AMD EPYC 9754 vs AMD Ryzen Threadripper 9980X Comparison

This comparison pits two extreme AMD processors against each other: the EPYC 9754, a 128-core Bergamo server chip built for scale-out density, and the Ryzen Threadripper 9980X, a 64-core Zen 5 desktop flagship aimed at high-frequency workstations. The data shows a clear split: the Threadripper dominates in nearly every compute-heavy benchmark, while the EPYC counters in memory-bandwidth-sensitive and server-specific workloads.

Head-to-Head Benchmarks

The most striking result is the Threadripper 9980X’s near-total sweep of the Cinebench suite. In Cinebench R15, R20, and R23, the Threadripper wins both single-core and multi-core tests by a uniform delta of 35.7%. Specifically, the 9980X scores 13,157 points in Cinebench R15 multi-core against the EPYC’s 8,460, and 130,529 in R23 multi-core versus 83,939. This is a substantial victory for the 9980X, despite having half the core count of the EPYC 9754. The single-core gap is even more lopsided in raw terms: 1,857 versus 1,194 in R15 single-core, and 18,427 versus 11,850 in R23 single-core.

The Passmark suite tells a similar story, but with a few notable exceptions. The Threadripper wins 14 of the 17 head-to-head tests. Its biggest wins come in single-threaded performance, where it scores 4,591 versus the EPYC’s 2,328, a 49.3% advantage. It also leads in prime number finding (837 vs. 604, a 27.8% delta), extended instructions (252,783 vs. 224,322, an 11.3% delta), and random string sorting (347,099 vs. 306,481, an 11.7% delta). The Passmark multi-thread score follows the Cinebench pattern: the Threadripper hits 153,564, while the EPYC manages 98,752, another 35.7% delta. Even in floating-point math, where the EPYC’s 128 cores might be expected to help, the Threadripper edges ahead, scoring 592,543 versus 588,187, a slim 0.7% margin.

However, the EPYC 9754 claims three decisive wins that highlight its server pedigree. The largest is in data encryption, where it scores 231,891 against the Threadripper’s 159,952, a 45% advantage. This is a massive gap, likely reflecting the EPYC’s Zen 4c architecture’s strength in cryptographic workloads. The EPYC also wins in data compression (3,558,043 vs. 3,204,633, an 11% delta) and integer math (1,026,896 vs. 900,760, a 14% delta). These three wins suggest that for specific server tasks, the EPYC’s raw core count still matters, even against a much faster per-core design.

Looking at the average benchmark scores, the EPYC 9754 holds a 5.7% lead over the Threadripper 9980X (364,371 vs. 344,569). This is a curious inversion of the individual benchmark results, which the Threadripper dominates. The explanation lies in the weight of the Passmark encryption and compression scores, where the EPYC’s wins are so large that they lift its average above the Threadripper’s despite the latter’s broader success. In the nearestRivals data, the EPYC is 5.7% ahead of the Threadripper, while the Threadripper is 5.4% behind the EPYC, confirming this aggregated view. Yet the head-to-head table shows the Threadripper winning 14 out of 17 tests, meaning the EPYC’s victories are concentrated in a few high-scoring categories.

Architecture Differences

The two chips represent divergent design philosophies within AMD’s lineup. The EPYC 9754 uses the Zen 4c architecture, codenamed Bergamo, built on a 5 nm process at TSMC. It packs 128 cores and 256 threads, a density-optimized design with a base clock of 2.25 GHz and a boost clock of 3.10 GHz. The Threadripper 9980X, by contrast, uses the newer Zen 5 architecture, codenamed Shimada Peak, on a 4 nm process. It has 64 cores and 128 threads, but runs at much higher clocks: 3.20 GHz base and 5.40 GHz boost. This clock advantage, combined with the architectural generational leap, explains the Threadripper’s dominant single-core and multi-core Cinebench results.

The die and transistor counts are also telling. The EPYC 9754 has 71,000 million transistors spread across eight 73 mm² dies. The Threadripper 9980X has 66,520 million transistors on eight 70.6 mm² dies. Despite the Threadripper having fewer cores, its transistor count is only slightly lower, indicating a more complex core design with more per-core logic. The cache structure is identical on paper: 64 KB L1 per core, 1 MB L2 per core, and 256 MB of shared L3. However, the effective use of that cache differs due to the core count and clock speeds.

Memory support is a major differentiator. The EPYC 9754 uses a twelve-channel DDR5 memory bus, delivering a peak bandwidth of 460.8 GB/s. The Threadripper 9980X uses a quad-channel bus, with a peak of 204.8 GB/s. This more than double bandwidth is likely the reason for the EPYC’s victory in data compression and encryption, which are often memory-latency and bandwidth-sensitive. The EPYC also supports PCIe Gen 5 with 128 lanes, while the Threadripper offers 80 lanes. Both support ECC memory, but the EPYC’s platform is clearly designed for massive I/O and memory throughput.

The sockets differ as well: the EPYC uses AMD Socket SP5, while the Threadripper uses sTR5. The EPYC is a server/workstation part, while the Threadripper is classified as a desktop part. The Threadripper has an unlocked multiplier, allowing overclocking, whereas the EPYC is locked. The EPYC’s TDP is 360 W, slightly higher than the Threadripper’s 350 W. The EPYC launched in June 2023 with a launch MSRP of $11900, while the Threadripper launched in July 2025 with a launch MSRP of $4999.

FAQ

Q: Which processor has more cores?

A: The AMD EPYC 9754 has 128 cores and 256 threads, double the Threadripper 9980X’s 64 cores and 128 threads.

Q: Why does the Threadripper 9980X win in Cinebench multi-core despite having fewer cores?

A: The Threadripper’s Zen 5 architecture runs at 5.40 GHz boost, versus the EPYC’s 3.10 GHz boost. This higher frequency, combined with the newer core design, allows it to outperform the EPYC’s higher core count in Cinebench tests, where it leads by 35.7% in every R15, R20, and R23 multi-core test.

Q: In which benchmarks does the EPYC 9754 beat the Threadripper?

A: The EPYC wins in three Passmark tests: data encryption (231,891 vs. 159,952, a 45% delta), data compression (3,558,043 vs. 3,204,633, an 11% delta), and integer math (1,026,896 vs. 900,760, a 14% delta).

Q: How does memory bandwidth differ between the two?

A: The EPYC 9754 has a twelve-channel DDR5 memory bus with 460.8 GB/s of bandwidth, while the Threadripper 9980X has a quad-channel bus with 204.8 GB/s. This difference is likely the key factor in the EPYC’s encryption and compression wins.

Q: What is the average benchmark score difference?

A: The EPYC 9754 has an average benchmark score of 364,371, which is 5.7% higher than the Threadripper’s 344,569. The Threadripper’s own nearestRivals data shows it is 5.4% behind the EPYC.

Q: Are both processors unlocked for overclocking?

A: No. The Ryzen Threadripper 9980X has an unlocked multiplier, while the EPYC 9754 does not.

The Verdict

The data points to two distinct use cases. The Ryzen Threadripper 9980X is the clear choice for compute-heavy workloads that favor high clock speeds and modern core architecture. It wins 14 of 17 head-to-head benchmarks, including every Cinebench test and the Passmark multi-thread score. For tasks like rendering, video encoding, and simulation, where Cinebench-style performance is representative, the Threadripper is decisively faster, despite having half the cores.

The EPYC 9754, however, is the pick for server workloads that leverage massive memory bandwidth and parallel I/O. Its wins in data encryption (45% ahead), data compression (11% ahead), and integer math (14% ahead) are significant, and its 128 PCIe Gen 5 lanes and twelve-channel memory bus make it suited for database, virtualized, and high-throughput server environments. The average benchmark score also favors the EPYC by 5.7%, indicating that for a balanced mix of server tasks, it edges ahead.

For a desktop workstation user, the Threadripper 9980X is the obvious choice. It offers superior single-threaded performance (4,591 vs. 2,328 in Passmark single-thread, a 49.3% delta) and wins every multi-threaded render test. The unlocked multiplier adds flexibility for enthusiasts. For a data center operator, the EPYC 9754’s core density, higher memory bandwidth, and lower launch MSRP per core make it a more appropriate fit. The choice is not about which is “better” overall, but which architecture aligns with the workload’s bottlenecks: raw frequency versus raw throughput.

Specification Differences

| Specification | AMD EPYC 9754 | AMD Ryzen Threadripper 9980X |

| :--- | :--- | :--- |

| Cores | 128 | 64 |

| Threads | 256 | 128 |

| Base Clock | 2.25 GHz | 3.20 GHz |

| Boost Clock | 3.10 GHz | 5.40 GHz |

| TDP | 360 W | 350 W |

| Socket | AMD Socket SP5 | AMD Socket sTR5 |

| Architecture | Zen 4 (Bergamo) | Zen 5 (Shimada Peak) |

| Process Node | 5 nm | 4 nm |

| Transistors | 71,000 million | 66,520 million |

| Die Size | 8x 73 mm² | 8x 70.6 mm² |

| Memory Bus | Twelve-channel | Quad-channel |

| Memory Bandwidth | 460.8 GB/s | 204.8 GB/s |

| PCIe Lanes | Gen 5, 128 Lanes | Gen 5, 80 Lanes |

| Multiplier Unlocked | No | Yes |

| Market Segment | Server/Workstation | Desktop |

| Release Date | 2023-06-12 | 2025-07-29 |

| Launch MSRP | $11900 | $4999 |