New
AMD EPYC 9575F 64c/128t 3.3GHz-5.0GHz 400W (100-000001554)
P/N: 100-000001554
4 395€ (excl. TAX)
5274 € RRP en Amd.com
The recommended retail price provided by the product’s manufacturer.
Delivery is made within 3-7 days
This is an estimated timeframe. Delivery times may vary depending on logistics and stock availability.
Warranty 1 year
The product is covered by the manufacturer’s standard warranty.
In stock
Guaranteed Safe Checkout:
Bussines pricing from
4 087€
This price is available to registered corporate/business customers. To access it, submit a request on the website to receive a discounted invoice.
Business customers: submit a request to get an ongoing extra 3–7% discount.
AMD EPYC 9575F processor with fast EU delivery and worldwide shipping. The best price in the European Union. Includes official warranty.
Expert support On-line
Our specialist will help you choose the right server components and ensure full compatibility with your system.
Technical Specifications Product
| Dimensions | 17 × 17 × 10 cm |
|---|---|
| Country of manufacture | Taiwan |
| Manufacturer's warranty (years) | 1 |
| Series | EPYC |
| Number of cores | 64 |
| Number of threads | 128 |
| Clock frequency (GHz) | 3.3 |
| Cache L3 (MB) | 256 |
| Process technology (nm) | 4 |
| Maximum Turbo Frequency (GHz) | 5 |
| Memory type | DDR5 |
| Maximum memory channels | 12 |
| Maximum memory frequency (MHz) | 6400 |
| Heat dissipation TDP (W) | 400 |
| PCI Express controller | PCIE 5.0 |
| Number of PCI Express lanes | 128 |
| Processors on a motherboard | 2 |
| Length (cm) | 17 |
| Width (cm) | 17 |
| Architecture | Zen 5 (Turin) |
| Socket | SP5 |
Product description
AMD EPYC 9575F Processor
The AMD EPYC 9575F, based on the advanced Zen 5 (Turin) architecture, represents a significant leap in the realm of server and workstation processing. With an impressive 64 cores and 128 threads, this processor is engineered to handle demanding workloads with exceptional efficiency and performance. The EPYC 9575F operates at a base frequency of 2.5 GHz, which can be dynamically boosted to a turbo frequency of 3.5 GHz, allowing for versatile performance across a range of applications, from data analytics to virtualization.
Designed for high-demand environments, the EPYC 9575F features a substantial thermal design power (TDP) of 400 W, ensuring that it can maintain optimal performance levels under sustained loads. The processor’s large 320 MB L3 cache significantly enhances data retrieval speeds, reducing latency and improving overall responsiveness in multi-threaded applications. This is particularly beneficial for cloud computing, big data processing, and high-performance computing tasks where rapid data access is critical.
Memory support is another strong suit of the EPYC 9575F, as it utilizes the latest DDR5 technology. This allows for increased bandwidth and efficiency, facilitating faster data transfer rates and improved memory performance. The support for large memory capacities makes the EPYC 9575F ideal for enterprises that require robust memory solutions for their applications, ensuring that businesses can scale their operations without performance bottlenecks.
The architectural enhancements of the Zen 5 (Turin) design further contribute to the EPYC 9575F’s performance advantages. With improvements in core efficiency and interconnect bandwidth, this processor provides unparalleled multi-threaded performance, making it a reliable choice for server deployments. Its scalability ensures that businesses can adapt to changing workloads and demands, making it a future-proof investment for organizations looking to enhance their IT infrastructure.
In addition to its technical specifications, the AMD EPYC 9575F is backed by our commitment to customer satisfaction, offering fast delivery across Europe and global shipping options. Furthermore, it comes with an official warranty of 1 to 3 years, ensuring peace of mind for your investment. Explore the potential of the AMD EPYC 9575F for your business and elevate your processing capabilities today. For inquiries or to make a purchase, please contact us.
Product reviews
0
50
40
30
20
10
Only logged in customers who have purchased this product may leave a review.
Product Berchmark
Benchmark overview — AMD EPYC 9575F
The AMD EPYC 9575F is a frequency-optimized high-core server CPU designed for workloads that require
both strong aggregate throughput and elevated per-core performance.
With 64 cores and 128 threads, it targets environments where classic high-core processors begin to
show diminishing returns due to latency, synchronization overhead, or licensing constraints.
Within the EPYC 9005 lineup, the 9575F occupies a distinct niche. Unlike balanced throughput models
such as the EPYC 9555, this SKU prioritizes higher sustained frequencies across a large core count.
The result is a processor that performs exceptionally well in mixed workloads where parallel
execution dominates but serial and coordination phases still have a meaningful impact on total
execution time.
From a benchmarking perspective, the EPYC 9575F demonstrates a clear advantage over non-F
counterparts at the same core count in both single-thread and lightly loaded multi-thread
scenarios, while maintaining competitive performance under full-core utilization. This makes it
especially attractive for enterprise and technical workloads that scale well but are not
perfectly parallel.
In-family performance comparison (64-core EPYC 9005)
| CPU | Cores / Threads | PassMark CPU Mark | Single-Thread Score |
|---|---|---|---|
| EPYC 9555 | 64 / 128 | ~185,000 | ~4,200 |
| EPYC 9575F | 64 / 128 | ~197,000 | ~4,150 |
| EPYC 9655P | 96 / 192 | ~210,000 | ~4,100 |
At identical core counts, the EPYC 9575F delivers a measurable uplift in aggregate throughput compared to the EPYC 9555. This gain is driven primarily by higher effective frequencies under load rather than architectural differences. The improvement becomes most visible in workloads with uneven core utilization, frequent synchronization points, or phases that cannot fully saturate all cores simultaneously.
Single-thread performance remains clearly above that of balanced and dense SKUs, which reduces coordination overhead and improves responsiveness in mixed execution pipelines. While the difference may appear modest on paper, it often translates into meaningful wall-clock reductions in real-world enterprise applications.
Cross-generation comparison (64-core frequency-oriented class)
| CPU | Architecture | PassMark CPU Mark | Single-Thread Score |
|---|---|---|---|
| EPYC 7763 | Zen 3 | ~145,000 | ~2,900 |
| EPYC 9554F | Zen 4 | ~175,000 | ~3,800 |
| EPYC 9575F | Zen 5 | ~197,000 | ~4,150 |
Zen 5 brings a clear uplift over previous generations in both per-core efficiency and aggregate
performance. Compared to Zen 4 frequency-oriented SKUs, the EPYC 9575F improves not only peak
performance but also sustained behavior under continuous load.
For workloads that alternate between serial and parallel phases, these gains reduce the impact
of bottlenecks and improve overall system efficiency.
Cost-aware alternatives (planning comparison)
| CPU | Indicative RRP (€) | PassMark CPU Mark | Performance per € |
|---|---|---|---|
| EPYC 9555 | 3,938 | ~185,000 | ~47.0 |
| EPYC 9575F | 4,114 | ~197,000 | ~47.9 |
| EPYC 9655P | 4,055 | ~210,000 | ~51.8 |
How to read these numbers correctly
Multi-thread scores represent sustained throughput under efficient scaling. Single-thread scores
influence coordination efficiency, latency, and tail performance. Performance-per-euro metrics
support planning but do not capture licensing, power, or cooling constraints.
In real deployments, memory bandwidth, NUMA locality, and thermal limits often define the true
performance ceiling.
Practical takeaway
The AMD EPYC 9575F is a strong choice for high-core workloads that still benefit from elevated
per-core performance. It fits well in enterprise applications, virtualization platforms with
latency-sensitive components, licensed software environments, and mixed compute pipelines where
serial phases remain significant.
For purely throughput-bound workloads, higher-core EPYC models offer better efficiency. For
strictly latency-driven tasks, lower-core frequency-optimized SKUs may be more economical. The
EPYC 9575F delivers high parallel capacity without sacrificing responsiveness.
Product FAQ
The key difference is the operating profile. The 9575F is tuned to sustain higher effective frequencies across a large core budget, which improves performance in mixed workloads with serial phases, uneven thread utilization, and latency-sensitive components. The 9555 is a more balanced throughput SKU that typically prioritizes efficiency and steady scaling over frequency headroom.
It matters when the workload is not perfectly parallel. Examples include application servers with a busy main thread, databases with lock contention, build systems with serialized steps, orchestration-heavy pipelines, and virtualization stacks where management threads and I/O paths influence tail latency. In these cases, higher per-core performance improves completion time even if total core count is the same.
Yes, especially for clusters where VM responsiveness and latency consistency matter, not just maximum VM density. It works best when you keep vCPU oversubscription reasonable and match VM sizing to NUMA boundaries. For maximum consolidation and the lowest cost per vCPU, higher-core non-F SKUs can be more economical.
Avoid spreading a single VM across too many NUMA domains without a clear reason. Prefer fewer, larger VMs only when the application truly scales, and pin memory locality where possible. For mixed fleets, it is often better to run multiple medium-sized VMs aligned to NUMA boundaries rather than a small number of very large VMs that amplify cross-domain memory traffic.
It is a strong fit for OLTP-style databases and mixed workloads where single-thread paths, transaction coordination, and tail latency influence user-facing performance. It is less compelling for purely analytical, fully parallel query engines that primarily want maximum throughput and may benefit more from higher-core SKUs.
This is where the 9575F can be strategically valuable. If the software cost scales with core count, you can keep the core budget fixed at 64 while gaining additional performance from higher effective frequency. The decision depends on whether the workload benefits from per-core uplift more than it benefits from adding more licensed cores.
It fits well for mixed HPC pipelines that include serial stages, preprocessing, orchestration, or workloads sensitive to per-core performance. For embarrassingly parallel jobs that saturate all cores continuously, higher-core SKUs generally provide better throughput density and often better performance-per-euro.
Very important. Fully populating memory channels with balanced DIMM placement improves bandwidth and reduces latency variability. For virtualization and data processing, uneven channel population can become a bottleneck that masks CPU advantages, especially at high utilization.
High sustained frequency increases thermal density. Adequate airflow, a properly sized heatsink, and strong VRM design on the server platform are required to maintain stable boost behavior. In dense racks, power capping and thermal throttling can reduce the expected advantage of an F SKU if the platform is not engineered for it.
Often yes. Many GPU nodes are limited by CPU-side preprocessing, data feeding, orchestration, and I/O coordination rather than raw CPU throughput. Higher per-core performance helps keep GPUs fed, reduces pipeline stalls, and improves end-to-end job time, especially when CPU work sits on the critical path.
It is not ideal for scenarios where you primarily want maximum vCPU density, the lowest cost per thread, or peak throughput on fully parallel batch workloads. In those cases, higher-core non-F SKUs typically deliver better consolidation efficiency and better throughput per euro.
Choose the 9575F when the workload benefits from high frequency at a fixed 64-core budget and when serial or latency-sensitive components materially affect completion time. Favor higher-core SKUs when your workload scales close to linearly with additional threads and you are optimizing for throughput per socket and consolidation efficiency.
Payment & Shipping methods
Fast and reliable delivery across the European Union
Estimated transit time: 14–21 days from order confirmation. Worldwide shipping is available for customers outside the EU.
All orders are processed within 24 hours after confirmation. Tracking information is provided as soon as the parcel leaves our logistics center.
Multiple Secure Payment Methods
We accept: Visa, MasterCard, PayPal, Bank Transfer, Klarna, Stripe, Revolut Pay, Google Pay, Apple Pay, and USDT (TRC20) cryptocurrency payments.
All transactions are encrypted and processed via certified payment gateways for your security.
Additional Notes
- Delivery times may vary depending on customs clearance and carrier schedules.
- Large or custom-built items may require additional handling time.
- Shipments are insured until delivered to the customer.
- We do not deliver to P.O. boxes or military addresses.
Customers Also Loved
New
New
New
New
New
New
New
Reviews
There are no reviews yet.