AMD Ryzen Threadripper PRO 5965WX vs Intel Xeon 6520P Comparison

The AMD Ryzen Threadripper PRO 5965WX and Intel Xeon 6520P are both 24-core, 48-thread workstation processors, but benchmark results show they are far from twins. The AMD part wins 12 of the 17 head-to-head comparisons, while the Intel part takes 5, yet the margins in those Intel victories are occasionally dramatic. This analysis breaks down where each processor dominates, what architectural choices drive those results, and who should choose which based strictly on the performance data.

Head-to-Head Benchmarks

The most decisive AMD victory comes in PassMark integer math, where the Ryzen Threadripper PRO 5965WX scores 281,247 against the Xeon 6520P’s 214,288 — a 31.2% advantage. This is the largest single-benchmark gap in either direction, and it signals a substantial edge in general-purpose compute tasks that rely on integer operations. The AMD chip also wins PassMark data encryption by a wide margin, scoring 63,113 versus 45,188, a 39.7% lead that suggests a significant difference in cryptographic workload throughput.

The AMD processor shows a consistent, smaller lead across Cinebench tests. In Cinebench R23 multi-core, it scores 56,400 versus 53,495, a 5.4% advantage, and the same 5.4% delta appears in Cinebench R15 multi-core (5,685 vs 5,392) and R20 multi-core (23,688 vs 22,467). Single-core Cinebench results follow the same pattern: R15 shows 802 vs 761 (5.4%), R20 shows 3,344 vs 3,171 (5.5%), and R23 shows 7,962 vs 7,552 (5.4%). These consistent margins indicate the AMD architecture holds a steady clock-for-clock advantage in rendering workloads.

PassMark data compression goes to AMD with 1,010,067 versus 841,518, a 20% lead, while extended instructions favor AMD by 4.2% (67,300 vs 64,557), and random string sorting by 3.7% (99,287 vs 95,736). PassMark multi-thread also goes to AMD, 66,353 vs 62,936, a 5.4% edge that mirrors the Cinebench multi-core deltas.

The Intel Xeon 6520P’s wins are fewer but striking. The standout is PassMark physics, where Intel scores 7,209 against AMD’s 4,251 — a 41% advantage, the largest margin in either direction. This suggests the Intel processor handles physics simulation workloads far more efficiently, likely due to architectural differences in how the cores handle the specific instruction patterns involved. Intel also wins floating-point math, scoring 162,862 vs 157,708, a 3.2% edge, and takes PassMark single-thread by a razor-thin 0.6% margin (3,356 vs 3,337). The Intel chip also edges out AMD in find prime numbers, 526 vs 523, a 0.6% lead.

Looking at the overall averages, the AMD Ryzen Threadripper PRO 5965WX posts an average benchmark score of 98,504, placing it in the 97th percentile of all CPUs, while the Intel Xeon 6520P averages 93,786, in the 96th percentile. The AMD chip’s nearest rivals include the AMD EPYC 4585PX (99,324, -0.8% delta) and the Ryzen Threadripper PRO 9955WX (101,041, -2.5% delta), while the Intel chip sits closest to the Intel Core Ultra 7 270K Plus (93,785, 0% delta) and trails the AMD EPYC 4564P by 1.5%.

Architecture Differences

The two processors come from fundamentally different design philosophies. The AMD Ryzen Threadripper PRO 5965WX is built on the Zen 3 architecture, codenamed Chagall PRO, using TSMC’s 7 nm process node with 16,600 million transistors spread across a 4x 81 mm² die configuration. The Intel Xeon 6520P uses the Granite Rapids architecture, codenamed Granite Rapids, on Intel’s 5 nm process node with a single 598 mm² die. This difference in process node and die layout contributes to different thermal and power characteristics, though both are active production parts for the server/workstation segment.

Cache hierarchies differ substantially. AMD provides 64 KB of L1 cache per core and 512 KB of L2 per core, with a large 128 MB L3 cache. Intel counters with a larger 112 KB L1 per core and 2 MB L2 per core, plus 144 MB of shared L3. The larger per-core L2 on the Intel part may explain its edge in physics simulations, where faster access to recently used data can matter more than raw throughput.

Memory support is another major divergence. The AMD chip uses DDR4 with eight-channel memory and a bandwidth of 204.8 GB/s, while the Intel chip uses DDR5 with the same eight-channel configuration but doubles the bandwidth to 409.6 GB/s. Both support ECC memory, which is expected for workstation platforms. PCIe connectivity also differs: AMD offers Gen 4 with 128 CPU-only lanes, while Intel offers Gen 5 with 88 CPU-only lanes. The newer PCIe standard on Intel provides higher per-lane bandwidth, though AMD offers more total lanes.

Clock speeds favor AMD. The Ryzen Threadripper PRO 5965WX has a base clock of 3.80 GHz and a boost clock of 4.50 GHz, versus the Xeon 6520P’s 2.40 GHz base and 4.00 GHz boost. The AMD chip also has a higher TDP of 280 watts versus Intel’s 210 watts, which correlates with its higher clock speeds. Neither processor has an unlocked multiplier, and neither includes integrated graphics, with Intel explicitly listing "N/A" for that field.

Release timing is also notable. The AMD processor launched on March 7, 2022, while the Intel processor arrived on February 23, 2025. The AMD part’s launch MSRP is $2399, while the Intel part’s is $1295. Both are locked to their respective sockets: AMD Socket WRX8 for the Ryzen part and Intel Socket 4710 for the Xeon.

Where Each One Wins

The AMD Ryzen Threadripper PRO 5965WX is the clear choice for integer-heavy workloads. Its 31.2% lead in integer math and 39.7% lead in data encryption make it the stronger option for cryptography, data processing pipelines, and general-purpose computation. The 20% advantage in data compression also points to strengths in archiving, database compression, and similar data-management tasks. For rendering, the consistent 5.4% Cinebench multi-core wins across R15, R20, and R23 suggest the AMD part is the safer pick for 3D rendering workloads, even if the margin is modest.

The Intel Xeon 6520P wins decisively in physics simulation, with its 41% advantage in PassMark physics being the single largest gap in the entire comparison. This makes it the better option for simulation-heavy scientific computing, physics engines, and similar workloads that stress that specific benchmark pattern. Intel also holds a smaller but real edge in floating-point math (3.2%), which could matter for certain numerical analysis tasks that rely heavily on floating-point operations rather than integers. The near-tie in single-thread performance (Intel wins by 0.6%) means the Intel part is essentially equal for lightly threaded tasks, with a marginal edge.

FAQ

Q: Which processor has the higher multi-core benchmark score?

A: The AMD Ryzen Threadripper PRO 5965WX wins all three Cinebench multi-core tests: R15 (5,685 vs 5,392), R20 (23,688 vs 22,467), and R23 (56,400 vs 53,495), each by a 5.4% margin. It also wins PassMark multi-thread by 5.4% (66,353 vs 62,936).

Q: Is there any benchmark where the Intel Xeon 6520P wins by a large margin?

A: Yes. The Intel Xeon 6520P wins PassMark physics by a 41% margin, scoring 7,209 versus the AMD chip’s 4,251. This is the largest delta in any benchmark between the two processors.

Q: How do the two compare in single-thread performance?

A: The Intel Xeon 6520P has a marginal 0.6% edge in PassMark single-thread (3,356 vs 3,337), but the AMD Ryzen Threadripper PRO 5965WX wins all three Cinebench single-core tests by 5.4-5.5% margins. The winner depends on the benchmark suite.

Q: What is the memory bandwidth difference?

A: The Intel Xeon 6520P supports DDR5 memory with 409.6 GB/s bandwidth, while the AMD Ryzen Threadripper PRO 5965WX supports DDR4 with 204.8 GB/s. Both use an eight-channel memory bus and support ECC memory.

Q: Which processor has more PCIe lanes?

A: The AMD Ryzen Threadripper PRO 5965WX offers 128 CPU-only lanes of PCIe Gen 4, while the Intel Xeon 6520P offers 88 CPU-only lanes of PCIe Gen 5. AMD has more lanes, but Intel’s are newer generation.

Q: How do the launch prices compare?

A: The AMD Ryzen Threadripper PRO 5965WX has a launch MSRP of $2399, while the Intel Xeon 6520P has a launch MSRP of $1295.

The Verdict

The data paints a clear picture for most workloads. The AMD Ryzen Threadripper PRO 5965WX should be the pick for anyone prioritizing rendering, data compression, encryption, or integer-heavy computation. Its 5.4% Cinebench multi-core edge and 31.2% integer math lead are consistent and meaningful across the benchmark suite. The 20% advantage in data compression and 39.7% in encryption further solidify its position for data-centric workloads. For users who need maximum PCIe lane count, the AMD part’s 128 Gen 4 lanes versus Intel’s 88 Gen 5 lanes may also be a deciding factor.

The Intel Xeon 6520P is the choice for physics-heavy simulation workloads, where its 41% PassMark physics advantage is too large to ignore. Its 3.2% lead in floating-point math also makes it viable for numerical analysis that leans on floating-point operations. The Intel part’s doubled memory bandwidth (409.6 GB/s vs 204.8 GB/s) and newer DDR5 support are significant for memory-bound tasks, though the benchmark data only shows a win in floating-point math and physics among the tested workloads. For a lower launch MSRP and superior simulation performance, the Intel chip has a clear niche.

Neither processor is a universal winner. AMD takes 12 of the 17 benchmarks, but Intel’s physics win is the most dramatic single result. The choice hinges on whether the workload matches AMD’s integer and rendering strengths or Intel’s physics and floating-point advantages. For general workstation use, the AMD processor’s broader set of wins and higher average benchmark score (98,504 vs 93,786) make it the more versatile recommendation.

Specification Differences

| Specification | AMD Ryzen Threadripper PRO 5965WX | Intel Xeon 6520P |

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

| Base Clock | 3.80 GHz | 2.40 GHz |

| Boost Clock | 4.50 GHz | 4.00 GHz |

| TDP | 280 W | 210 W |

| Socket | AMD Socket WRX8 | Intel Socket 4710 |

| Architecture | Zen 3 | Granite Rapids |

| Codename | Chagall PRO | Granite Rapids |

| Process Node | 7 nm (TSMC) | 5 nm (Intel) |

| Die Size | 4x 81 mm² | 598 mm² |

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

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

| L3 Cache | 128 MB | 144 MB (shared) |

| Memory Support | DDR4 | DDR5 |

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

| PCIe | Gen 4, 128 Lanes (CPU only) | Gen 5, 88 Lanes (CPU only) |

| Release Date | March 7, 2022 | February 23, 2025 |

| Launch MSRP | $2399 | $1295 |

| Integrated Graphics | None | N/A |

| Transistors | 16,600 million | Not specified |