When adding a second CPU to an existing single-processor ST650 V2, verify that memory DIMMs are balanced across both CPU memory controllers to avoid asymmetric NUMA configurations that degrade performance on parallel workloads.

For enterprise server configurations running virtualization stacks (VMware vSphere, Microsoft Hyper-V) or high-density rack and tower server deployments, the Xeon Platinum 8358 tier provides the core count and cache depth to host production-grade workloads without stepping up to a rack server chassis. Organizations evaluating network video recorders at scale — particularly AI-accelerated VMS platforms that offload analytics to server CPUs — will find the 32-core, 64-thread configuration appropriate for running multiple concurrent deep-learning inference streams. Pair with additional validated DDR4 server memory to fully populate the memory channels enabled by this processor.

Frequently Asked Questions

Q: What server is the Lenovo 4XG7A72958 compatible with?

A: The 4XG7A72958 is a processor option kit validated for the Lenovo ThinkSystem ST650 V2 tower server. It is not a universal upgrade — verify your specific ST650 V2 board revision and firmware level before installing.

Q: Does the 4XG7A72958 include a heatsink or thermal solution?

A: Package contents are not detailed in the available evidence. Contact your pre-sales team to confirm whether a processor heatsink is included or must be ordered separately for the ST650 V2 platform.

Q: How many cores and threads does the Intel Xeon Platinum 8358 provide?

A: The Xeon Platinum 8358 delivers 32 physical cores and 64 threads via Intel Hyper-Threading, with a 2.6 GHz base clock and 3.4 GHz Turbo Boost peak frequency.

Q: What is the TDP of the Xeon Platinum 8358 in the 4XG7A72958?

A: The processor has a 250W thermal design power rating. Ensure your ST650 V2 power supply configuration and chassis cooling are validated for dual 250W-class processors if adding a second socket.

Q: What generation of Intel Xeon Scalable is the 4XG7A72958?

A: The Xeon Platinum 8358 is a 3rd Generation Intel Xeon Scalable processor (Ice Lake-SP) using the LGA 4189 socket on Intel's 10 nm process node.

Q: What is the cache size on the Xeon Platinum 8358?

A: The processor features 48 MB of L3 (last-level) cache, which helps reduce main-memory latency on large in-memory workloads such as databases and analytics engines.