AMD EPYC 4585PX vs AMD Ryzen Threadripper PRO 5965WX Comparison

Head-to-Head Benchmarks

The benchmark data tells a clear story of two very different design philosophies. The AMD EPYC 4585PX wins 11 of the 17 head-to-head benchmarks, while the AMD Ryzen Threadripper PRO 5965WX takes 6. More telling than the raw win count is the magnitude of the victories in each direction. The EPYC 4585PX dominates in single-threaded and lightly-threaded workloads with margins that reach 43.7%, while the Threadripper PRO 5965WX counters in memory-bandwidth-sensitive and multi-threaded integer tasks, with its largest edge at 21.8%.

Starting with the Cinebench suite, the EPYC 4585PX sweeps all six tests. In Cinebench R15, R20, and R23, the EPYC leads by a consistent 7.2% in both single-core and multi-core runs. The R23 multi-core score of 60451 versus 56400 puts the EPYC ahead by 4051 points, a meaningful gap in a heavily threaded render workload. The single-core R23 delta of 8534 versus 7962 is equally notable, showing that the EPYC’s architectural advantage extends to per-thread performance. This consistency across all three Cinebench versions suggests a fundamental performance-per-clock advantage, not a workload-specific quirk.

The Passmark suite reveals a more nuanced picture. The EPYC 4585PX wins Passmark single-thread with a score of 4795 versus 3337, a staggering 43.7% advantage. That is the largest margin in the entire head-to-head set. It also wins Passmark multithread (71552 versus 66353, a 7.8% edge), Passmark physics (7140 versus 4251, a 68% blowout), and Passmark extended instructions (67572 versus 67300, a razor-thin 0.4% margin). The physics result is particularly striking — the EPYC nearly doubles the Threadripper’s score, indicating a major advantage in simulated physics or rigid-body calculations.

The Threadripper PRO 5965WX, however, flexes its 24-core muscle in several Passmark workloads. It wins data compression by 11.5% (1010067 versus 893899), data encryption by 21.8% (63113 versus 49374), and integer math by 9% (281247 versus 255898). It also edges out the EPYC in floating-point math (157708 versus 154641, a 1.9% margin), find prime numbers (523 versus 514, a 1.7% margin), and random string sorting (99287 versus 98839, a 0.5% margin). These wins are not trivial — data encryption at a 21.8% gap is a substantial lead — but they are concentrated in specific algorithm types rather than broad workloads.

The average benchmark score positions the EPYC 4585PX at 100819, which is 2.3% higher than the Threadripper PRO 5965WX’s 98504. Both CPUs sit in the 98th percentile of all CPUs tracked, so this is a battle between two top-tier parts. In the nearestRivals data, the EPYC 4585PX sits between the Ryzen Threadripper PRO 9955WX (0.6% higher) and the EPYC 7513 (1.4% lower), while the Threadripper PRO 5965WX trails the 9955WX by 1.7% and leads the EPYC 4564P by 2.3%. The head-to-head delta of 2.3% favoring the EPYC is consistent with the overall average score gap.

FAQ

Q: Which processor has the higher single-threaded performance?

A: The AMD EPYC 4585PX is decisively ahead. It wins Passmark single-thread by 43.7% (4795 versus 3337) and Cinebench R23 single-core by 7.2% (8534 versus 7962). This is the EPYC’s largest categorical advantage.

Q: Where does the AMD Ryzen Threadripper PRO 5965WX outperform the EPYC 4585PX?

A: The Threadripper wins in data compression (1010067 versus 893899, an 11.5% margin), data encryption (63113 versus 49374, a 21.8% margin), and integer math (281247 versus 255898, a 9% margin). It also narrowly wins floating-point math, find prime numbers, and random string sorting, though those margins are under 2%.

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

A: The EPYC 4585PX leads in all Cinebench multi-core tests by 7.2%. The R23 multi-core scores are 60451 for the EPYC versus 56400 for the Threadripper. In Passmark multithread, the EPYC also wins by 7.8% (71552 versus 66353). The Threadripper’s extra 8 cores do not translate to a win in these rendering benchmarks.

Q: What is the most lopsided benchmark result in the head-to-head comparison?

A: The Passmark physics test. The EPYC 4585PX scores 7140 versus the Threadripper PRO 5965WX’s 4251, a 68% advantage. This is by far the largest delta in either direction across all 17 benchmarks.

Q: Are these processors in the same performance class?

A: Yes. Both are in the 98th percentile of all CPUs. The EPYC 4585PX has an average benchmark score of 100819, which is 2.3% higher than the Threadripper’s 98504. The nearestRivals data places them as direct competitors, with the EPYC 4585PX sitting 2.3% above the Threadripper PRO 5965WX.

Q: How close are the extended instruction benchmarks?

A: Extremely close. The EPYC 4585PX wins Passmark extended instructions with 67572 versus 67300, a margin of just 0.4%. This is the narrowest head-to-head result in the entire dataset.

Architecture Differences

The architectural divide between these two CPUs is generational and substantial. The EPYC 4585PX is built on Zen 5 architecture, codenamed Grado, while the Threadripper PRO 5965WX uses Zen 3 architecture, codenamed Chagall PRO. This is a two-generation gap in core design, which explains the EPYC’s commanding single-thread lead.

The process node tells part of the story. The EPYC 4585PX is fabricated on a 4 nm process at TSMC, while the Threadripper PRO 5965WX uses a 7 nm process, also at TSMC. The smaller node allows for denser transistors and lower power per operation. Interestingly, the transistor counts are nearly identical — 16,630 million for the EPYC versus 16,600 million for the Threadripper — but the die layout differs. The EPYC uses two dies at 70.6 mm² each, while the Threadripper uses four dies at 81 mm² each. The smaller, newer dies on the EPYC pack similar transistor counts into less silicon area.

Cache hierarchies also diverge. The EPYC 4585PX has 80 KB of L1 cache per core and 1 MB of L2 cache per core, with 128 MB of shared L3 cache. The Threadripper PRO 5965WX has 64 KB of L1 per core and 512 KB of L2 per core, also with 128 MB of shared L3. The EPYC’s larger per-core L1 and L2 caches contribute to its single-thread advantage, while the equal L3 capacity helps both in shared-data workloads.

Memory architecture is a major differentiator. The EPYC 4585PX supports DDR5 memory on a dual-channel bus, delivering 89.6 GB/s of bandwidth. The Threadripper PRO 5965WX uses DDR4 on an eight-channel bus, offering 204.8 GB/s. The Threadripper’s 2.3x bandwidth advantage explains its wins in memory-sensitive tasks like data compression and encryption. However, the EPYC’s DDR5 support brings newer memory technology, even if the channel count is lower.

PCIe capabilities also differ significantly. The EPYC 4585PX provides PCIe Gen 5 with 24 lanes (CPU only), while the Threadripper PRO 5965WX provides PCIe Gen 4 with 128 lanes (CPU only). The Threadripper’s massive lane count is clearly aimed at workstation expansion, while the EPYC’s Gen 5 support offers higher per-lane bandwidth.

The EPYC 4585PX includes integrated Radeon Graphics, while the Threadripper PRO 5965WX has no integrated graphics. This is a feature advantage for the EPYC in systems that do not require a discrete GPU. Both CPUs support ECC memory, and neither has an unlocked multiplier.

Specification Differences

The two processors differ across nearly every core specification. The EPYC 4585PX has 16 cores and 32 threads, while the Threadripper PRO 5965WX has 24 cores and 48 threads — a 50% core advantage for the Threadripper. Yet the EPYC’s clock speeds are higher: base clock of 4.30 GHz versus 3.80 GHz, and boost clock of 5.70 GHz versus 4.50 GHz. The EPYC’s boost clock is 1.2 GHz higher, a substantial margin.

Thermal design power differs sharply. The EPYC 4585PX is rated at 170 W TDP, while the Threadripper PRO 5965WX is rated at 280 W TDP. The Threadripper draws 64.7% more power at the TDP rating, despite being on an older, less efficient process node.

Socket and platform are completely different. The EPYC 4585PX uses AMD Socket AM5, while the Threadripper PRO 5965WX uses AMD Socket WRX8. This means they are not interchangeable in any system. The EPYC’s part number is 100-000001561, and the Threadripper’s is 100-000000446.

Release dates are separated by over three years. The EPYC 4585PX launched on 2025-05-12, while the Threadripper PRO 5965WX launched on 2022-03-07. The EPYC’s newer release date aligns with its Zen 5 architecture and 4 nm process. Both are listed as Active in production status.

Memory support differs in both type and bus width. The EPYC uses DDR5 on a dual-channel bus with 89.6 GB/s bandwidth. The Threadripper uses DDR4 on an eight-channel bus with 204.8 GB/s bandwidth. The EPYC’s PCIe implementation is Gen 5 with 24 lanes, while the Threadripper’s is Gen 4 with 128 lanes.

The EPYC 4585PX includes integrated Radeon Graphics; the Threadripper PRO 5965WX has none. The EPYC’s market segment is listed as Server/Workstation, as is the Threadripper’s. Both are server-class parts, but the EPYC’s AM5 socket and dual-channel memory suggest a different system design point than the WRX8 platform.

Where Each One Wins

The AMD EPYC 4585PX is the clear choice for workloads that depend on single-thread performance, per-core efficiency, and modern architecture. Its 43.7% lead in Passmark single-thread and 7.2% lead across all Cinebench tests make it the better pick for lightly threaded applications, real-time physics simulation, and any software that does not scale perfectly across many cores. The 68% advantage in Passmark physics is a strong indicator for engineering simulation or game-physics workloads. The EPYC also wins Passmark multithread by 7.8% despite having 8 fewer cores, proving that its architectural efficiency can overcome a core-count deficit in many threaded tasks. Systems built around this CPU benefit from DDR5 memory, PCIe Gen 5, integrated graphics, and a lower 170 W TDP, making it a more power-conscious option for dense server environments.

The AMD Ryzen Threadripper PRO 5965WX is the better choice for workloads that are heavily memory-bandwidth-bound or that exhibit high parallelism with simple per-thread operations. Its 204.8 GB/s eight-channel DDR4 bandwidth is 2.3x that of the EPYC, and this shows in the benchmark data. The Threadripper wins data compression by 11.5% and data encryption by 21.8%, both of which are classic memory-throughput-sensitive tasks. It also leads integer math by 9%, a workload that benefits from having 24 cores and 48 threads available. The 128 PCIe Gen 4 lanes make it the superior platform for systems requiring massive I/O expansion, such as GPU compute clusters or multi-NVMe storage arrays. Its 280 W TDP is higher, but for a workstation that needs maximum core count and expansion capability, the Threadripper’s design point is justified.

The overall average benchmark score favors the EPYC 4585PX by 2.3%, and it wins 11 of 17 head-to-head tests. However, the Threadripper’s wins are in specific, identifiable workload categories that matter for certain professional users. For a general-purpose server CPU with modern features and strong single-thread performance, the EPYC 4585PX is the better all-rounder. For a workstation CPU that prioritizes raw core count, memory bandwidth, and I/O lanes over per-thread speed, the Threadripper PRO 5965WX retains a clear niche. The