Skip to main content

Server CPU Benchmarks:
Intel® Xeon® 6th Generation vs AMD EPYC™ 5th Generation

CPU Benchmark Comparison for High-Performance Computing
Based on empirically measured LINPACK, STREAM, and HPCG Benchmarks for Scientific Computing Centers and HPC Clusters

What to expect from this CPU comparison

Comprehensive CPU benchmark comparison of current Intel® Xeon® and AMD EPYC™ processors for HPC applications –
conducted under controlled conditions at the MEGWARE Benchmark Center.

 

HPC-Relevant Benchmarks

Comparison of the computational and memory performance of selected Intel® Xeon® and AMD EPYC™ processors based on the established LINPACK, STREAM, and HPCG benchmarks, with a focus on typical requirements of scientific computing centers.

Memory Bandwidth Analysis

Specific investigation of the impact of memory technology on measured bandwidth, including a direct comparison of DDR5-6400 RDIMMs and DDR5-8800 MRDIMMs.

HPC Workload-Specific Interpretation

Interpretation of benchmark results within the context of compute-intensive and memory-bound HPC workloads.

Tested Configurations

Dual-socket HPC systems with fully populated memory channels and DDR5 ECC memory for servers

 

Intel® Xeon® Processors


Intel® Xeon® 6972P 
Cores: 96 pro CPU
TDP: 500 W
Base Frequency: 2.0 GHz
Memory: DDR5-6400
Intel® Xeon® 6980P  
Cores: 128 pro CPU
TDP: 500 W
Base Frequency: 2.0 GHz
Memory: DDR5-6400
Intel® Xeon® 6980P + MRDIMM 
Cores: 128 pro CPU
TDP: 500 W
Base Frequency: 2.0 GHz
Memory: DDR5-8800

 

AMD EPYC™ Processors


AMD EPYC™ 9655 
Cores: 96 pro CPU
TDP: 400 W
Base Frequency: 2.6 GHz
Memory: DDR5-6400
AMD EPYC™ 9745 
Cores: 128 pro CPU
TDP: 400 W
Base Frequency: 2.4 GHz
Memory: DDR5-6400
AMD EPYC™ 9965 
Cores: 192 pro CPU
TDP: 500 W
Base Frequency: 2.25 GHz
Memory: DDR5-6400

 

Three crucial benchmarks in a single whitepaper

Our CPU comparison includes LINPACK, STREAM, and HPCG – for a complete performance evaluation of all tested Intel® Xeon® and AMD EPYC™ systems

 

LINPACK

Evaluation of the maximum achievable computational performance. LINPACK provides metrics for peak performance in TFLOPS as well as energy efficiency relative to TDP.
 

STREAM

Measurement of the maximum achievable memory bandwidth at both system and core levels. STREAM illustrates how efficiently data is delivered from main memory and highlights potential bandwidth bottlenecks.
 

HPCG

Analysis of performance in sparse linear algebra with irregular memory access patterns. HPCG is suitable for evaluating system performance in memory-bound HPC workloads beyond peak computational capability.

Download the CPU comparison as a whitepaper

The complete whitepaper documents the conducted measurements,
the used system configurations, and the underlying test methodology
in a clear and reproducible manner.

Complete Measurement Results
Detailed benchmark results of all tested Intel® Xeon® and AMD EPYC™ systems, including configurations, measurements, and key metrics.

 

Transparent Test Methodology
Description of the used benchmarks, the software and BIOS configurations, and the test conditions at the MEGWARE Benchmark Center.

 

Interpretations of the Results
Objective assessment of the measured results in the context of compute- and memory-intensive HPC applications.

 

CPU Benchmark 2026 Whitepaper

I accept the privacy policy and consent to the processing of my data.