AMD EPYC 4564P vs Intel Xeon 6731P Comparison

Where Each One Wins

The benchmark record splits these two server processors into clearly different use cases. The AMD EPYC 4564P wins 13 of the 17 head-to-head comparisons, while the Intel Xeon 6731P takes 4. That raw count understates the nature of the split: AMD dominates single-thread, multi-thread, and memory-sensitive workloads, while Intel wins in specific math and physics-oriented tasks.

The AMD EPYC 4564P is the general-purpose throughput champion. Its wins include every Cinebench test (R15, R20, R23, both single and multi-core), PassMark multithread, integer math, data compression, data encryption, random string sorting, and both single-thread tests. The margins are often substantial. In PassMark single-thread, it leads by 103.7%, which is the largest delta in the entire comparison. In data encryption, it leads by 26.5%. In Cinebench R23 multicore, it leads by 23.9%. This is a processor that dominates anything clock-speed sensitive or latency-bound.

The Intel Xeon 6731P wins in four specific areas: PassMark extended instructions, find prime numbers, floating point math, and physics. The physics win is enormous at 52.3% over AMD. The find prime numbers win is even larger at 32.5% over AMD. These are not trivial workloads; they represent integer-heavy, branch-predictor-stressing, and floating-point-simulation tasks. For scientific computing, physics simulation, or prime-number research, Intel has a clear edge. However, the floating point math lead is more modest at 10.4%, and the extended instructions lead is just 3.8%.

The pattern is consistent across all Cinebench versions. AMD leads by 23.9% or 24.0% in every single Cinebench test, regardless of core count. That consistency suggests the advantage comes from per-core performance, not just core count. The AMD has 16 cores versus Intel's 32, yet still wins multicore tests by nearly a quarter. This is a case where fewer, faster cores beat more, slower cores in the majority of measured scenarios.

Architecture Differences

The two processors come from fundamentally different design philosophies. The AMD EPYC 4564P uses the Zen 4 architecture on the Raphael codename, built on a 5 nm TSMC process. The Intel Xeon 6731P uses the Granite Rapids architecture on the Granite Rapids-SP codename, also on a 5 nm process but from Intel's own foundry. Both are active production parts.

The core counts differ dramatically. AMD packs 16 cores and 32 threads, while Intel doubles that to 32 cores and 64 threads. Yet AMD's base clock of 4.50 GHz and boost clock of 5.70 GHz dwarf Intel's 2.50 GHz base and 4.10 GHz boost. This explains the single-thread results: AMD's clock advantage is enormous, and it shows in every single-core benchmark.

Cache hierarchies also differ. AMD uses 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3. Intel uses 112 KB of L1 per core, 2 MB of L2 per core, and 144 MB of shared L3. Intel has more cache per core and more total L3, but AMD's higher clocks compensate in most tests. The die size tells a story too: Intel's die is 598 mm², while AMD's is 2x 71 mm² (two chiplets). AMD's transistor count is listed at 13,140 million; Intel's is not recorded.

Memory architecture separates them further. Both support DDR5 and ECC, but AMD uses a dual-channel memory bus with 83.2 GB/s bandwidth, while Intel uses an eight-channel bus with 409.6 GB/s bandwidth. Intel has nearly five times the memory bandwidth, which should help in memory-bound workloads. Yet the benchmark data shows AMD winning data compression and encryption, which are often memory-sensitive. This suggests AMD's clock advantage and memory latency characteristics overcome Intel's raw bandwidth.

PCIe connectivity also differs. AMD provides Gen 5 with 28 lanes (CPU only), while Intel provides Gen 5 with 136 lanes (CPU only). Intel has far more expansion capability for GPU or storage arrays. AMD integrates Radeon Graphics; Intel has no integrated graphics. Both are locked multipliers, and both target the server/workstation segment.

Head-to-Head Benchmarks

The Cinebench results are uniformly in AMD's favor. In R15 multicore, AMD scores 5603 versus Intel's 4522, a 23.9% lead. In R15 single-core, AMD scores 791 versus 638, a 24.0% lead. R20 multicore shows 23348 versus 18845, again 23.9%. R20 single-core shows 3296 versus 2660, again 23.9%. R23 multicore shows 55592 versus 44871, 23.9%. R23 single-core shows 7848 versus 6334, 23.9%. The consistency is remarkable: AMD wins every Cinebench test by essentially the same margin.

PassMark multithread goes to AMD at 64357 versus 52790, a 21.9% lead. Integer math goes to AMD at 228761 versus 198761, a 15.1% lead. Random string sorting goes to AMD at 103202 versus 88019, a 17.2% lead. Data compression goes to AMD at 858105 versus 799474, a 7.3% lead. Data encryption goes to AMD at 50708 versus 40087, a 26.5% lead. The single-thread tests show the biggest gap: AMD scores 4292 versus Intel's 2107, a 103.7% lead. That is more than double Intel's score.

Intel's wins are concentrated but significant. PassMark physics shows Intel at 7105 versus AMD's 3392, a 52.3% lead. Find prime numbers shows Intel at 541 versus 365, a 32.5% lead. Floating point math shows Intel at 157330 versus 141017, a 10.4% lead. Extended instructions shows Intel at 65656 versus 63136, a 3.8% lead. The physics and prime number results are particularly notable because they suggest Intel's architecture handles certain algorithmic patterns much better, even with lower clocks.

The overall average benchmark scores reflect the split. AMD's average is 95183, while Intel's is 87756. AMD's nearest rival is the AMD EPYC 9175F at 95615, just 0.5% higher. Intel's nearest rival is the Intel Xeon 6736P at 87864, just 0.1% higher. Both sit at the 96th percentile of all CPUs, so both are top-tier parts.

FAQ

Q: Which processor has higher single-thread performance?

A: The AMD EPYC 4564P is dramatically faster in single-thread tests. In PassMark single-thread, it scores 4292 versus Intel's 2107, a 103.7% advantage. In Cinebench R23 single-core, it scores 7848 versus 6334, a 23.9% lead.

Q: Does the Intel Xeon 6731P win any benchmarks?

A: Yes, it wins 4 of 17 head-to-head tests: PassMark physics, find prime numbers, floating point math, and extended instructions. The physics win is the largest at 52.3% over AMD.

Q: How do the core counts compare?

A: The Intel Xeon 6731P has 32 cores and 64 threads, while the AMD EPYC 4564P has 16 cores and 32 threads. Despite having half the cores, AMD wins the multicore Cinebench tests by 23.9%.

Q: What is the memory bandwidth difference?

A: The Intel Xeon 6731P has an eight-channel memory bus with 409.6 GB/s bandwidth. The AMD EPYC 4564P has a dual-channel bus with 83.2 GB/s. Intel has roughly five times the bandwidth.

Q: Which processor has more PCIe lanes?

A: The Intel Xeon 6731P provides Gen 5 with 136 lanes (CPU only). The AMD EPYC 4564P provides Gen 5 with 28 lanes (CPU only). Intel has nearly five times the expansion capacity.

Q: What are the launch MSRP values?

A: The AMD EPYC 4564P has a launch MSRP of $699. The Intel Xeon 6731P has a launch MSRP of $2700.

Specification Differences

| Specification | AMD EPYC 4564P | Intel Xeon 6731P |

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

| Cores | 16 | 32 |

| Threads | 32 | 64 |

| Base Clock | 4.50 GHz | 2.50 GHz |

| Boost Clock | 5.70 GHz | 4.10 GHz |

| TDP | 170 W | 245 W |

| Socket | AMD Socket AM5 | Intel Socket 4710 |

| Architecture | Zen 4 | Granite Rapids |

| Codename | Raphael | Granite Rapids |

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

| Die Size | 2x 71 mm² | 598 mm² |

| Transistors | 13,140 million | Not recorded |

| 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, 136 Lanes |

| Integrated Graphics | Radeon Graphics | N/A |

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

| Launch MSRP | $699 | $2700 |

| Part Number | 100-000001476 | SRVNR |

The Verdict

The data points to a clear conclusion: the AMD EPYC 4564P is the superior processor for the majority of workloads. It wins 13 of 17 benchmarks, including all Cinebench tests, all single-thread tests, and most PassMark workloads. Its single-thread advantage of 103.7% is the single largest margin in the comparison. Even with half the cores of Intel, it wins multicore tests by 23.9%. For general server workloads, database operations, encryption, compression, and any clock-sensitive task, AMD is the choice.

The Intel Xeon 6731P is the better option for a narrower set of workloads. Its physics performance is 52.3% ahead, and its prime number finding is 32.5% ahead. For scientific simulation, physics engines, or floating-point-heavy computation, Intel's architecture delivers results AMD cannot match. Its eight-channel memory bus and 409.6 GB/s bandwidth also provide a platform advantage for memory-hungry applications, even if the benchmark data does not consistently reflect that in the tested scenarios.

The average scores put AMD at 95183 versus Intel's 87756, an 8.5% overall gap. Both are at the 96th percentile of all CPUs. The launch MSRP difference is substantial, but the performance data alone justifies AMD's position as the general-purpose winner. Intel wins where its architecture specializes, but those wins are the exception, not the rule. For most buyers, the AMD EPYC 4564P is the safer choice based on the recorded measurements. For physics and prime-number workloads specifically, the Intel Xeon 6731P is the only rational pick.