AMD EPYC 9475F vs AMD Ryzen Threadripper PRO 9995WX Comparison

Head-to-Head Benchmarks

The head-to-head data presents a decisive picture: the AMD Ryzen Threadripper PRO 9995WX wins 10 of 11 benchmark comparisons against the AMD EPYC 9475F, with the EPYC taking only a single, narrow victory. The most striking margin comes in integer math, where the Threadripper scores 1,203,634 against 605,696, a 98.7% advantage — effectively double the throughput. Floating-point math follows a similar pattern, with the Threadripper at 725,066 versus 406,524, a 78.4% lead. These two results alone signal that the 96-core Threadripper is operating in a different performance tier for compute-heavy workloads.

Data compression and encryption also show substantial gaps. The Threadripper scores 3,723,652 in compression against 2,156,305 for the EPYC, a 72.7% delta. Encryption shows a 77.2% advantage, with scores of 206,662 and 116,648 respectively. The extended instructions benchmark, which exercises SIMD and specialized instruction paths, favors the Threadripper by 56.3% (270,647 vs 173,169). Random string sorting, a test sensitive to memory subsystem and cache behavior, shows a 65% delta (418,973 vs 253,936).

The multithreaded PassMark score — a broader aggregate — puts the Threadripper at 171,200 versus 122,476, a 39.8% advantage. This is smaller than the math-focused deltas, suggesting that the EPYC's higher base clock (3.65 GHz vs 2.50 GHz) partially compensates for its lower core count in some mixed workloads. Single-thread performance tells a similar story: the Threadripper leads 4,542 to 3,779, a 20.2% margin, driven by its 5.40 GHz boost clock versus 4.80 GHz on the EPYC.

The EPYC's sole win is in prime number finding, scoring 1,507 against 1,476 — a 2.1% edge. This is a trivial margin, likely reflecting per-core efficiency in a workload that does not scale linearly with core count. Physics simulation is nearly a tie: 16,860 for the Threadripper versus 16,443, a 2.5% delta. These two results indicate that in latency-bound, low-thread scenarios, the EPYC's higher base clock keeps it competitive, though it cannot overcome the Threadripper's core-count advantage elsewhere.

The Verdict

The data is unambiguous: the AMD Ryzen Threadripper PRO 9995WX is the superior processor for nearly every measured workload. Its 96 cores and 192 threads, combined with a 5.40 GHz boost clock, deliver dominant results across math, compression, encryption, and multithreaded aggregates. The 10-to-1 win tally, with deltas ranging from 2.5% to 98.7%, leaves little room for interpretation. The EPYC 9475F, with 48 cores and 96 threads, is not a peer in raw throughput; it is a different class of part aimed at different priorities.

Who should choose the EPYC 9475F? Strictly from the data, only a user whose workload is dominated by prime-number finding or similar latency-sensitive, low-thread tasks would prefer it — and even then, the 2.1% delta is within noise. The EPYC's higher base clock (3.65 GHz vs 2.50 GHz) does not translate into a meaningful benchmark advantage, and its single-thread score trails by 20.2%. The EPYC's twelve-channel memory bus (576.0 GB/s) versus the Threadripper's eight-channel (409.6 GB/s) offers more bandwidth, but no benchmark in this dataset reflects that advantage.

The verdict for most buyers is the Threadripper. It wins single-thread, multithread, math, compression, and encryption. The only scenario where the EPYC makes sense from this data is a deployment where the 48-core part's lower core count is a requirement — perhaps due to software licensing or power constraints — but the 400 W TDP versus 350 W TDP does not suggest a power advantage. The Threadripper is simply faster in nearly every measurable way.

Architecture Differences

Both processors are built on TSMC's 4 nm process node and use the Zen 5 architecture, but they diverge significantly in implementation. The Threadripper PRO 9995WX, codenamed Shimada Peak, is part of the 9000 series and uses the AMD Socket sTR5. It packs 96 cores and 192 threads, with a base clock of 2.50 GHz and a boost clock of 5.40 GHz. The EPYC 9475F, codenamed Turin, belongs to the EPYC 9005 series and uses AMD Socket SP5. It has 48 cores and 96 threads, with a higher base clock of 3.65 GHz but a lower boost clock of 4.80 GHz.

Cache hierarchies differ substantially. The Threadripper has 64 KB of L1 per core, 1 MB of L2 per core, and 384 MB of L3 cache. The EPYC has 80 KB of L1 per core, 1 MB of L2 per core, and 256 MB of shared L3. The Threadripper's larger L3 is a direct consequence of its higher core count and contributes to its 98.7% integer math lead. Transistor counts reflect the same disparity: 99,780 million for the Threadripper versus 66,520 million for the EPYC. Die sizes are 12x 70.6 mm² versus 8x 70.6 mm², confirming that the Threadripper uses more identical chiplets.

Memory support differs in channel count. The Threadripper uses eight-channel DDR5 with 409.6 GB/s bandwidth; the EPYC uses twelve-channel DDR5 with 576.0 GB/s bandwidth. Both support ECC memory and offer Gen 5 PCIe with 128 CPU-only lanes. Neither has integrated graphics. The Threadripper has an unlocked multiplier, enabling overclocking, while the EPYC is locked. The Threadripper's launch MSRP is $11700; the EPYC's is $7592. The Threadripper also has a more recent release date (2025-07-22) versus the EPYC (2024-10-09). Both are active production parts in the Server/Workstation segment.

FAQ

Q: Which processor has more cores and threads?

A: The AMD Ryzen Threadripper PRO 9995WX has 96 cores and 192 threads, while the AMD EPYC 9475F has 48 cores and 96 threads.

Q: How much faster is the Threadripper in integer math?

A: The Threadripper scores 1,203,634 in PassMark integer math versus 605,696 for the EPYC, a 98.7% advantage.

Q: Does the EPYC win any benchmark?

A: Yes, the EPYC wins the PassMark find prime numbers test with a score of 1,507 versus 1,476, a 2.1% delta.

Q: What are the base clock speeds of each processor?

A: The Threadripper has a base clock of 2.50 GHz, while the EPYC has a base clock of 3.65 GHz.

Q: Which processor has higher memory bandwidth?

A: The EPYC has higher memory bandwidth at 576.0 GB/s with a twelve-channel bus, versus 409.6 GB/s with an eight-channel bus for the Threadripper.

Q: Are both processors based on the same architecture?

A: Yes, both use the Zen 5 architecture on TSMC's 4 nm process node, but they have different codenames: Shimada Peak for the Threadripper and Turin for the EPYC.

Where Each One Wins

The Threadripper PRO 9995WX wins in every high-thread, compute-intensive category. Its 98.7% integer math lead and 78.4% floating-point advantage make it the clear choice for scientific computing, financial modeling, and any workload that saturates all cores. Data compression (72.7% ahead) and encryption (77.2% ahead) favor it for database workloads, backup systems, and secure communications processing. Extended instructions (56.3% ahead) and random string sorting (65% ahead) indicate strong performance in parsing, encoding, and data transformation tasks. The multithreaded aggregate (39.8% ahead) confirms broad superiority across mixed workloads. Even single-thread performance, where the EPYC's higher base clock might help, goes to the Threadripper by 20.2%, thanks to its 5.40 GHz boost clock.

The EPYC 9475F wins exactly one benchmark: find prime numbers, by 2.1%. This is a narrow, likely noise-level margin in a workload that is latency-bound and does not scale with core count. In physics simulation, the EPYC is nearly tied (16,443 vs 16,860, a 2.5% delta), which suggests that in lightly threaded, physics-based simulations, the two are indistinguishable. For any user primarily running prime-number sieving or similar single-thread integer loops, the EPYC offers a marginal edge. However, the delta is so small that it would not justify choosing the EPYC over the Threadripper for any broader workload mix.

The EPYC's twelve-channel memory bus and higher 576.0 GB/s bandwidth could theoretically benefit memory-bandwidth-bound tasks, but no benchmark in this dataset shows such an effect. The Threadripper's 384 MB L3 cache versus 256 MB on the EPYC likely explains its dominance in data-heavy tests like compression and sorting. In summary, the Threadripper is the winner for rendering, encoding, compilation, data analytics, and any parallel compute task. The EPYC is a niche choice for specific latency-sensitive workloads, and even then, the margin is negligible.