AMD EPYC 9475F vs AMD Ryzen Threadripper 9980X Comparison

The benchmark data shows a clear overall winner, but the EPYC 9475F holds two decisive, workload-specific victories that make it indispensable for certain tasks. The AMD Ryzen Threadripper 9980X wins 9 of 11 head-to-head tests, leveraging its higher core count and clock speeds to dominate most compute-heavy PassMark workloads. However, the AMD EPYC 9475F posts a 96% advantage in prime number finding and a 105.5% lead in physics simulations, showcasing its architectural optimization for specific instruction patterns. The Threadripper's average benchmark score of 321,753 places it in the 99th percentile of all CPUs, while the EPYC achieves a perfect 100th percentile ranking despite a lower average score of 350,933, a distinction driven by the EPYC's consistency across a broader range of workloads.

Head-to-Head Benchmarks

The Threadripper 9980X establishes its dominance in raw multithreaded throughput, scoring 141,641 in PassMark's multithread test, which is 13.5% higher than the EPYC's 122,476. This advantage extends to integer math, where the Threadripper's 872,071 score is 30.5% higher than the EPYC's 605,696. The Threadripper also leads in floating-point math, posting 559,003 versus 406,524, a 27.3% margin. In data compression, the Threadripper delivers 2,974,534, beating the EPYC's 2,156,305 by 27.5%. These results correlate with the Threadripper's 64 cores and 128 threads, which outnumber the EPYC's 48 cores and 96 threads.

The Threadripper's single-thread performance is equally impressive, scoring 4,537 in PassMark's single-thread test, 16.7% ahead of the EPYC's 3,779. This advantage is reflected in its 5.40 GHz boost clock, which exceeds the EPYC's 4.80 GHz. Data encryption also favors the Threadripper, with a score of 157,137 versus 116,648, representing a 25.8% lead. Extended instruction workloads show a 24.4% advantage for the Threadripper, scoring 228,959 against the EPYC's 173,169. Random string sorting completes the Threadripper's sweep of general-purpose tasks, with 292,083 versus 253,936, a 13.1% difference.

The EPYC 9475F claims victory in two highly specialized tests. In PassMark's find prime numbers test, the EPYC scores 1,507, which is 96% higher than the Threadripper's 769. This is a staggering margin that points to a fundamental difference in how each processor handles this specific mathematical operation. The EPYC also wins the physics test with 16,443 versus 8,001, a 105.5% advantage. These two wins are not anomalies; they represent a consistent pattern where the EPYC's server-oriented design excels at certain types of integer and physics calculations, likely due to its larger L3 cache and different memory architecture. The EPYC's wins account for its 100th percentile ranking, while the Threadripper's nine wins position it at the 99th percentile.

FAQ

Q: Which processor has the higher average benchmark score?

A: The AMD EPYC 9475F has a higher average benchmark score of 350,933, compared to the AMD Ryzen Threadripper 9980X's 321,753. This is notable because the Threadripper wins more individual tests, but the EPYC's performance in its two strongest tests boosts its average.

Q: How does the EPYC 9475F compare to its nearest rivals?

A: The EPYC's nearest rival is the AMD EPYC 9754, which has an average score of 364,371, 3.7% higher. The Intel Xeon 6960P scores 365,194, 3.9% higher, and the AMD EPYC 9655 scores 373,479, 6% higher. The AMD EPYC 9535 is 7.5% higher with a score of 379,408.

Q: Where does the Threadripper 9980X rank among its competitors?

A: The Threadripper's closest rival is the AMD Ryzen Threadripper PRO 9985WX, which scores 320,749, just 0.3% lower. The Intel Xeon 6781P scores 315,524, 2% lower, while the AMD EPYC 9575F scores 311,774, 3.2% lower. The AMD EPYC 9734 scores 310,619, 3.6% lower.

Q: What is the difference in core and thread counts?

A: The Threadripper 9980X has 64 cores and 128 threads, while the EPYC 9475F has 48 cores and 96 threads. This gives the Threadripper a 33% advantage in core count and a 33% advantage in thread count.

Q: Which processor has the higher boost clock speed?

A: The Threadripper 9980X has a higher boost clock of 5.40 GHz, while the EPYC 9475F has a boost clock of 4.80 GHz. This contributes to the Threadripper's 16.7% lead in single-thread performance.

Q: How does memory bandwidth differ between the two?

A: The EPYC 9475F has a memory bandwidth of 576.0 GB/s, while the Threadripper 9980X has a memory bandwidth of 204.8 GB/s. This is a significant difference, with the EPYC offering nearly three times the bandwidth.

Architecture Differences

Both processors are built on the Zen 5 architecture, but they belong to different families with distinct design goals. The EPYC 9475F, codenamed Turin, is part of the EPYC 9005 series designed for server and workstation markets. The Threadripper 9980X, codenamed Shimada Peak, is part of the 9000 series aimed at desktop enthusiasts. Both use a 4 nm process node from TSMC and feature 66,520 million transistors with a die size of 8x 70.6 mm². This shared foundation means the core architecture is identical, but the surrounding infrastructure differs substantially.

The cache hierarchy shows a key difference in L1 cache size. The EPYC 9475F has 80 KB of L1 cache per core, while the Threadripper 9980X has 64 KB per core. Both processors share 1 MB of L2 cache per core and 256 MB of L3 cache, but the EPYC's L3 is listed as shared while the Threadripper's is not specified as shared. The larger L1 cache on the EPYC may contribute to its superior performance in prime number finding and physics simulations, as these workloads can benefit from faster access to frequently used data.

Memory architecture diverges significantly. The EPYC supports twelve-channel memory, while the Threadripper supports quad-channel memory. This translates to a memory bandwidth of 576.0 GB/s for the EPYC versus 204.8 GB/s for the Threadripper. Both support DDR5 memory and ECC, but the EPYC's wider memory bus is a critical advantage for server workloads that require high throughput. The PCIe configuration also differs, with the EPYC offering 128 Gen 5 lanes (CPU only) compared to the Threadripper's 80 Gen 5 lanes (CPU only).

Specification Differences

The core count is a primary differentiator, with the EPYC 9475F offering 48 cores and 96 threads, while the Threadripper 9980X offers 64 cores and 128 threads. Base clocks differ, with the EPYC running at 3.65 GHz and the Threadripper at 3.20 GHz. Boost clocks reverse this trend, with the EPYC reaching 4.80 GHz and the Threadripper reaching 5.40 GHz. Thermal design power (TDP) also differs, with the EPYC rated at 400 and the Threadripper at 350.

The socket types are incompatible: the EPYC uses AMD Socket SP5, while the Threadripper uses AMD Socket sTR5. Memory channels differ, with the EPYC supporting twelve channels and the Threadripper supporting four. This leads to the memory bandwidth difference of 576.0 GB/s versus 204.8 GB/s. PCIe lane counts differ as well, with the EPYC providing 128 lanes and the Threadripper providing 80 lanes. The EPYC's multiplier is locked, while the Threadripper's multiplier is unlocked, allowing for overclocking. The EPYC was released on 2024-10-09, while the Threadripper was released on 2025-07-29. The EPYC has a launch MSRP of $7592, while the Threadripper has a launch MSRP of $4999.

Where Each One Wins

The Threadripper 9980X wins in nine of the eleven benchmark tests, making it the clear choice for general-purpose high-performance computing. Its 30.5% lead in integer math and 27.3% lead in floating-point math indicate strong performance in scientific computing, financial modeling, and simulation workloads. The 27.5% advantage in data compression and 25.8% lead in data encryption make it suitable for database operations, file archiving, and secure communications. The 16.7% single-thread advantage, driven by its 5.40 GHz boost clock, benefits legacy software and applications that rely on single-core performance. The Threadripper's higher multithread score of 141,641 also makes it a strong candidate for video rendering, 3D animation, and other creative workloads that scale with core count.

The EPYC 9475F wins in two highly specific tests, but these wins are so pronounced that they define its use case. The 96% advantage in find prime numbers suggests exceptional performance in cryptography, number theory, and certain types of mathematical research. The 105.5% lead in physics simulations indicates superior handling of physics engines, particle simulations, and computational fluid dynamics. The EPYC's twelve-channel memory with 576.0 GB/s bandwidth provides a massive throughput advantage over the Threadripper's 204.8 GB/s, making it ideal for memory-intensive server workloads such as large-scale virtualization, in-memory databases, and high-performance computing clusters. The EPYC's 128 PCIe lanes also support more expansion cards, GPUs, and NVMe storage devices, making it the better choice for dense server configurations.

The Verdict

The data is unambiguous in general performance: the AMD Ryzen Threadripper 9980X is the faster processor for the majority of tasks. Its 64 cores, higher boost clock, and 30.5% lead in integer math make it the superior choice for content creation, software development, and any workload that benefits from parallel processing. The Threadripper's 99th percentile ranking reflects its broad applicability, and its 13.5% multithread advantage over the EPYC confirms its strength in mainstream high-performance computing. The Threadripper's unlocked multiplier and lower TDP of 350 also make it a more flexible option for desktop users who may want to overclock or build a system with lower cooling requirements.

The AMD EPYC 9475F, despite winning only two tests, is the undisputed champion in those specific domains. The 96% lead in prime number finding and 105.5% lead in physics simulations are not marginal differences; they are transformative advantages for researchers and engineers working in these areas. The EPYC's 100th percentile ranking, achieved despite a lower average score, suggests that its performance is more consistent across the full spectrum of benchmark tests. For server deployments requiring maximum memory bandwidth of 576.0 GB/s, twelve-channel memory support, and 128 PCIe lanes, the EPYC is the only logical choice. Its launch MSRP of $7592 is higher than the Threadripper's $4999, but the architectural benefits for server workloads justify the cost for enterprise buyers. Ultimately, the Threadripper wins the head-to-head on volume of victories, but the EPYC wins where it matters most for its intended market.