AMD Ryzen Threadripper PRO 3995WX vs Intel Xeon 6740P Comparison

The Intel Xeon 6740P and AMD Ryzen Threadripper PRO 3995WX represent two distinct approaches to high-core-count computing: one is a 2025-era server/workstation chip built on Intel’s Granite Rapids architecture, the other a 2020-era desktop/workstation behemoth based on AMD’s Zen 2 design. Benchmark data shows a clear split in strengths, with the Xeon dominating compute-heavy and single-threaded workloads while the Threadripper takes the lead in memory-intensive and encryption tasks. The Xeon wins 12 of 16 direct head-to-head comparisons, yet the average benchmark score across all tests tells a closer story: the Xeon posts 176,227 against the Threadripper’s 171,748, a 2.6% gap that flatters the newer chip but does not erase the AMD part’s specialized advantages.

Head-to-Head Benchmarks

The most decisive victory for the Intel Xeon 6740P comes in PassMark’s physics test, where it scores 9705 against the Threadripper’s 5498 — a 76.5% lead. This is the single largest delta in the entire comparison, and it suggests the Xeon’s architecture handles the simulation-style workload with far greater efficiency per thread. Similarly, in PassMark’s find-prime-numbers test, the Xeon scores 822 versus 566, a 45.2% advantage. These are not marginal wins; they indicate a fundamental difference in how each chip executes integer-heavy, latency-sensitive loops.

Single-thread performance also favors the Xeon decisively. In PassMark single-thread, the Xeon scores 2975 against 2595, a 14.6% lead. Cinebench R15 single-core shows the same pattern: 1065 for Intel versus 1014 for AMD, a 5% edge. The Xeon’s advantage persists across every Cinebench iteration — R15, R20, and R23 — with multi-core deltas consistently at 4.9% and single-core deltas at 5% or 4.9%. In Cinebench R23 multi-core, the Xeon hits 74851 versus 71338, a 4.9% margin that translates to roughly 3,500 points. The Xeon also wins PassMark floating-point math (296102 vs 277715, +6.6%) and extended instructions (116992 vs 106423, +9.9%), showing strength in both scalar and vector workloads.

The AMD Ryzen Threadripper PRO 3995WX, however, owns the memory-bound and data-movement categories. Its biggest win is in PassMark data encryption, where it scores 124433 versus 82028 — a 34.1% lead for AMD. This is a massive gap, and it points to the Threadripper’s superior per-core encryption throughput. In data compression, the AMD chip scores 1841292 against 1537129, a 16.5% advantage. Integer math also goes to AMD: 490116 versus 388500, a 20.7% lead. Random string sorting is another AMD win, 188574 versus 175015, a 7.2% margin. These four wins are substantial, but they are isolated to specific workload types rather than general-purpose compute.

Where Each One Wins

The Intel Xeon 6740P is the choice for compute-heavy, latency-sensitive applications. Its 76.5% physics win and 45.2% prime-number win make it the stronger candidate for scientific simulation, financial modeling, or any workload that stresses branch prediction and integer throughput. The 14.6% single-thread lead in PassMark also gives it an edge in lightly threaded tasks like compilation or scripting, where per-core speed matters more than raw core count. The consistent 4.9% Cinebench multi-core margin across three versions suggests that in rendering or video-encoding workloads, the Xeon will deliver roughly 5% more performance than the Threadripper — a modest but consistent advantage.

The AMD Ryzen Threadripper PRO 3995WX is the better fit for data-heavy workflows. Its 34.1% encryption win and 16.5% compression win make it superior for database encryption, file archiving, or any task that moves large volumes of data through the CPU. The 20.7% integer-math advantage reinforces this: integer-heavy workloads like hash computation or data validation will run noticeably faster on AMD. The 7.2% random-string-sorting win is another data-processing signal. If the workload involves moving, sorting, or encrypting data rather than computing mathematical results, the Threadripper’s 64 cores and higher memory bandwidth per core give it the edge.

Architecture Differences

The two processors are built on fundamentally different foundations. The Intel Xeon 6740P uses Granite Rapids architecture on Intel’s 5 nm process, with a die size of 2x 598 mm². It packs 48 cores and 96 threads, with a base clock of 2.10 GHz and a boost clock of 3.80 GHz. The Threadripper PRO 3995WX uses Zen 2 architecture on TSMC’s 7 nm process, with a die size of 8x 74 mm² and 30,400 million transistors. It offers 64 cores and 128 threads, running at a 2.70 GHz base and 4.20 GHz boost. Despite having 16 fewer cores, the Xeon manages to stay ahead in most multi-threaded tests, suggesting its per-core IPC and clock efficiency compensate for the core deficit.

Cache configurations differ significantly. The Xeon provides 112 KB of L1 per core, 2 MB of L2 per core, and a massive 288 MB of shared L3 cache. The Threadripper has 64 KB of L1 per core, 512 KB of L2 per core, and 256 MB of L3. The Xeon’s larger per-core L2 and overall L3 likely contribute to its single-thread and physics-test dominance. Memory support also diverges: the Xeon uses DDR5 with eight-channel memory and a bandwidth of 409.6 GB/s, while the Threadripper uses DDR4 with eight-channel memory and 204.8 GB/s bandwidth — exactly half the Xeon’s theoretical peak. The Xeon’s PCIe implementation is Gen 5 with 88 lanes, whereas the Threadripper offers Gen 4 with 128 lanes. The Xeon’s higher memory bandwidth is notable, but the Threadripper’s encryption and compression wins suggest that raw bandwidth alone does not determine data-movement performance.

The Xeon’s TDP is 270 W versus the Threadripper’s 280 W, a negligible difference given the core-count disparity. The Xeon is a server/workstation part with no integrated graphics, while the Threadripper is classified as desktop and also lacks integrated graphics. Both support ECC memory. The Xeon was released on 2025-02-23, while the Threadripper launched on 2020-07-13 — a gap of over four years that explains the process-node and memory-technology differences.

FAQ

Q: Which processor has more cores and threads?

A: The AMD Ryzen Threadripper PRO 3995WX has 64 cores and 128 threads, while the Intel Xeon 6740P has 48 cores and 96 threads — a 16-core, 32-thread advantage for AMD.

Q: Why does the Intel Xeon win most benchmarks despite having fewer cores?

A: Benchmark results show the Xeon wins 12 of 16 head-to-head tests, including all Cinebench versions and PassMark multithread, physics, and single-thread. Its advantage appears tied to its 288 MB L3 cache, larger per-core L2, and higher memory bandwidth (409.6 GB/s vs 204.8 GB/s), which boost per-core efficiency and latency-sensitive workloads.

Q: In which workloads does the AMD Threadripper clearly outperform the Xeon?

A: The Threadripper wins in four specific PassMark tests: data encryption (124433 vs 82028, +34.1%), data compression (1841292 vs 1537129, +16.5%), integer math (490116 vs 388500, +20.7%), and random string sorting (188574 vs 175015, +7.2%). These indicate a strong advantage in data-movement and encryption tasks.

Q: What is the memory bandwidth difference between the two?

A: The Intel Xeon 6740P supports DDR5 with eight-channel memory and a bandwidth of 409.6 GB/s. The AMD Ryzen Threadripper PRO 3995WX supports DDR4 with eight-channel memory and a bandwidth of 204.8 GB/s — exactly half the Xeon’s figure.

Q: How do their average benchmark scores compare?

A: The Xeon’s average benchmark score is 176,227, while the Threadripper’s is 171,748. This gives the Xeon a 2.6% higher average score. The Xeon’s nearest rival, the AMD EPYC 7763, scores 179,916, which is 2.1% higher than the Xeon.

Q: What are the launch MSRPs?

A: The Intel Xeon 6740P has a launch MSRP of $4650. The AMD Ryzen Threadripper PRO 3995WX has a launch MSRP of $5489.

Specification Differences

| Specification | Intel Xeon 6740P | AMD Ryzen Threadripper PRO 3995WX |

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

| Cores | 48 | 64 |

| Threads | 96 | 128 |

| Base Clock | 2.10 GHz | 2.70 GHz |

| Boost Clock | 3.80 GHz | 4.20 GHz |

| TDP | 270 W | 280 W |

| Socket | Intel Socket 4710 | AMD Socket WRX8 |

| Architecture | Granite Rapids | Zen 2 |

| Codename | Granite Rapids | Castle Peak |

| Process Node | 5 nm | 7 nm |

| Foundry | Intel | TSMC |

| Die Size | 2x 598 mm² | 8x 74 mm² |

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

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

| L3 Cache | 288 MB (shared) | 256 MB |

| Memory Support | DDR5 | DDR4 |

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

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

| Market Segment | Server/Workstation | Desktop |

| Release Date | 2025-02-23 | 2020-07-13 |

| Part Number | SRV5R | 100-000000087100-100000087WOF |