AMD EPYC 9455P vs AMD Ryzen Threadripper PRO 9985WX Comparison

The AMD Ryzen Threadripper PRO 9985WX and the AMD EPYC 9455P are both Zen 5 processors from AMD, but the recorded benchmarks place them in different performance tiers. The Threadripper PRO 9985WX wins 16 of the 17 head-to-head tests in the database, with the EPYC 9455P taking a single field in the PassMark physics test. Across all six Cinebench workloads, the Threadripper holds a uniform 33.9% lead, and its PassMark extended instructions score is 63.9% higher. The EPYC's only win comes with a 20.4% margin in physics. The Threadripper is the clear compute winner, while the EPYC offers advantages in memory bandwidth and a specific physics workload.

Head-to-Head Benchmarks

The Threadripper PRO 9985WX dominates the benchmark record. Its largest margin is in PassMark extended instructions, where it scores 225,340 against the EPYC's 137,485, a 63.9% advantage. Floating-point math is nearly as decisive: 553,348 versus 357,783, a 54.7% lead. Data compression shows a 51% gap, 2,912,972 versus 1,928,897, and integer math follows with 872,710 versus 606,239, a 44% advantage. Random string sorting goes to the Threadripper by 35.6%, 329,014 versus 242,701, and data encryption is 34.2% ahead, 154,824 versus 115,403.

The Cinebench series is remarkable for its consistency. Every Cinebench test, from R15 to R23, both single-core and multi-core, shows the Threadripper exactly 33.9% ahead of the EPYC. Cinebench R23 multi-core records 132,859 for the Threadripper and 99,206 for the EPYC. R23 single-core is 18,756 versus 14,005. R20 multi-core is 55,800 versus 41,666, and R20 single-core is 7,877 versus 5,882. R15 multi-core is 13,392 versus 9,999, and R15 single-core is 1,890 versus 1,411. The pattern is identical in every case.

In the PassMark suite, the Threadripper also wins the multithread test, 150,071 versus 116,927, a 28.3% margin. Single-thread performance is 4,482 versus 3,745, a 19.7% lead. Find prime numbers is the closest contest, with the Threadripper ahead by only 2.8%, 1,138 versus 1,107. The EPYC's sole victory is in PassMark physics, where it scores 17,315 against the Threadripper's 13,783, a 20.4% advantage. That is the only head-to-head test the EPYC wins out of 17.

The average benchmark scores confirm the overall pattern. The Threadripper PRO 9985WX averages 320,749 across all recorded tests, while the EPYC 9455P averages 217,854. Both processors sit at the 99th percentile of all CPUs in the database, but their competitive contexts differ. The Threadripper's nearest rival, the AMD Ryzen Threadripper 9980X, is only 0.3% away in average score, and the Intel Xeon 6781P is 1.7% behind. The EPYC's nearest rival, the Intel Xeon w9-3595X, is 3.8% behind, with the Intel Xeon 6747P sitting 8.6% ahead of it. These numbers place the Threadripper in a tightly contested tier and the EPYC in a more spread-out field.

FAQ

Q: Which processor wins the majority of head-to-head benchmarks?

A: The AMD Ryzen Threadripper PRO 9985WX wins 16 of the 17 recorded head-to-head tests. The EPYC 9455P wins only the PassMark physics test.

Q: How large is the Threadripper's lead in Cinebench?

A: The lead is exactly 33.9% across all six Cinebench results: R15 multi-core and single-core, R20 multi-core and single-core, and R23 multi-core and single-core.

Q: What is the EPYC's single benchmark win?

A: The EPYC 9455P wins PassMark physics, scoring 17,315 against the Threadripper's 13,783, a 20.4% advantage.

Q: Do both processors have the same memory bandwidth?

A: No. The EPYC has twelve-channel memory with 576.0 GB/s bandwidth, while the Threadripper has eight-channel memory with 409.6 GB/s bandwidth.

Q: Are both chips built on the same manufacturing process?

A: Yes. Both use TSMC's 4nm process, with the same transistor count of 66,520 million and the same die size of 8x 70.6 mm².

Q: Which one has an unlocked multiplier?

A: The Threadripper PRO 9985WX has an unlocked multiplier; the EPYC 9455P has a locked multiplier.

The Verdict

The data is decisive: the AMD Ryzen Threadripper PRO 9985WX is the faster processor in nearly every benchmark. It wins all six Cinebench workloads, all PassMark math and compression tests, multithread, single-thread, encryption, random string sorting, and find prime numbers. Its 64 cores and 128 threads give it a 33.9% lead across Cinebench, and its 5.40 GHz boost clock supports a 19.7% single-thread advantage. For rendering, data compression, encryption, and general compute workloads, the Threadripper is the stronger choice.

The EPYC 9455P is not a general compute winner, but it has specific strengths. Its twelve-channel memory delivers 576.0 GB/s of bandwidth, and it wins the PassMark physics test by 20.4%. For workloads that rely heavily on memory bandwidth or physics calculations, the EPYC may be the better fit. The Threadripper carries a launch MSRP of $7999, while the EPYC has a launch MSRP of $4819.

Specification Differences

• Cores: Threadripper 64, EPYC 48
• Threads: 128 versus 96
• Base clock: 3.20 GHz versus 3.15 GHz
• Boost clock: 5.40 GHz versus 4.40 GHz
• TDP: 350W versus 300W
• Socket: AMD Socket sTR5 versus AMD Socket SP5
• Memory bus: Eight-channel versus Twelve-channel
• Memory bandwidth: 409.6 GB/s versus 576.0 GB/s
• L1 cache: 64 KB per core versus 80 KB per core
• L3 cache: 256 MB versus 256 MB shared
• Multiplier: Unlocked versus Locked
• Release date: 2025-07-22 versus 2024-10-09
• Part number: 100-000000722 versus 100-000001563

Architecture Differences

Both processors share the Zen 5 architecture, the 4nm TSMC process, 66,520 million transistors, and an 8x 70.6 mm² die layout. The Threadripper uses the Shimada Peak codename, while the EPYC uses Turin. The L1 cache per core differs, with the Threadripper at 64 KB and the EPYC at 80 KB. The L2 cache is 1 MB per core for both, and the L3 cache is 256 MB for both, though the EPYC specifies its L3 as shared. The Threadripper is designed for the sTR5 socket, the EPYC for SP5. Both support DDR5 memory, both provide 128 PCIe Gen 5 lanes, and neither has integrated graphics. The Threadripper has an unlocked multiplier, while the EPYC is locked. The memory controller is the notable architectural split: the Threadripper uses eight channels at 409.6 GB/s, and the EPYC uses twelve channels at 576.0 GB/s.

Where Each One Wins

The Threadripper PRO 9985WX wins in all of the following benchmark categories:

• All Cinebench tests (R15, R20, R23, both single-core and multi-core) by 33.9%
• PassMark extended instructions by 63.9%
• PassMark floating-point math by 54.7%
• PassMark data compression by 51%
• PassMark integer math by 44%
• PassMark random string sorting by 35.6%
• PassMark data encryption by 34.2%
• PassMark multithread by 28.3%
• PassMark single-thread by 19.7%
• PassMark find prime numbers by 2.8%

The EPYC 9455P wins in:

• PassMark physics by 20.4%
• Memory bandwidth, with twelve-channel 576.0 GB/s versus the Threadripper's eight-channel 409.6 GB/s

For multi-core compute, rendering, and compression, the Threadripper PRO 9985WX is the clear winner. For physics-heavy workloads or memory-bandwidth-bound applications, the EPYC's 20.4% physics advantage and twelve-channel memory make it the better choice. The recorded data shows the Threadripper in a tightly contested performance tier with its nearest rival 0.3% away, while the EPYC's nearest rival sits 3.8% behind, confirming the Threadripper's position as the higher-performing of the two processors.