AMD EPYC 9654

AMD EPYC™ 9654 server CPU

AMD EPYC 9654 server CPU (96-core/192-thread, 2.40GHz base and up to 3.70GHz max boost frequency, 384MB L3 Cache, 360W TDP, up to 6TB DDR5-4800 memory, 2P/1P)

Realize exceptional time-to-results and energy efficiency for your business-critical applications with AMD EPYC 9004 Series processors for modern data centers.

Supercharge your applications and accelerate your business results in the data center with leadership performance and energy-efficiency.

We take the mystery out of CPU selection with AMD EPYC processors: Just choose the core count, frequency, and L3 cache size your workload requires. Memory capacity, security features, I/O bandwidth, and the rest are included at no extra cost. Fourth-generation AMD EPYC processors build upon our innovative hybrid multi-chip architecture to deliver on the defining metric of our day: the ultimate combination of performance and efficiency.

Enhance data center performance and energy efficiency to accelerate business results. Get high-performance density, energy efficiency and full-cloud compatibility for on-premises and cloud-native environments.

    Advanced Data Center Performance and Efficiency

  • BREAKTHROUGH PERFORMANCE: Accelerate productivity, make more informed decisions, and speed time to market with a common platform that delivers the performance you need.
  • MODERNIZE YOUR INFRASTRUCTURE: When you upgrade your infrastructure, you want the most innovative design, high density, excellent energy efficiency, and low total cost of ownership.
  • COMPUTE WITH CONFIDENCE: To propel your business forward, you need to confidently navigate today’s risks, complexities, and requirements.
  • High Frequency and Core Count: High frequency (3 GHz+) allows the CPU to process instructions faster, while a high core count (8+ cores) enables parallel processing of multiple tasks simultaneously. This is crucial for handling extensive data tasks. Additionally, it facilitates faster post-processing of results generated by the AI model.
  • Large Cache Size: A large cache size (20MB+) reduces the need to fetch data from slower main memory, improving performance. This is essential for facilitating fast access to massive datasets in training and running AI models and helps minimize latency, leading to quicker training times and faster execution.
  • High Memory Bandwidth and High Performance I/O: CPUs with high memory bandwidth (50GB/s+) and high-performance I/O interfaces are well-suited for AI tasks enable efficient data flow, enabling rapid data transfer between the CPU and GPU, and helping minimize bottlenecks during AI model training and inference.
  • Energy-Efficient Cores: AMD EPYC processors have energy-efficient cores that enable efficient GPU-accelerated systems, optimizing high performance, energy efficiency, memory, and IO scalability. AMD EPYC processors power the world’s most energy-efficient servers, delivering exceptional performance and helping reduce energy costs.
  • GPU Ecosystem Compatibility: GPU compatibility enables seamless communication and optimized performance between different components. This allows for efficient utilization of GPU resources and facilitates smooth software integration, leading to faster training times and better model performance
  • 1P and 2P capable 4th Gen AMD EPYC processor based on 5nm process technology and “ Zen 4” or ultra-efficient “ Zen 4c” chiplet microarchitecture
    • Up to 128 cores and 256 threads (“Zen 4c”); up to 96 cores and 192 threads (“Zen 4”)
  • Up to 160 lanes of PCIe 5.0 I/O1
  • Up to 6 TB of DDR5-4800 ECC memory channels across 24 DIMM slots
  • Up to 384 MB L3 cache
  • Up to 64 lanes of CXL™ 2.0 industry standard cache-coherent interconnect
  • Industry-standard AVX-512 support (2x256b data path)
  • AMD Infinity Guard