AMD EPYC 7F72 vs Intel Xeon 6731P Comparison

The Intel Xeon 6731P and AMD EPYC 7F72 are both high-end server/workstation processors that land in the 96th percentile of all CPUs, but they achieve that standing through very different means. The Xeon 6731P wins 12 of the 17 head-to-head benchmark comparisons, while the EPYC 7F72 takes 5, yet the margin of victory in many of those wins is razor-thin. The data reveals a fundamental split: the Intel part dominates raw compute throughput and extended instruction workloads, while the AMD part counters with superior single-thread performance and specific memory-latency-sensitive tasks. Average benchmark scores place the Xeon 6731P at 87,756 against the EPYC 7F72's 85,072, a 3.2% gap that flips the expected narrative of a newer part easily eclipsing an older one.

Where Each One Wins

The Intel Xeon 6731P is the clear winner for compute-heavy, parallel workloads. Its most decisive victories come in floating-point math, where it scores 157,330 against the EPYC's 108,437, a 45.1% advantage. Extended instruction performance follows a similar pattern, with the Xeon scoring 65,656 versus 46,936, a 39.9% lead. The Xeon also wins in prime number finding (541 vs 498, an 8.6% lead), integer math (198,761 vs 181,103, a 9.8% lead), and physics simulations (7,105 vs 6,459, a 10% lead). These results indicate the Xeon 6731P is better suited for scientific computing, financial modeling, or any workload that stresses arithmetic logic units and vectorized instructions.

The AMD EPYC 7F72 wins a different set of categories. Its most significant victory is in data encryption, where it scores 56,261 against the Xeon's 40,087, a commanding 28.7% advantage. It also wins in random string sorting (102,436 vs 88,019, a 14.1% lead), data compression (808,795 vs 799,474, a 1.2% edge), and single-thread performance (2,384 vs 2,107, an 11.6% lead). The single-thread win is particularly noteworthy, as it comes despite the EPYC having a lower boost clock. The EPYC's wins suggest it is better for database operations, encryption-heavy applications, and workloads that depend on low-latency memory access rather than raw throughput.

The Cinebench results are essentially a tie. Across all six Cinebench R15, R20, and R23 tests, the Xeon wins by margins of 0.1% to 0.2%, which is within noise. This means the two processors are functionally equivalent for general rendering workloads, despite the Xeon having 32 cores versus the EPYC's 24. The PassMark multithread score also lands at a near-deadlock: 52,790 for Intel versus 52,740 for AMD. The data shows that the EPYC's higher base clock and larger L3 cache compensate for its core count disadvantage in these specific tests.

Architecture Differences

The architectural divide is stark. The Intel Xeon 6731P uses Granite Rapids architecture on a 5 nm process fabricated by Intel, with a die size of 598 mm². It packs 32 cores and 64 threads, with a base clock of 2.50 GHz and a boost clock of 4.10 GHz. Its cache hierarchy consists of 112 KB of L1 per core, 2 MB of L2 per core, and 144 MB of shared L3 cache. It supports DDR5 memory over an eight-channel bus, delivering 409.6 GB/s of memory bandwidth, and uses Intel Socket 4710 with PCIe Gen 5 (136 lanes).

The AMD EPYC 7F72 is built on the older Zen 2 architecture using a 7 nm process from TSMC. It has 24 cores and 48 threads, with a higher base clock of 3.20 GHz but a lower boost clock of 3.70 GHz. The die size is just 74 mm², with 3,800 million transistors. Its cache layout differs significantly: 96 KB of L1 per core, 512 KB of L2 per core, and a much larger 192 MB of shared L3 cache. The EPYC uses DDR4 memory across an eight-channel bus, providing 204.8 GB/s of bandwidth—exactly half of the Intel part's figure. It uses AMD Socket SP3 and PCIe Gen 4.

These architectural choices explain the benchmark splits. The Xeon's newer 5 nm process and higher boost clock drive its floating-point and integer math wins. The EPYC's larger L3 cache and higher base clock contribute to its single-thread and encryption advantages. The memory bandwidth difference is dramatic—409.6 GB/s versus 204.8 GB/s—yet the EPYC still manages to win in data compression and random string sorting, suggesting that for those workloads, cache size matters more than raw memory throughput. The core count difference (32 vs 24) is offset by the EPYC's higher base clock, which explains the near-identical Cinebench scores.

Head-to-Head Benchmarks

The largest single win for the Intel Xeon 6731P is in PassMark floating-point math, where it outperforms the EPYC 7F72 by 45.1%. This is an enormous margin that dwarfs any other result in the comparison. The Xeon also posts a 39.9% win in extended instructions. These two results alone establish the Xeon as the superior part for number-crunching and SIMD-heavy code. The Xeon's other wins are more modest: 10% in physics, 9.8% in integer math, and 8.6% in prime number finding.

The AMD EPYC 7F72's biggest win is in data encryption, where it leads by 28.7%. This is a substantial margin that indicates a hardware advantage in cryptographic operations. The EPYC also wins single-thread performance by 11.6%, which is notable given that the Xeon has a 0.40 GHz higher boost clock. Random string sorting goes to the EPYC by 14.1%, and data compression by 1.2%. In the Cinebench tests, the Xeon wins all six by margins of 0.1% or 0.2%, which are effectively negligible. The same applies to the PassMark multithread test, where the Xeon wins by 0.1%.

The pattern is clear: the Xeon wins big where it wins, and loses big where it loses. The EPYC's wins are more evenly distributed across encryption, sorting, and single-thread tasks. The 28.7% encryption gap is the second-largest delta in the entire comparison, and it is a significant factor for any server handling encrypted traffic or secure communications. The Xeon's 45.1% floating-point win is the largest, but it is offset by the EPYC's ability to win in several practical server workloads.

FAQ

Q: Which processor has more cores?

A: The Intel Xeon 6731P has 32 cores and 64 threads, while the AMD EPYC 7F72 has 24 cores and 48 threads.

Q: Does the AMD EPYC 7F72 have higher clock speeds?

A: The EPYC has a higher base clock at 3.20 GHz versus the Xeon's 2.50 GHz, but the Xeon has a higher boost clock at 4.10 GHz versus the EPYC's 3.70 GHz.

Q: Why does the EPYC win single-thread performance despite a lower boost clock?

A: The EPYC scores 2,384 in PassMark single-thread versus the Xeon's 2,107, an 11.6% advantage. This is likely due to the EPYC's larger 192 MB shared L3 cache and higher base clock, which can benefit latency-sensitive single-threaded workloads.

Q: Which processor has better memory bandwidth?

A: The Intel Xeon 6731P has 409.6 GB/s of memory bandwidth, exactly double the EPYC's 204.8 GB/s. The Xeon also supports DDR5 memory, while the EPYC uses DDR4.

Q: What is the biggest performance gap between the two?

A: The Xeon wins floating-point math by 45.1%, while the EPYC wins data encryption by 28.7%. These are the two largest deltas in the comparison.

Q: Are the Cinebench scores meaningfully different?

A: No. The Xeon wins all six Cinebench tests by 0.1% or 0.2%, which is within measurement noise. The two CPUs are functionally equivalent for rendering workloads.

Specification Differences

| Specification | Intel Xeon 6731P | AMD EPYC 7F72 |

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

| Cores | 32 | 24 |

| Threads | 64 | 48 |

| Base Clock | 2.50 GHz | 3.20 GHz |

| Boost Clock | 4.10 GHz | 3.70 GHz |

| TDP | 245 W | 240 W |

| Socket | Intel Socket 4710 | AMD Socket SP3 |

| Architecture | Granite Rapids | Zen 2 |

| Process Node | 5 nm | 7 nm |

| Foundry | Intel | TSMC |

| Die Size | 598 mm² | 74 mm² |

| L1 Cache | 112 KB (per core) | 96 KB (per core) |

| L2 Cache | 2 MB (per core) | 512 KB (per core) |

| L3 Cache | 144 MB (shared) | 192 MB (shared) |

| Memory Support | DDR5 | DDR4 |

| Memory Bandwidth | 409.6 GB/s | 204.8 GB/s |

| PCIe | Gen 5, 136 Lanes | Gen 4 |

| Launch MSRP | $2700 | N/A |

The Verdict

The data supports a clear split decision based on workload type. For scientific computing, engineering simulations, or any task dominated by floating-point operations, the Intel Xeon 6731P is the definitive choice. Its 45.1% lead in floating-point math and 39.9% lead in extended instructions are overwhelming advantages that no other benchmark result can counterbalance. The Xeon also offers double the memory bandwidth (409.6 GB/s vs 204.8 GB/s) and a more modern DDR5 platform, which future-proofs it for memory-intensive applications.

The AMD EPYC 7F72 is the better pick for security and data-processing workloads. Its 28.7% advantage in data encryption is a decisive factor for servers handling SSL/TLS, VPN traffic, or any cryptographic operations. The 14.1% win in random string sorting and the 11.6% single-thread advantage make it better suited for database indexing, text processing, and workloads with irregular memory access patterns. The EPYC's larger 192 MB L3 cache clearly compensates for its older DDR4 memory and lower core count in these scenarios.

For general rendering or mixed workloads, the choice is a toss-up. The Cinebench scores are within 0.2% across the board, and the PassMark multithread score differs by just 0.1%. The 3.2% average benchmark gap in favor of the Xeon is small enough to be irrelevant in real-world use. The Xeon's 32-core advantage does not translate into meaningful wins in these tests, suggesting that the EPYC's higher base clock and cache configuration are equally effective at keeping threads fed.

The Xeon 6731P is the newer part, released in 2025 versus the EPYC's 2020 release, and it shows in the process node (5 nm vs 7 nm) and memory support. However, the EPYC proves that architecture efficiency can overcome generational gaps. The final recommendation: choose the Xeon for raw compute and memory bandwidth, choose the EPYC for encryption and latency-sensitive tasks, and consider either for balanced server workloads. The Xeon's $2700 launch MSRP is the only pricing data available, but the EPYC's lack of a listed MSRP prevents a direct cost comparison.