AMD EPYC 4585PX vs Intel Xeon 6736P Comparison

FAQ

Q: Which processor wins more benchmark comparisons in the database?

A: The AMD EPYC 4585PX wins 16 of the 17 head-to-head tests recorded, while the Intel Xeon 6736P wins only 1. The sole Intel victory comes in PassMark random string sorting, where it scores 103723 against AMD's 95210, a margin of 8.2%.

Q: How do the two compare in multi-threaded rendering workloads?

A: The AMD EPYC 4585PX leads decisively across all Cinebench multi-core tests. In Cinebench R23 multi-core, AMD scores 60451 versus Intel's 42561, a 42% advantage. The same 42% delta appears in R15 and R20 multi-core, with scores of 6093 vs 4290 and 25389 vs 17875 respectively.

Q: What about single-thread performance?

A: The AMD part is far ahead in single-core tests. In PassMark single-thread, AMD scores 4538 versus Intel's 2024, a massive 124.2% lead. Cinebench R23 single-core shows AMD at 8534 against Intel's 6008, a 42% gap, and R15 single-core shows 860 vs 605, a 42.1% margin.

Q: Which processor has higher overall benchmark standing?

A: The AMD EPYC 4585PX sits at the 97th percentile among all CPUs in the database, with an average benchmark score of 99324. The Intel Xeon 6736P is at the 96th percentile, with an average score of 87864.

Q: How does the AMD processor compare to its nearest rivals?

A: The AMD EPYC 4585PX's nearest rival is the AMD Ryzen Threadripper PRO 5965WX, which scores 98504, putting AMD 0.8% ahead. It is 1.7% behind the Ryzen Threadripper PRO 9955WX (101041) and 1.9% ahead of the Ryzen 9 9955HX3D (97453), while trailing the EPYC 7513 (102244) by 2.9%.

Q: How does the Intel processor fare against its closest competitors?

A: The Intel Xeon 6736P is essentially tied with the Intel Xeon 6731P, which scores 87756, a 0.1% difference. It sits 0.3% ahead of the Xeon w7-2575X (88172), 3% behind the AMD Ryzen AI Max+ 392 (90541), and 3.3% ahead of the AMD EPYC 7F72 (85072).

Architecture Differences

The two processors come from fundamentally different design philosophies. The AMD EPYC 4585PX uses the Zen 5 architecture, codenamed Grado, built on a 4 nm process at TSMC. It packs 16 cores and 32 threads in a dual-chiplet design with a die size of 2x 70.6 mm². The Intel Xeon 6736P uses Granite Rapids architecture, built on Intel's 5 nm process, with a monolithic die measuring 598 mm².

The transistor counts tell a story of density versus scale. AMD's chip contains 16,630 million transistors across its two chiplets, while Intel's larger die has no transistor count listed in the database. The core counts diverge sharply: Intel offers 36 cores and 72 threads, more than double AMD's 16 cores and 32 threads. Yet the AMD part compensates with much higher clocks: 4.30 GHz base and 5.70 GHz boost versus Intel's 2.00 GHz base and 4.10 GHz boost.

Cache hierarchies differ in structure. AMD provides 80 KB of L1 per core, 1 MB of L2 per core, and a large 128 MB L3 cache. Intel gives 112 KB L1 per core, 2 MB L2 per core, and 144 MB of shared L3. AMD's L3 is larger per core (8 MB per core versus 4 MB per core for Intel), which matters for workloads with high data locality.

Memory architecture is a major differentiator. Both support DDR5 and ECC memory, but Intel runs an eight-channel memory bus with 409.6 GB/s bandwidth, while AMD uses a dual-channel bus with 89.6 GB/s. That is a 4.6x bandwidth advantage for Intel, yet the benchmark data shows AMD winning most memory-sensitive tests, suggesting latency and clock speed compensate for the narrower bus.

PCIe connectivity also differs. Intel provides 88 Gen 5 lanes from the CPU, while AMD offers 24 Gen 5 lanes. AMD includes integrated Radeon Graphics; Intel has no integrated graphics. Both are locked multipliers, target the server/workstation segment, and are in active production. AMD launched on 2025-05-12, Intel on 2025-02-23. The AMD part carries launch MSRP of $699, while Intel's launch MSRP is $3351.

The Verdict

The benchmark data points to a clear split: the AMD EPYC 4585PX dominates in almost every measured workload, while the Intel Xeon 6736P wins only in random string sorting. For buyers prioritizing raw compute throughput in rendering, encryption, math operations, and single-threaded responsiveness, the AMD part is the stronger choice across the board. Its 42% lead in all Cinebench multi-core tests and 124.2% lead in PassMark single-thread make it the database's preferred option for those workloads.

The Intel Xeon 6736P makes sense only for workloads that specifically depend on its eight-channel memory bandwidth and 88 PCIe lanes, since those features do not translate into benchmark wins in the recorded tests. Its 36 cores and 72 threads provide more parallelism on paper, but the measured results show AMD's 16-core chip outperforming it in multi-threaded tests anyway. The 8.2% win in random string sorting suggests Intel has some advantage in sorting algorithms, but that single result does not offset 16 losses.

Given the average benchmark scores, AMD sits 13.1% higher overall (99324 vs 87864). The percentile rankings are close, 97th versus 96th, but the delta in actual scores is substantial. For most server and workstation buyers, the AMD EPYC 4585PX delivers better measured performance across the database's test suite.

Specification Differences

| Specification | AMD EPYC 4585PX | Intel Xeon 6736P |

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

| Cores | 16 | 36 |

| Threads | 32 | 72 |

| Base clock | 4.30 GHz | 2.00 GHz |

| Boost clock | 5.70 GHz | 4.10 GHz |

| TDP | 170 W | 205 W |

| Socket | AMD Socket AM5 | Intel Socket 4710 |

| Architecture | Zen 5 | Granite Rapids |

| Process node | 4 nm | 5 nm |

| Foundry | TSMC | Intel |

| Die size | 2x 70.6 mm² | 598 mm² |

| L1 cache | 80 KB per core | 112 KB per core |

| L2 cache | 1 MB per core | 2 MB per core |

| L3 cache | 128 MB | 144 MB shared |

| Memory bus | Dual-channel | Eight-channel |

| Memory bandwidth | 89.6 GB/s | 409.6 GB/s |

| PCIe | Gen 5, 24 lanes | Gen 5, 88 lanes |

| Integrated graphics | Radeon Graphics | N/A |

| Launch MSRP | $699 | $3351 |

Head-to-Head Benchmarks

The AMD EPYC 4585PX wins every Cinebench test by the same approximate margin. In R15 multi-core, AMD scores 6093 against Intel's 4290, a 42% lead. R15 single-core shows 860 vs 605, a 42.1% gap. R20 multi-core repeats the 42% pattern with 25389 vs 17875, and R20 single-core shows 3584 vs 2523, again 42.1%. R23 multi-core gives 60451 vs 42561, a 42% margin, while R23 single-core gives 8534 vs 6008, also 42%.

PassMark results show a more varied picture. The largest single advantage for AMD is in single-thread performance: 4538 vs 2024, a 124.2% lead. The smallest AMD wins are narrow: floating point math at 153219 vs 149770 (2.3%), data encryption at 47224 vs 46236 (2.1%), and physics at 6612 vs 6531 (1.2%). Mid-range AMD wins include extended instructions at 69457 vs 55563 (25%), integer math at 248563 vs 206833 (20.2%), find prime numbers at 547 vs 392 (39.5%), and multithread at 68908 vs 50072 (37.6%).

Data compression shows AMD ahead at 884774 vs 796658, an 11.1% margin. The only Intel win is random string sorting: 103723 vs 95210, meaning AMD trails by 8.2% in that specific test. The wins tally is 16 for AMD and 1 for Intel.

Where Each One Wins

The AMD EPYC 4585PX wins in every category except one. Its strengths lie in single-thread responsiveness, where the 124.2% PassMark lead is decisive, and in rendering workloads, where the consistent 42% Cinebench margins across three versions of the benchmark indicate a structural advantage rather than a test-specific quirk. Integer math, extended instructions, and prime number finding all show AMD ahead by 20% to 39.5%, pointing to strong general-purpose compute. Even in multithreaded tests, where Intel has 36 cores versus AMD's 16, AMD wins by 37.6%, demonstrating that clock speed and architecture efficiency overcome core count.

The Intel Xeon 6736P wins only in random string sorting, with a score of 103723 versus AMD's 95210. That 8.2% advantage suggests Intel's larger L3 cache (144 MB vs 128 MB) or its memory subsystem helps in sorting workloads with large data sets. Intel also has the theoretical advantage of 409.6 GB/s memory bandwidth versus AMD's 89.6 GB/s, plus 88 PCIe lanes versus 24, but the recorded benchmarks do not show these translating into wins elsewhere.

For buyers running sorting-heavy workloads or applications that can exploit eight-channel memory and vast PCIe connectivity, the Intel part has a narrow niche. For everything else in the database's test suite, the AMD EPYC 4585PX is the measured winner. The data shows AMD ahead by double digits in most tests and by over 100% in single-thread performance, making it the default recommendation for mixed server and workstation workloads.