Need a 128 core dedicated server? The first question is whether the requirement means 128 physical cores, 128 hardware threads or 128 vCPU. StreamData’s current public dedicated catalogue includes an AMD EPYC 7C13 system with 64 physical cores and 128 threads. That is a high-core server, but it is not a 128-physical-core machine.
Core, thread and vCPU are different
| Term | Meaning |
|---|---|
| Physical core | Actual execution core on the processor |
| Hardware thread | Logical execution context exposed by SMT |
| vCPU | Virtual CPU allocation presented to a VM |
What StreamData currently publishes
The dedicated catalogue reaches the AMD EPYC 7C13 at 64 cores, 256 GB DDR4 and 2 TB NVMe. StreamData’s CPU comparison article describes the 7C13 as 64 cores / 128 threads and contrasts it with the 16-core / 32-thread Ryzen 9 9950X.
If a tender or workload genuinely requires 128 physical cores, ask for a custom configuration rather than relying on the phrase “128 threads.”
When very high core count is useful
- dense Proxmox/VMware/Hyper-V hosts;
- parallel build farms;
- many independent application workers;
- rendering and batch processing;
- large container fleets;
- high-throughput data processing.
When a 9950X can be faster
StreamData’s own EPYC 7C13 vs Ryzen 9 9950X article shows the trade-off: EPYC wins on aggregate multi-thread capacity in the cited PassMark comparison, while the 9950X has much stronger single-thread performance. If the workload has one hot thread, buying more cores may do very little.
Memory and storage must scale too
A high-core server becomes inefficient if it lacks RAM or storage throughput. For virtualisation, sum guest memory, host overhead and storage cache. For databases, profile the working set, query latency and disk response. For build farms, measure both CPU parallelism and artifact/cache I/O.
One large host vs multiple smaller hosts
One 128-core-class system can simplify management and some licensing, but it creates a large failure domain. Two 64-core hosts can provide better maintenance flexibility and failover if the application can split across nodes.
Buyer checklist
- Specify physical cores vs threads vs vCPU.
- Record required RAM and storage.
- Confirm socket count and CPU model.
- Check software licensing rules.
- Test whether the workload scales past 32/64 cores.
- Define port speed, transfer and DDoS scope.
- Design backup and failover.
Bottom line
StreamData’s current highest-core public system is a 64-core / 128-thread EPYC 7C13 server. If you truly need 128 physical cores, treat that as a custom architecture requirement rather than confusing thread count with core count.
Next step: review the dedicated catalogue and send the exact physical-core, RAM, storage and network requirement.
Frequently asked questions
Why does StreamData show 64 cores but 128 threads?
The EPYC 7C13 exposes simultaneous multithreading, so each of its 64 physical cores can present two hardware threads. Thread count is useful for scheduling parallel work but should not be reported as 128 physical cores.
Would two 64-core StreamData servers be better than one custom 128-core box?
They can be if the software supports horizontal scaling and you value maintenance/failure isolation. A single larger host can be simpler when a workload needs one shared-memory system or licensing favours consolidation.
Does a 128-vCPU VM equal a 128-core server?
No. A vCPU is a virtual scheduling unit. Performance depends on the underlying host, allocation policy and contention.
When should I request custom high-core hardware?
When measurements show that the current 64-core/128-thread class is insufficient and the application continues to scale efficiently beyond it. Include RAM, storage and network requirements in the same request.