AMD Ryzen Threadripper PRO 5965WX vs AMD Ryzen Threadripper PRO 9955WX Comparison

The AMD Ryzen Threadripper PRO 9955WX and the AMD Ryzen Threadripper PRO 5965WX represent two distinct generations of AMD’s workstation-class processors, yet their overall benchmark averages land within 1.7% of each other. The 9955WX posts an average benchmark score of 100227, while the 5965WX scores 98504, a gap that hides a dramatic split in workload behavior. Across 17 head-to-head tests, the 9955WX wins 12 while the 5965WX takes 5, but those wins are not evenly distributed—they cluster around entirely different performance domains. This analysis breaks down where each processor dominates, the architectural reasons behind that split, and what the raw numbers indicate for real-world workstation use.

Where Each One Wins

The 9955WX is the clear winner in anything that stresses single-threaded execution or lightly-threaded responsiveness. Its single-core scores are consistently ahead by a wide margin: 846 vs 802 in Cinebench R15 single-core, 3527 vs 3344 in R20, and 8399 vs 7962 in R23, each a 5.5% advantage. The PassMark single-thread test shows a far larger gap, with the 9955WX scoring 4561 against the 5965WX’s 3337, a 36.7% lead. Physics simulation also favors the newer chip dramatically—7056 vs 4251, a 66% advantage. Random string sorting, another latency-sensitive task, goes to the 9955WX at 109915 vs 99287, a 10.7% edge. In every Cinebench iteration, both multi-core and single-core, the 9955WX wins by exactly 5.5%, suggesting a consistent per-clock efficiency advantage that scales across rendering workloads.

The 5965WX, by contrast, dominates in throughput-heavy, parallel integer workloads. PassMark data compression shows the older chip at 1010067 vs 890743, an 11.8% win. Data encryption is even more lopsided: 63113 vs 48476, a 23.2% advantage. Integer math goes to the 5965WX at 281247 vs 246111, a 12.5% lead, and floating-point math follows suit at 157708 vs 148042, a 6.1% edge. The most extreme difference is in prime number finding, where the 5965WX scores 523 vs 340, a 35% advantage. These are classic multi-threaded, cache-hungry tasks where the 5965WX’s 24 cores and 48 threads outmuscle the 9955WX’s 16 cores and 32 threads. The only PassMark test the 5965WX loses among these heavy parallel workloads is extended instructions (67300 vs 70809, a 5.2% win for the 9955WX), but the overall pattern is clear: the 5965WX is the brute-force parallel processor, while the 9955WX is the efficiency-focused modern design.

Architecture Differences

The two chips sit on entirely different platforms. The 9955WX uses AMD Socket sTR5, while the 5965WX uses AMD Socket WRX8. The 9955WX is built on the Zen 5 architecture with the codename Shimada Peak, whereas the 5965WX uses Zen 3 under the codename Chagall PRO. The process node tells the manufacturing story: the 9955WX is fabricated on a 4 nm process at TSMC, while the 5965WX uses a 7 nm process, also at TSMC. That node advantage translates directly into the 9955WX’s ability to hit higher clocks at a lower core count—its base clock is 4.50 GHz with a boost of 5.40 GHz, compared to the 5965WX’s 3.80 GHz base and 4.50 GHz boost. The 9955WX’s TDP is 350 W, higher than the 5965WX’s 280 W, reflecting the effort to push those clocks.

Cache configurations differ substantially. Both share 64 KB of L1 per core, but the 9955WX has 1 MB of L2 per core while the 5965WX has 512 KB per core. The L3 cache is the bigger split: the 9955WX has 64 MB shared, while the 5965WX has 128 MB. That larger L3 on the 5965WX helps explain its dominance in compression and encryption, which often benefit from larger working sets. The die layout also differs—the 9955WX uses 2x 70.6 mm² dies, while the 5965WX uses 4x 81 mm² dies. Transistor counts are nearly identical, but the 9955WX packs its transistors into fewer, smaller dies on a denser process.

Memory support is a generation apart. The 9955WX uses DDR5 with an eight-channel memory bus and 409.6 GB/s of bandwidth, while the 5965WX uses DDR4 on the same eight-channel configuration but only 204.8 GB/s. That doubling of memory bandwidth is a major architectural upgrade for the newer chip. PCIe capability also advances: the 9955WX offers Gen 5 with 128 lanes (CPU only), while the 5965WX offers Gen 4 with the same 128 lanes. Both support ECC memory, but the 5965WX is classified as Server/Workstation market segment while the 9955WX is listed as Desktop. The 9955WX has an unlocked multiplier, while the 5965WX does not. The 9955WX was released on 2025-07-22, while the 5965WX came out on 2022-03-07.

Head-to-Head Benchmarks

The biggest wins for the 9955WX come in workloads that reward high clocks and modern instruction efficiency. PassMark physics shows a 66% delta, the largest of any test, with the 9955WX scoring 7056 against 4251. Single-thread performance is nearly as impressive at 36.7% (4561 vs 3337). Random string sorting adds a 10.7% win, and extended instructions posts a 5.2% gain. Cinebench tests, both single and multi-core, all show exactly 5.5% deltas in favor of the 9955WX, from R15’s 5996 vs 5685 multi-core and 846 vs 802 single-core, through R20’s 24987 vs 23688 and 3527 vs 3344, up to R23’s 59494 vs 56400 and 8399 vs 7962. These consistent 5.5% margins suggest the 9955WX’s per-core advantage translates uniformly across rendering workloads.

The 5965WX’s wins are larger in percentage terms but confined to specific parallel tasks. Prime number finding shows a 35% delta (523 vs 340), data encryption a 23.2% delta (63113 vs 48476), and integer math a 12.5% delta (281247 vs 246111). Data compression comes in at 11.8% (1010067 vs 890743), and floating-point math at 6.1% (157708 vs 148042). The pattern is striking: the 5965WX’s 24 cores and 128 MB of L3 cache give it a decisive edge in workloads that can saturate all threads and leverage large cache pools, while the 9955WX’s higher clocks and newer architecture win everywhere else. In PassMark multithread, the 9955WX still wins at 69993 vs 66353, a 5.5% gain, despite having 8 fewer cores—evidence of its per-core efficiency. The overall benchmark average favors the 9955WX at 100227 vs 98504, a 1.7% lead, but that average masks the fact that the two chips are optimized for different task profiles.

FAQ

Q: Which processor is faster in single-threaded performance?

A: The 9955WX wins every single-core benchmark. PassMark single-thread shows 4561 vs 3337, a 36.7% advantage, and Cinebench R23 single-core shows 8399 vs 7962, a 5.5% lead.

Q: Does the 5965WX ever win in a rendering benchmark?

A: No. The 9955WX wins all Cinebench tests, including R15, R20, and R23 in both single-core and multi-core, each by exactly 5.5%.

Q: Why does the 5965WX win in data compression and encryption?

A: The 5965WX has 24 cores and 48 threads versus 16 cores and 32 threads on the 9955WX, plus 128 MB of L3 cache versus 64 MB. That combination drives wins of 11.8% in compression and 23.2% in encryption.

Q: What is the memory bandwidth difference?

A: The 9955WX supports DDR5 with 409.6 GB/s of bandwidth, while the 5965WX supports DDR4 with 204.8 GB/s. Both use an eight-channel memory bus.

Q: Which chip has a higher boost clock?

A: The 9955WX boosts to 5.40 GHz, while the 5965WX boosts to 4.50 GHz. The base clocks are 4.50 GHz and 3.80 GHz, respectively.

Q: Are both processors in the same market segment?

A: No. The 9955WX is listed as Desktop, while the 5965WX is listed as Server/Workstation. Both have ECC memory support and 128 PCIe lanes.

Specification Differences

| Specification | AMD Ryzen Threadripper PRO 9955WX | AMD Ryzen Threadripper PRO 5965WX |

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

| Cores | 16 | 24 |

| Threads | 32 | 48 |

| Base Clock | 4.50 GHz | 3.80 GHz |

| Boost Clock | 5.40 GHz | 4.50 GHz |

| TDP | 350 W | 280 W |

| Socket | AMD Socket sTR5 | AMD Socket WRX8 |

| Architecture | Zen 5 | Zen 3 |

| Codename | Shimada Peak | Chagall PRO |

| Process Node | 4 nm | 7 nm |

| Die Size | 2x 70.6 mm² | 4x 81 mm² |

| L2 Cache | 1 MB (per core) | 512 KB (per core) |

| L3 Cache | 64 MB (shared) | 128 MB |

| Memory Support | DDR5 | DDR4 |

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

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

| Market Segment | Desktop | Server/Workstation |

| Release Date | 2025-07-22 | 2022-03-07 |

| Launch MSRP | $1649 | $2399 |

| Multiplier Unlocked | Yes | No |

| Part Number | 100-000000725 | 100-000000446 |