AMD EPYC 4564P vs Intel Xeon 6520P Comparison

The AMD EPYC 4564P and Intel Xeon 6520P are both 96th-percentile server/workstation processors, but the benchmark data shows a clear split in their strengths. The AMD EPYC 4564P wins the majority of head-to-head tests (13 of 17), dominating in single-threaded and general multi-threaded workloads, while the Intel Xeon 6520P counters with decisive victories in physics, prime number finding, and floating-point math. The average benchmark scores place the EPYC 4564P at 95,183 versus 93,786 for the Xeon 6520P, a 1.5% gap in favor of AMD.

Head-to-Head Benchmarks

The AMD EPYC 4564P’s most striking advantage comes in single-thread performance. In PassMark single-thread testing, the EPYC scores 4,292 against the Xeon’s 3,356, a commanding 27.9% lead. This edge carries into Cinebench single-core tests: the EPYC wins all three Cinebench single-core variants by a consistent 3.9%, with scores of 791, 3,296, and 7,848 in R15, R20, and R23, respectively, versus 761, 3,171, and 7,552 for Intel.

Multi-core results tell a similar story, though with smaller margins. The EPYC 4564P leads every Cinebench multi-core test by exactly 3.9%: 5,603 vs 5,392 in R15, 23,348 vs 22,467 in R20, and 55,592 vs 53,495 in R23. The PassMark multithread score shows a 2.3% win for AMD (64,357 vs 62,936), while integer math goes to AMD by 6.8% (228,761 vs 214,288). Random string sorting also favors AMD at 7.8% (103,202 vs 95,736), and data compression is 2% ahead (858,105 vs 841,518). The largest single multi-threaded win for AMD is data encryption, where the EPYC scores 50,708 against 45,188, a 12.2% advantage.

The Intel Xeon 6520P’s wins are fewer but substantial. The most dramatic is PassMark physics, where Intel scores 7,209 versus AMD’s 3,392 — a 52.9% blowout. Prime number finding goes to Intel by 30.6% (526 vs 365), and floating-point math favors Intel by 13.4% (162,862 vs 141,017). The only other Intel win is extended instructions, where the Xeon scores 64,557 against 63,136, a modest 2.2% edge. Notably, the Xeon 6520P achieves these wins despite having a lower average benchmark score, highlighting its specialized strengths.

The Verdict

The data supports a split decision based on workload. The AMD EPYC 4564P is the clear choice for single-threaded responsiveness, general server tasks, encryption, and integer-heavy workloads. Its 27.9% lead in single-thread PassMark and 12.2% lead in data encryption make it superior for database transactions, web serving, and any application where per-core speed matters. The consistent 3.9% Cinebench multi-core advantage also makes it the better pick for mainstream rendering and compilation workloads.

The Intel Xeon 6520P is the definitive winner for physics simulation and floating-point math. The 52.9% margin in PassMark physics and 13.4% edge in floating-point math indicate the Xeon’s architecture is better suited for scientific computing, financial modeling, and simulation workloads. Its 30.6% win in prime number finding further reinforces this, suggesting strength in integer-heavy algorithmic tasks that involve modular arithmetic.

For mixed workloads, the EPYC 4564P’s higher average score (95,183 vs 93,786) and 13 benchmark wins give it the overall edge. However, buyers whose primary applications are physics-based or floating-point intensive should strongly consider the Xeon 6520P, as its specialized wins are far larger than AMD’s general-purpose advantages. Both processors sit at the 96th percentile, but the AMD part offers more consistent performance across a broader range of tests.

Architecture Differences

The AMD EPYC 4564P uses the Zen 4 architecture, codenamed Raphael, built on a 5 nm TSMC process. It features 16 cores and 32 threads, with a base clock of 4.50 GHz and a boost clock of 5.70 GHz. The Intel Xeon 6520P uses the Granite Rapids architecture, also on a 5 nm process but fabricated by Intel, with 24 cores and 48 threads at a 2.40 GHz base and 4.00 GHz boost. The EPYC has a TDP of 170 W, while the Xeon has a TDP of 210 W.

Cache configurations differ significantly. The EPYC 4564P has 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3. The Xeon 6520P has 112 KB of L1 per core, 2 MB of L2 per core, and 144 MB of shared L3. This gives Intel a 2.25x advantage in L3 cache, which likely contributes to its floating-point and physics wins. The EPYC uses a dual-channel DDR5 memory bus with 83.2 GB/s bandwidth, while the Xeon uses an eight-channel DDR5 bus with 409.6 GB/s — a nearly 5x bandwidth advantage for Intel that explains its memory-heavy workload performance.

PCIe connectivity also differs: the EPYC provides 28 Gen 5 lanes, while the Xeon provides 88 Gen 5 lanes. The EPYC includes integrated Radeon Graphics, while the Xeon has no integrated graphics. Both support ECC memory, but the EPYC uses AMD Socket AM5 while the Xeon uses Intel Socket 4710. The EPYC has a die size of 2x 71 mm² with 13,140 million transistors, while the Xeon has a single 598 mm² die with transistor count not listed. The EPYC launched on 2024-05-20 with a launch MSRP of $699; the Xeon launched on 2025-02-23 with a launch MSRP of $1295.

FAQ

Q: Which processor has better single-thread performance?

A: The AMD EPYC 4564P wins decisively. It scores 4,292 in PassMark single-thread versus 3,356 for the Xeon 6520P, a 27.9% advantage. Cinebench R23 single-core also goes to AMD at 7,848 vs 7,552.

Q: Why does the Intel Xeon 6520P win in physics benchmarks?

A: The Xeon scores 7,209 in PassMark physics versus 3,392 for the EPYC, a 52.9% margin. This correlates with its larger 144 MB L3 cache and 409.6 GB/s memory bandwidth, which likely benefit physics simulation workloads.

Q: Which processor supports more memory bandwidth?

A: The Intel Xeon 6520P has an eight-channel DDR5 memory bus delivering 409.6 GB/s. The AMD EPYC 4564P has a dual-channel bus at 83.2 GB/s, making the Xeon nearly 5x faster in memory bandwidth.

Q: What is the core and thread count difference?

A: The Xeon 6520P has 24 cores and 48 threads, while the EPYC 4564P has 16 cores and 32 threads. Despite fewer cores, the EPYC wins most multi-threaded benchmarks due to higher clocks.

Q: Which processor has more PCIe lanes?

A: The Intel Xeon 6520P has 88 Gen 5 lanes from the CPU, while the AMD EPYC 4564P has 28 Gen 5 lanes. This gives Intel a significant advantage for systems requiring many expansion cards or NVMe drives.

Q: Do both processors support ECC memory?

A: Yes, both the AMD EPYC 4564P and Intel Xeon 6520P support ECC memory, making them suitable for error-sensitive server and workstation applications.

Where Each One Wins

The AMD EPYC 4564P wins in 13 of 17 head-to-head benchmarks, making it the broader performer. Its biggest wins are in single-thread PassMark (27.9% ahead), data encryption (12.2%), random string sorting (7.8%), and integer math (6.8%). These results indicate it is the better choice for general-purpose server workloads, database applications, web services, and any task requiring fast per-core execution. Its consistent 3.9% lead across all Cinebench tests also makes it the better option for rendering and content creation, despite having fewer cores than the Xeon.

The Intel Xeon 6520P wins in 4 benchmarks, but its margins are substantial. The physics win at 52.9% is the largest single delta in the comparison, making the Xeon the clear choice for simulation, physics engines, and scientific computing. The floating-point math win at 13.4% and prime number finding at 30.6% further solidify its position for numerical analysis, financial risk modeling, and computational research. The Xeon’s 2.2% win in extended instructions also suggests it handles vectorized code slightly better than the EPYC.

For memory-intensive workloads, the Xeon’s eight-channel 409.6 GB/s bandwidth versus the EPYC’s dual-channel 83.2 GB/s is a decisive architectural advantage, even if the benchmark data only partially reflects it. Similarly, the Xeon’s 88 PCIe lanes versus 28 for the EPYC makes it the better platform for GPU-heavy compute nodes or storage servers. The EPYC, however, offers integrated Radeon Graphics, eliminating the need for a discrete GPU in basic display scenarios.

Specification Differences

| Specification | AMD EPYC 4564P | Intel Xeon 6520P |

|---|---|---|

| Cores | 16 | 24 |

| Threads | 32 | 48 |

| Base Clock | 4.50 GHz | 2.40 GHz |

| Boost Clock | 5.70 GHz | 4.00 GHz |

| TDP | 170 W | 210 W |

| Socket | AMD Socket AM5 | Intel Socket 4710 |

| Architecture | Zen 4 | Granite Rapids |

| Process Node | 5 nm (TSMC) | 5 nm (Intel) |

| Die Size | 2x 71 mm² | 598 mm² |

| L1 Cache | 64 KB (per core) | 112 KB (per core) |

| L2 Cache | 1 MB (per core) | 2 MB (per core) |

| L3 Cache | 64 MB (shared) | 144 MB (shared) |

| Memory Bus | Dual-channel | Eight-channel |

| Memory Bandwidth | 83.2 GB/s | 409.6 GB/s |

| PCIe | Gen 5, 28 Lanes | Gen 5, 88 Lanes |

| Integrated Graphics | Radeon Graphics | N/A |

| Launch Date | 2024-05-20 | 2025-02-23 |

| Launch MSRP | $699 | $1295 |

| Part Number | 100-000001476 | SRVNQ |