AMD EPYC 9575F vs AMD Ryzen Threadripper PRO 9995WX Comparison

FAQ

Q: What is the core count difference between the AMD Ryzen Threadripper PRO 9995WX and the AMD EPYC 9575F?

A: The Threadripper PRO 9995WX has 96 cores and 192 threads, while the EPYC 9575F has 64 cores and 128 threads.

Q: Which processor has the higher boost clock?

A: The Threadripper PRO 9995WX boosts to 5.40 GHz, which is higher than the EPYC 9575F's 5.00 GHz boost. However, the EPYC has a higher base clock at 3.30 GHz versus 2.50 GHz.

Q: How do the two compare in overall benchmark averages?

A: The Threadripper PRO 9995WX has an average benchmark score of 412068, putting it in the 100th percentile of all CPUs. The EPYC 9575F averages 311774, which places it in the 99th percentile.

Q: Does the EPYC 9575F win any benchmark against the Threadripper PRO 9995WX?

A: Yes, in PassMark physics testing, the EPYC 9575F scores 20652 versus 16860, a lead of 18.4 percent.

Q: Are both processors based on the same architecture?

A: Yes, both use Zen 5 architecture, but they come from different product lines and codenames. The Threadripper uses the Shimada Peak codename from the 9000 series, while the EPYC is from the 9005 series with the Turin codename.

Q: What is the memory channel configuration difference?

A: The EPYC 9575F has a twelve-channel memory bus with 576.0 GB/s bandwidth, while the Threadripper PRO 9995WX has an eight-channel bus with 409.6 GB/s bandwidth.

Architecture Differences

Both processors are built on Zen 5 architecture at TSMC's 4 nm process node, but they diverge significantly in physical design and platform integration. The Threadripper PRO 9995WX, codenamed Shimada Peak, uses 12 compute dies, each 70.6 mm², totaling 99,780 million transistors. The EPYC 9575F, codenamed Turin, uses 8 dies of the same individual size, with 66,520 million transistors total. This die count difference directly explains the core count gap: 96 cores on the Threadripper versus 64 on the EPYC.

Cache hierarchies also differ. The Threadripper features 64 KB of L1 cache per core and 1 MB of L2 per core, with 384 MB of L3 cache. The EPYC has a larger L1 allocation at 80 KB per core, the same 1 MB L2 per core, but substantially less L3 at 256 MB shared. For workloads that depend on large pooled caches, the Threadripper's 384 MB L3 gives it a structural advantage.

The platforms themselves are incompatible. The Threadripper PRO 9995WX mounts on AMD Socket sTR5, while the EPYC 9575F uses AMD Socket SP5. Both provide 128 PCIe Gen 5 lanes from the CPU, so expansion capacity is equal on paper. Neither has integrated graphics. Both support DDR5 memory with ECC, but the EPYC's twelve-channel memory controller delivers 576.0 GB/s versus the Threadripper's eight-channel 409.6 GB/s, a 40.6 percent bandwidth advantage for the server part.

Power envelopes differ as well. The EPYC 9575F has a TDP of 400 watts, while the Threadripper PRO 9995WX is rated at 350 watts. The Threadripper also has an unlocked multiplier, making it a workstation-class overclocking target, whereas the EPYC is multiplier-locked for datacenter stability. Release timing separates them by roughly nine months: the EPYC launched on 2024-10-09, and the Threadripper followed on 2025-07-22.

Where Each One Wins

The benchmark data splits cleanly. The Threadripper PRO 9995WX wins 16 of 17 head-to-head comparisons, making it the default choice for heavily threaded compute. Cinebench results are uniform: across R15, R20, and R23, both multi-core and single-core tests show the Threadripper ahead by exactly 16.3 percent. That consistency across render workloads indicates a broad architectural advantage rather than a test-specific quirk.

PassMark results reinforce the pattern. The Threadripper leads by 34.3 percent in data compression, 26.9 percent in data encryption, 37 percent in extended instructions, 21.5 percent in prime number finding, 37.8 percent in floating point math, 35 percent in integer math, 15.7 percent in multithreaded performance, and 20.2 percent in random string sorting. These are large, double-digit margins across the board, covering everything from cryptography to sorting algorithms.

Single-thread performance also favors the Threadripper, though by a smaller margin: 8.8 percent in PassMark single-thread tests. That advantage aligns with its higher 5.40 GHz boost clock versus 5.00 GHz on the EPYC.

The EPYC 9575F has exactly one win: PassMark physics, where it scores 20652 versus 16860, an 18.4 percent margin. This result is notable because the EPYC has fewer cores but a higher base clock (3.30 GHz versus 2.50 GHz) and more memory bandwidth. Physics simulations that are latency-sensitive or bandwidth-bound can benefit from that combination. The EPYC also holds the memory bandwidth crown at 576.0 GB/s, which matters for workloads that stream large datasets, even if that advantage does not translate into a broader benchmark sweep.

Specification Differences

| Specification | AMD Ryzen Threadripper PRO 9995WX | AMD EPYC 9575F |

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

| Series | 9000 series | EPYC 9005 series |

| Cores | 96 | 64 |

| Threads | 192 | 128 |

| Base Clock | 2.50 GHz | 3.30 GHz |

| Boost Clock | 5.40 GHz | 5.00 GHz |

| TDP | 350 W | 400 W |

| Socket | AMD Socket sTR5 | AMD Socket SP5 |

| Codename | Shimada Peak | Turin |

| Transistors | 99,780 million | 66,520 million |

| Die Count | 12x 70.6 mm² | 8x 70.6 mm² |

| L1 Cache | 64 KB (per core) | 80 KB (per core) |

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

| L3 Cache | 384 MB | 256 MB (shared) |

| Memory Bus | Eight-channel | Twelve-channel |

| Memory Bandwidth | 409.6 GB/s | 576.0 GB/s |

| Multiplier | Unlocked | Locked |

| Release Date | 2025-07-22 | 2024-10-09 |

| Launch MSRP | $11700 | $11791 |

The table highlights the trade-off. The Threadripper offers more cores, more cache, and a higher boost clock. The EPYC counters with a higher base clock, more memory channels, higher bandwidth, and a 50-watt higher TDP. Both use the same 4 nm TSMC process and share Zen 5 cores, so IPC is identical; the differences come from configuration and scale.

Head-to-Head Benchmarks

The Cinebench suite is a clean sweep for the Threadripper PRO 9995WX. In R15 multi-core, it scores 14978 versus 12876, a 16.3 percent win. Single-core R15 follows the same pattern: 2114 versus 1817, also 16.3 percent. The R20 results are 62412 versus 53650 for multi-core and 8811 versus 7573 for single-core, both again at 16.3 percent. R23 closes the Cinebench set with 148601 versus 127739 multi-core and 20979 versus 18033 single-core, maintaining the identical 16.3 percent margin. The uniformity suggests the Threadripper's advantage scales linearly with core count in these render workloads, with no single-thread penalty.

PassMark data compression shows a bigger gap: 3723652 versus 2773634, a 34.3 percent lead for the Threadripper. Data encryption is 206662 versus 162818, a 26.9 percent margin. Extended instructions jump to 37 percent ahead at 270647 versus 197484. Prime number finding is 1476 versus 1215, a 21.5 percent lead. Floating point math delivers the largest margin of the entire comparison: 725066 versus 526003, or 37.8 percent. Integer math is close behind at 35 percent, with 1203634 versus 891817.

The multithread PassMark score gives the Threadripper 171200 versus 147998, a 15.7 percent edge, which is smaller than the Cinebench margins but still decisive. Random string sorting shows 418973 versus 348500, a 20.2 percent win. Single-thread performance narrows to 8.8 percent: 4542 versus 4173.

The single EPYC win comes in PassMark physics. The EPYC 9575F scores 20652, beating the Threadripper's 16860 by 18.4 percent. This is the only test where the EPYC's higher base clock and memory bandwidth overcome the Threadripper's core and cache advantages. It is worth remembering the physics test is a specialized workload, and the EPYC's loss everywhere else is consistent with its lower core count and smaller L3 cache.

The overall record stands at 16 wins for the Threadripper PRO 9995WX and 1 win for the EPYC 9575F. The Threadripper's average benchmark score of 412068 places it in the 100th percentile, while the EPYC's 311774 places it in the 99th. In the nearest rival comparisons, the Threadripper sits 3.3 percent below the EPYC 9745, 3.6 percent above the EPYC 9655P, 8.6 percent above the EPYC 9535, and 10.3 percent above the EPYC 9655. The EPYC 9575F, by contrast, is 0.4 percent above the EPYC 9734, 1.2 percent below the Intel Xeon 6781P, 2.8 percent below the Threadripper PRO 9985WX, and 3.1 percent below the Threadripper 9980X.

For buyers weighing these two, the decision hinges on workload. The Threadripper PRO 9995WX dominates rendering, compression, encryption, and math-heavy tasks. The EPYC 9575F wins in physics simulation and offers substantially more memory bandwidth for data streaming, but its overall benchmark profile trails across nearly every measured category.