AMD EPYC 9535 vs AMD Ryzen Threadripper 9980X Comparison

Head-to-Head Benchmarks

The benchmark data presents a decisive picture: the AMD Ryzen Threadripper 9980X wins 10 of the 11 head-to-head comparisons, with the AMD EPYC 9535 securing only a single victory. The most striking margin comes in `passmark_physics`, where the Threadripper scores 8001 against the EPYC's 3834, a 52.1% advantage. This is the largest delta in the entire dataset and signals a fundamental difference in how the two processors handle real-time physics calculations.

In `passmark_extended_instructions`, the Threadripper again dominates with a score of 228959 versus 175784, a 23.2% lead. This workload, typically involving SIMD and specialized instruction paths, shows the Threadripper's higher boost clock of 5.40 GHz against the EPYC's 4.30 GHz translating directly into throughput. The `passmark_data_compression` test follows a similar pattern: 2974534 for the Threadripper versus 2308822 for the EPYC, a 22.4% gap. Compression workloads are often memory-bandwidth-sensitive, yet the EPYC's twelve-channel memory bus (576.0 GB/s) does not compensate for the Threadripper's per-core clock advantage.

The `passmark_floating_point_math` and `passmark_integer_math` results show consistent Threadripper advantages of 18.2% and 16.3%, respectively. Scores of 559003 versus 457047 in floating-point, and 872071 versus 730281 in integer math, reinforce the pattern of higher single-core performance driving aggregate gains. The `passmark_multithread` score of 141641 for the Threadripper against 114528 for the EPYC represents a 19.1% lead, even though both chips have identical core counts of 64 and thread counts of 128. This is a notable result because multithreaded workloads typically scale with core count and memory bandwidth, yet the Threadripper still prevails.

In `passmark_data_encryption`, the Threadripper scores 157137 versus 127372, an 18.9% advantage. Encryption workloads often rely on AES-NI and other cryptographic instructions, where clock speed plays a significant role. The `passmark_random_string_sorting` test shows a 15.3% lead for the Threadripper (292083 versus 247506), while `passmark_single_thread` confirms the trend with 4537 versus 3720, an 18% delta. The EPYC's only win comes in `passmark_find_prime_numbers`, where it scores 874 against the Threadripper's 769, a 13.7% advantage. This workload, which involves iterative integer division and primality testing, appears to benefit from the EPYC's memory subsystem or cache hierarchy despite the clock deficit.

Where Each One Wins

The Threadripper 9980X wins across nearly every workload category represented in the benchmark suite. For users running physics simulations, the 52.1% lead in `passmark_physics` makes it the clear choice. For data compression tasks, the 22.4% advantage means faster archive creation and decompression cycles. Encryption workloads see an 18.9% improvement, which is material for any server handling TLS or disk encryption. The Threadripper's 18% lead in single-threaded performance also makes it superior for lightly threaded applications, legacy software, and any workload that cannot fully utilize 128 threads.

The EPYC 9535's single win in `passmark_find_prime_numbers` is narrow but real. At 13.7% ahead, it suggests that workloads dominated by integer division and modular arithmetic—common in cryptography key generation and some scientific computing—may favor the EPYC's architecture. However, this is an isolated data point. The EPYC's twelve-channel memory bus and 576.0 GB/s bandwidth versus the Threadripper's quad-channel 204.8 GB/s does not translate into wins in the memory-sensitive benchmarks like data compression or random string sorting. In fact, the Threadripper leads in both of those tests, indicating that raw memory bandwidth is not the limiting factor in these PassMark workloads.

For server workloads that emphasize instruction-level parallelism and branch prediction, the EPYC's 256 MB shared L3 cache is matched by the Threadripper's 256 MB L3, so cache capacity does not explain the divergence. The data suggests that the Threadripper's higher base clock (3.20 GHz vs 2.40 GHz) and boost clock (5.40 GHz vs 4.30 GHz) provide a per-thread throughput advantage that compounds across all 128 threads. The EPYC's 13.7% win in prime number finding is the exception that proves the rule: this particular algorithm is likely sensitive to memory latency rather than clock speed, and the EPYC's server-oriented memory controller may offer lower latency per access despite lower aggregate bandwidth.

Architecture Differences

Both processors are built on the same Zen 5 architecture, manufactured on TSMC's 4 nm process, with identical transistor counts of 66,520 million and die sizes of 8x 70.6 mm². The core counts are identical at 64 cores and 128 threads. The critical differences lie in clock speeds, memory configuration, socket, and platform features. The EPYC 9535 operates at a base clock of 2.40 GHz and boosts to 4.30 GHz, while the Threadripper 9980X runs at 3.20 GHz base and 5.40 GHz boost. This 1.10 GHz boost delta is the most significant architectural differentiator for performance.

The EPYC uses AMD Socket SP5, which supports twelve-channel DDR5 memory and delivers 576.0 GB/s of bandwidth. The Threadripper uses AMD Socket sTR5, which supports quad-channel DDR5 memory at 204.8 GB/s. Despite having one-third of the memory bandwidth, the Threadripper wins the memory-sensitive benchmarks, suggesting that the benchmarks are not bandwidth-limited at these thread counts. The EPYC provides 128 PCIe Gen 5 lanes (CPU only), while the Threadripper provides 80 lanes. For expansion-heavy server configurations, the EPYC's additional lanes are significant.

Cache layouts differ slightly: the EPYC has 80 KB of L1 per core, while the Threadripper has 64 KB of L1 per core. Both have 1 MB L2 per core and 256 MB of shared L3. The EPYC's larger L1 may contribute to its prime number finding win, as that workload benefits from tighter data locality. The Threadripper is multiplier-unlocked, enabling overclocking, while the EPYC is locked. The Threadripper's TDP is 350 W, higher than the EPYC's 300 W, reflecting its higher clock speeds. The EPYC is a server/workstation part released on 2024-10-09, while the Threadripper is a desktop part released on 2025-07-29. Both support ECC memory and have no integrated graphics.

The EPYC's part number is 100-000001147, and the Threadripper's is 100-000001593. The EPYC's codename is Turin, while the Threadripper's is Shimada Peak. The EPYC 9535's average benchmark score is 379408, placing it at the 100th percentile of all CPUs. The Threadripper's average is 321753, at the 99th percentile. This is a counterintuitive result: the Threadripper wins most head-to-head tests but has a lower average score, likely because the average includes a broader set of benchmarks where the EPYC's server-oriented design performs differently.

The Verdict

The data is unambiguous: the AMD Ryzen Threadripper 9980X outperforms the AMD EPYC 9535 in 10 of 11 benchmarks, with margins ranging from 15.3% to 52.1%. For any workload represented in this dataset, the Threadripper is the faster processor. The single exception is prime number finding, where the EPYC leads by 13.7%, but this is a narrow, specialized task. The Threadripper's higher clock speeds—3.20 GHz base and 5.40 GHz boost versus 2.40 GHz and 4.30 GHz—provide a consistent advantage across integer, floating-point, compression, encryption, and physics workloads. The EPYC's twelve-channel memory bus and 576.0 GB/s bandwidth do not translate into benchmark wins, suggesting that the PassMark suite favors clock speed over memory throughput.

The Threadripper's 52.1% lead in physics is particularly striking, as it indicates a fundamental efficiency advantage in that workload. Its 22.4% lead in data compression and 18.9% lead in encryption make it the superior choice for those tasks. The EPYC's 100th percentile ranking versus the Threadripper's 99th is explained by the average benchmark score calculation, which weights all benchmarks equally and may include tests not in the head-to-head set. However, for the specific comparisons available, the Threadripper is faster. The Threadripper also offers an unlocked multiplier for overclocking, while the EPYC does not, further extending its potential lead.

For server deployments requiring maximum PCIe lane count (128 vs 80) or twelve-channel memory bandwidth, the EPYC has a platform advantage. The EPYC's launch MSRP is $8992, while the Threadripper's launch MSRP is $4999. The Threadripper is also unlocked, allowing users to push beyond stock clocks. The verdict from the data is that the Threadripper 9980X is the higher-performing processor in almost every measurable way, with the EPYC retaining a narrow niche in prime number finding and offering more platform expansion options.

FAQ

Q: Which processor has a higher boost clock?

A: The AMD Ryzen Threadripper 9980X boosts to 5.40 GHz, while the AMD EPYC 9535 boosts to 4.30 GHz.

Q: How many cores and threads does each processor have?

A: Both the AMD EPYC 9535 and the AMD Ryzen Threadripper 9980X have 64 cores and 128 threads.

Q: What is the largest performance gap between the two in the head-to-head benchmarks?

A: The largest gap is in `passmark_physics`, where the Threadripper scores 8001 versus the EPYC's 3834, a 52.1% difference.

Q: Does the EPYC 9535 win any benchmark against the Threadripper 9980X?

A: Yes, the EPYC 9535 wins `passmark_find_prime_numbers` with a score of 874 versus 769, a 13.7% advantage.

Q: What memory bandwidth does each processor support?

A: The EPYC 9535 supports twelve-channel memory with 576.0 GB/s bandwidth, while the Threadripper 9980X supports quad-channel memory with 204.8 GB/s.

Q: Which processor is unlocked for overclocking?

A: The AMD Ryzen Threadripper 9980X has an unlocked multiplier; the AMD EPYC 9535 does not.