AMD EPYC 9575F vs AMD Ryzen Threadripper PRO 9985WX Comparison

The AMD Ryzen Threadripper PRO 9985WX and AMD EPYC 9575F are both 64-core, 128-thread Zen 5 processors built on TSMC’s 4 nm process, each with 66,520 million transistors and 8x 70.6 mm² dies. They share the same 256 MB shared L3 cache and 128 PCIe Gen 5 lanes, yet they target different sockets — sTR5 for the Threadripper and SP5 for the EPYC — and the benchmark data reveals a consistent, if narrow, performance gap.

Head-to-Head Benchmarks

The Threadripper PRO 9985WX wins 13 of the 17 head-to-head benchmark comparisons, with its largest margins appearing in single-threaded workloads. In PassMark single-thread testing, it scores 4586 against the EPYC’s 4244, a delta of 8.1% — the widest gap in the entire comparison. This advantage carries into Cinebench single-core tests: the Threadripper leads by 4% in Cinebench R15 (1890 vs 1817), R20 (7877 vs 7573), and R23 (18756 vs 18033). The data suggests the Threadripper’s higher boost clock of 5.40 GHz, compared to the EPYC’s 5.00 GHz, translates directly into measurable single-thread superiority across every Cinebench iteration.

Multi-core results follow a similar pattern, though with tighter margins. The Threadripper leads Cinebench R23 multi-core with 132859 points versus 127739 for the EPYC, a 4% advantage. That same 4% delta repeats in Cinebench R15 multi-core (13392 vs 12876) and R20 multi-core (55800 vs 53650). PassMark multithread shows 156305 for the Threadripper against 150282 for the EPYC, also a 4% edge. The consistency of this 4% gap across all three Cinebench versions and PassMark multithread indicates a structural advantage rather than a workload-specific anomaly.

The EPYC 9575F does claim four wins, but they are concentrated in specific PassMark subtests. Its largest victory comes in PassMark physics, where it scores 22021 against the Threadripper’s 19284 — a 12.4% margin that stands as the single biggest delta in either direction. It also edges ahead in PassMark find prime numbers (1238 vs 1196, a 3.4% lead), PassMark integer math (899937 vs 888032, a 1.3% lead), and PassMark random string sorting (354092 vs 352614, a narrow 0.4% lead). Despite these wins, the EPYC’s overall average benchmark score of 313982 trails the Threadripper’s 318888 by 1.5%, according to the nearestRivals data.

Looking at the broader rival landscape, the Threadripper sits at the 100th percentile among all CPUs, while the EPYC ranks at the 99th. The Threadripper’s average score puts it 1.6% ahead of the EPYC and 2.7% ahead of the AMD EPYC 9734, but 7.5% behind the AMD Ryzen Threadripper 9980X. The EPYC, for its part, is 1.1% ahead of the EPYC 9734 and 7.7% ahead of the AMD EPYC 9555P, with a 10% lead over the AMD EPYC 9565. These figures place both processors in the same performance tier, with the Threadripper holding a slight but repeatable edge.

Where Each One Wins

The Threadripper PRO 9985WX dominates in rendering and content-creation workloads, as evidenced by its clean sweep of all six Cinebench tests. Its single-core lead of 4% across R15, R20, and R23 makes it the stronger choice for applications that rely heavily on per-thread performance, such as legacy software, lightly threaded design tools, or any workflow where clock speed matters more than core count. The 8.1% PassMark single-thread advantage reinforces this — the Threadripper’s 5.40 GHz boost clock is a genuine asset.

The Threadripper also wins in memory-sensitive compression and encryption tasks. PassMark data compression shows 2851752 for the Threadripper versus 2790810 for the EPYC, a 2.2% margin, while data encryption comes in at 169124 against 165699, a 2.1% edge. Extended instructions, which often indicate AVX-512 or similar workload performance, favor the Threadripper by 4.7% (204637 vs 195472). Floating-point math also goes to the Threadripper at 538413 versus 527967, a 2% lead.

The EPYC 9575F wins where the workload favors its architecture’s strengths. The 12.4% physics benchmark victory is the standout — PassMark physics typically reflects rigid-body simulation and collision detection, which can benefit from specific scheduling or cache behavior. The EPYC’s 3.4% lead in find prime numbers and 1.3% lead in integer math suggest it handles certain algorithmic or integer-heavy computations more efficiently, even if the overall margins are modest. Random string sorting is nearly a tie at 0.4% in the EPYC’s favor, indicating neither processor has a meaningful advantage in that specific task.

The memory subsystem differs significantly — the EPYC uses a twelve-channel DDR5 bus with 576.0 GB/s bandwidth, while the Threadripper uses eight channels at 409.6 GB/s — yet this does not translate into benchmark wins for the EPYC in bandwidth-sensitive tests like data compression. The Threadripper’s wins there suggest its higher clock speeds compensate for the narrower memory bus. The EPYC’s physics win remains its most pronounced edge, but it is isolated to that single workload.

The Verdict

For users prioritizing single-threaded performance, rendering, or encryption, the AMD Ryzen Threadripper PRO 9985WX is the clear choice from the data. It leads in every Cinebench test, every single-thread PassMark test, and in compression, encryption, extended instructions, floating-point math, and multithreaded PassMark. Its 100th percentile ranking among all CPUs and 1.6% average score advantage over the EPYC make it the higher-performing part overall.

The AMD EPYC 9575F appeals to a narrower set of workloads. Its 12.4% physics win is substantial, and its leads in integer math, prime number finding, and random string sorting — while small — show it handles certain compute patterns differently. The EPYC’s 99th percentile ranking and its position 1.1% ahead of the EPYC 9734 confirm it is a strong server part, but the head-to-head data gives the Threadripper the overall performance crown.

The choice hinges on workload. Rendering, content creation, and general workstation tasks favor the Threadripper. Physics simulation and integer-heavy server workloads may see marginal gains from the EPYC, but those gains are confined to a few PassMark subtests. The Threadripper’s 4% multi-core Cinebench advantage is more broadly applicable, and its 8.1% single-thread lead is decisive for any software that is not fully parallelized.

FAQ

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

A: The AMD Ryzen Threadripper PRO 9985WX wins all single-thread benchmarks. It scores 4586 in PassMark single-thread versus 4244 for the EPYC 9575F, an 8.1% lead, and also leads by 4% in Cinebench R15, R20, and R23 single-core tests.

Q: How do the multi-core scores compare?

A: The Threadripper leads in every multi-core test. Cinebench R23 multi-core shows 132859 for the Threadripper versus 127739 for the EPYC, a 4% margin, and PassMark multithread shows 156305 versus 150282, also a 4% lead.

Q: In which benchmark does the EPYC 9575F have its largest advantage?

A: The EPYC wins PassMark physics by 12.4%, scoring 22021 against the Threadripper’s 19284. This is the largest delta in either direction across all head-to-head benchmarks.

Q: What is the average benchmark score difference between the two?

A: The Threadripper PRO 9985WX has an average benchmark score of 318888, while the EPYC 9575F scores 313982. The nearestRivals data lists the Threadripper as 1.6% ahead of the EPYC.

Q: Do both processors have the same number of cores and threads?

A: Yes, both have 64 cores and 128 threads. They also share the same 256 MB shared L3 cache and 128 PCIe Gen 5 lanes.

Q: Which processor has a higher boost clock?

A: The Threadripper PRO 9985WX has a boost clock of 5.40 GHz, compared to 5.00 GHz for the EPYC 9575F. Its base clock is 3.20 GHz versus the EPYC’s 3.30 GHz.

Architecture Differences

Both processors are built on the Zen 5 architecture and manufactured by TSMC on a 4 nm process, with identical transistor counts of 66,520 million and die configurations of 8x 70.6 mm². The Threadripper uses the codename Shimada Peak, while the EPYC uses Turin. The L1 cache differs: the Threadripper has 64 KB per core, while the EPYC has 80 KB per core. L2 cache is identical at 1 MB per core, and both share the same 256 MB L3 cache.

The memory controllers are a major architectural split. The Threadripper uses an eight-channel DDR5 bus with 409.6 GB/s bandwidth, while the EPYC uses a twelve-channel DDR5 bus with 576.0 GB/s bandwidth. Both support ECC memory. The Threadripper is socketed for AMD Socket sTR5, and the EPYC for AMD Socket SP5. The Threadripper has an unlocked multiplier, the EPYC does not. Neither processor includes integrated graphics or 3D V-Cache.

The EPYC’s larger L1 cache and wider memory bus are notable architectural differences, yet the benchmark data shows the Threadripper winning most tests despite the EPYC’s higher memory bandwidth. This suggests the Threadripper’s higher boost clock (5.40 GHz vs 5.00 GHz) is the more decisive factor in the measured workloads.

Specification Differences

The two processors differ in several specification fields. The Threadripper PRO 9985WX has a base clock of 3.20 GHz and a boost clock of 5.40 GHz, while the EPYC 9575F has a base clock of 3.30 GHz and a boost clock of 5.00 GHz. Thermal design power differs: the Threadripper is rated at 350 W, the EPYC at 400 W. The Threadripper’s launch MSRP is $7999; the EPYC’s is $11791.

Memory bandwidth differs significantly: 409.6 GB/s for the Threadripper versus 576.0 GB/s for the EPYC, reflecting the eight-channel versus twelve-channel memory bus. L1 cache is 64 KB per core on the Threadripper versus 80 KB per core on the EPYC. The Threadripper has an unlocked multiplier, the EPYC is locked. Release dates differ, with the Threadripper listed as later than the EPYC. Part numbers also differ: 100-000000722 for the Threadripper and 100-000001554 for the EPYC. The Threadripper ranks at the 100th percentile among all CPUs, the EPYC at the 99th.