The HPE ProLiant DL360 Gen10 is a compact 1U server designed to support a wide range of Intel server processors. Depending on the configuration, it can use an entry-level Xeon with a modest core count or two high-end CPUs delivering dozens of physical cores. Clock speed also varies considerably: a processor may prioritize a high base frequency, a large number of cores, or short bursts of turbo performance. Understanding these differences helps you select the right system for virtualization, databases, applications, cloud infrastructure, or general-purpose computing without judging performance by GHz alone.
The DL360 Gen10 supports Intel Xeon Scalable processors from both the 1st Generation Intel Xeon Scalable family, commonly associated with Skylake, and the 2nd Generation Intel Xeon Scalable family, commonly associated with Cascade Lake. Hewlett Packard Enterprise documents supported combinations through server-specific configurations, firmware, and system options. These processors use the LGA3647 platform and provide a much broader range of cores, memory capabilities, and features than older Xeon E5 generations.
Compatibility is not determined only by whether a CPU fits the socket. BIOS support, heatsink requirements, memory population, and HPE configuration rules can also matter. A processor listed for a generic LGA3647 platform is not automatically approved for every DL360 Gen10 system, so check the server documentation before replacing or upgrading a CPU.
The four principal tiers are Intel Xeon Bronze, Silver, Gold, and Platinum. Bronze models generally provide the lowest cost and performance, while Silver adds more capability for mainstream servers. Gold processors cover a broad range of core counts and frequencies, and Platinum models usually target the highest core counts, memory bandwidth, and enterprise workloads.
The tier is a useful starting point, but it does not completely define speed. Some Gold chips have fewer cores and higher frequencies than certain Platinum chips. Compare the exact processor model, supported features, power rating, and price rather than choosing only by the family name.
A single-socket configuration uses one processor and is often sufficient for file services, smaller databases, web applications, and lightly virtualized environments. A dual-socket configuration installs two compatible CPUs, increasing available cores, memory channels, and total memory capacity.
Two sockets do not automatically make every application twice as fast. Software must be able to use additional threads, and workloads that cross sockets can experience NUMA latency. Dual-socket systems also consume more power and may require matched processors, correct memory placement, and additional licensing consideration.
The core count depends on the installed Xeon model rather than simply the server generation. Supported options range from low-core Bronze processors, such as the six-core Xeon 3204, to high-end Platinum processors with up to 28 physical cores. Silver and Gold models span the middle of that range, giving administrators flexibility to prioritize purchase cost, frequency, capacity, or parallel processing.
For example, the Intel Xeon 4210 has 10 cores, the Intel Xeon 5218 has 16, and the Intel Xeon 6230 has 20. The Intel Xeon 8280 reaches 28 cores. These figures describe physical cores inside one CPU. They are more useful than a family label when estimating how many virtual machines, database threads, or batch jobs a server can handle.
With two supported 28-core processors, the DL360 Gen10 can reach a maximum of 56 physical CPU cores. This is the headline maximum for a two-processor configuration using the highest-core-count compatible Xeon options. The actual supported maximum should still be confirmed against the server's specific QuickSpecs revision and configuration rules.
Adding a second CPU also activates additional memory channels and can expand total memory capacity, depending on the DIMMs installed. However, the result is a NUMA system. Virtualization platforms and operating systems generally manage NUMA well, but performance-sensitive applications may need processor and memory affinity configured carefully.
Many supported Xeon Scalable CPUs use Intel Hyper-Threading Technology. A physical core can expose two logical processors, allowing the operating system to schedule more threads. Therefore, a 28-core CPU may appear as 56 logical processors, and two such CPUs may appear as 112 logical processors.
Logical processors are not equivalent to additional physical cores. Hyper-Threading can improve utilization when one thread is waiting, but it does not double execution resources in every workload. For licensing, capacity planning, and performance estimates, distinguish physical cores from logical processors and test the applications that matter to you.
The base clock speed is the nominal frequency a processor is designed to sustain under defined power and thermal conditions. The maximum turbo frequency is the highest short-term frequency the CPU may reach under suitable conditions, usually on one or a small number of active cores. These are different specifications and should not be treated as two constant operating speeds.
A processor can have a relatively low base frequency but deliver a high turbo frequency for interactive or lightly threaded work. Another model may have more cores and a lower base frequency, making it better for heavily parallel tasks. When comparing DL360 Gen10 CPUs, read both values along with core count, cache, memory support, and power requirements.
Turbo Boost dynamically raises frequency when the processor has available power, current, and thermal headroom. The advertised maximum applies only to specific conditions, often a limited number of active cores. When all cores are busy, the sustainable frequency may be lower than the maximum turbo figure.
For example, a CPU listed with a 3.9 GHz maximum turbo speed may reach that rate during a short, lightly threaded task but operate below it during a sustained rendering or virtualization load. This is normal behavior rather than a fault. Benchmark results should therefore identify whether they measure single-thread, all-core, burst, or long-duration performance.
Clock behavior is influenced by workload intensity, active core count, temperature, power limits, cooling, and firmware policy. The processor's thermal design power (TDP) is a design target for cooling and typical sustained power, not a guaranteed maximum electricity draw or a direct performance score.
In a dense 1U server, airflow and ambient temperature can affect sustained performance. A well-configured DL360 Gen10 with appropriate heatsinks, clean airflow, and current firmware gives the CPU more opportunity to maintain its intended frequencies. Memory errors, power restrictions, or an overloaded chassis can produce a different result from the specification sheet.
Bronze processors suit basic infrastructure where purchase cost and modest utilization are more important than maximum throughput. They can handle lightweight file services, simple web servers, management tools, and small office applications. Intel Xeon Silver processors are a stronger choice for mainstream general-purpose workloads, offering more cores, memory capability, and performance headroom.
These tiers can be sensible when the server will run only a few services or when software licensing is tied to core count. They are also easier to cool and may reduce energy use compared with high-end CPUs. Avoid under-sizing, however, if future virtualization growth or database demand is likely.
Intel Xeon Gold processors provide a broad balance of cores, frequency, cache, memory features, and expansion capability. This makes them popular for virtualization clusters, application servers, medium-sized databases, analytics, and mixed workloads. Gold models vary widely, so an individual chip may emphasize high frequency or a larger core count.
For virtualization, examine the number of virtual machines, their vCPU allocations, peak CPU demand, and memory requirements. A moderate-core Gold CPU with strong frequency can be preferable for latency-sensitive guests, while a higher-core model may be better for consolidation. Keep enough CPU capacity for the hypervisor and expected growth.
Intel Xeon Platinum processors target demanding enterprise deployments that benefit from high core counts, large memory configurations, and strong platform features. They are appropriate for dense virtualization, large databases, scientific computing, intensive analytics, and other workloads that can keep many threads busy.
Platinum does not guarantee the best result for every application. A highly threaded workload may benefit greatly from 28 cores, while a lightly threaded application may prefer a faster lower-core Gold model. Platinum CPUs can also carry higher purchase and power costs, so confirm that the workload and software licensing justify the upgrade.
Comparing a processor model number is more useful than comparing Xeon tiers alone. The model tells you which specification sheet to consult for cores, base frequency, maximum turbo frequency, cache, memory support, and TDP. A lower-core CPU can have a higher base frequency, while a high-core CPU can deliver greater total throughput at a lower per-core speed.
The figures below are representative published specifications for commonly discussed DL360 Gen10 options. Exact HPE support can depend on the system configuration, firmware, and documentation revision. Turbo values are maximum ratings, not guaranteed all-core frequencies.
The Intel Xeon 3204 is a six-core Bronze processor with a 1.90 GHz base frequency and no standard turbo boost. The Intel Xeon 4210 has 10 cores, a 2.20 GHz base frequency, and up to 3.20 GHz turbo. The Intel Xeon 5218 provides 16 cores, a 2.30 GHz base frequency, and up to 3.90 GHz turbo.
The Intel Xeon 6230 has 20 cores, a 2.10 GHz base frequency, and up to 3.90 GHz turbo. At the top of this example set, the Intel Xeon 8280 offers 28 cores, a 2.70 GHz base frequency, and up to 4.00 GHz turbo. These examples show why core count and frequency must be evaluated together.
Clock speed is only one part of CPU performance. Cache can reduce trips to system memory, additional cores improve parallel throughput, and a higher TDP may accompany greater sustained capability. Memory speed, DIMM population, storage latency, and network throughput can also become bottlenecks before the processor reaches its limit.
For a lightly threaded application, the Xeon 4210 or 5218 may provide a better value than a 28-core chip. For consolidation or parallel computation, the Xeon 6230 or 8280 may finish more work per server. Compare benchmark results that resemble your workload, then account for energy use, licensing, cooling, and purchase price.
| Xeon 3204 | 6 cores; 1.90 GHz base; no turbo; Bronze class |
| Xeon 4210 | 10 cores; 2.20 GHz base; up to 3.20 GHz turbo; Silver class |
| Xeon 5218 | 16 cores; 2.30 GHz base; up to 3.90 GHz turbo; Gold class |
| Xeon 6230 | 20 cores; 2.10 GHz base; up to 3.90 GHz turbo; Gold class |
| Xeon 8280 | 28 cores; 2.70 GHz base; up to 4.00 GHz turbo; Platinum class |
Start with the workload rather than the processor tier. Databases, application servers, and virtual machines often benefit from a balance of cores and frequency, while rendering, scientific computing, compression, and large-scale analytics can use many cores efficiently. If response time matters most, prioritize strong per-core performance. If throughput matters most, prioritize physical cores and sustained all-core capacity.
Review current utilization with monitoring tools before upgrading. A server that rarely exceeds 30 percent CPU usage may gain little from a high-end Platinum processor, while a host running constant parallel workloads may need more cores even if its base frequency is lower.
CPU selection should be made alongside memory planning. More cores can support more virtual machines, but each guest also needs sufficient RAM, storage I/O, and network capacity. Two-socket systems may unlock more memory channels, although NUMA-aware configuration becomes more important.
Software licensing can change the economics substantially. Some databases, virtualization products, and commercial applications license by core or socket. A lower-core processor with higher frequency may cost less to license, while a high-core CPU may consolidate more workloads onto one server. Include support, energy, cooling, and future capacity in the comparison.
Before ordering a processor, check the HPE QuickSpecs for the exact DL360 Gen10 model and intended configuration. Confirm supported generation, socket count, BIOS or firmware requirements, heatsink options, memory restrictions, and whether a second CPU requires additional hardware. HPE documentation is the final authority for supported combinations.
Also verify that two processors are a supported pair and that the server has the correct power supplies and cooling arrangement. When buying refurbished hardware, ask for the installed CPU, firmware version, heatsink type, and complete part number. These checks help prevent a physically compatible processor from becoming an unsupported or poorly balanced upgrade.
Choosing a DL360 Gen10 processor is a balance between physical cores, per-core speed, turbo behavior, memory capacity, software licensing, and power consumption. Use the workload to determine whether you need frequency, parallel throughput, or a mixture of both. Then compare exact Xeon model specifications and verify the configuration in HPE documentation before buying. A carefully matched CPU usually delivers better value than simply selecting the processor with the highest advertised GHz or core count.
The server can support up to 56 physical cores when configured with two compatible 28-core Intel Xeon Scalable processors. The exact supported processor and configuration should be confirmed in the applicable HPE QuickSpecs.
Yes. The DL360 Gen10 supports selected 1st Generation and 2nd Generation Intel Xeon Scalable processors. Compatibility depends on the exact system configuration, firmware, heatsink, and HPE-supported options.
Not necessarily. Clock speed is only one factor. Core count, cache, memory configuration, turbo behavior, workload type, TDP, and software optimization also affect real-world performance.
Physical cores are the actual processing cores inside the CPU. Hyper-Threading can expose two logical processors per physical core, but logical processors do not provide the same performance as additional physical cores.
No. The Xeon 8280 offers more physical cores and is often better for heavily parallel workloads, while the Xeon 5218 can be a better value for workloads that need balanced performance, lower cost, or fewer licensed cores.
Dual-socket systems generally require compatible, matched processors with supported characteristics. Do not assume that two physically fitting CPUs can be mixed; check HPE documentation and the relevant QuickSpecs first.