AMD EPYC 9655P vs AMD EPYC 9755 Comparison

The AMD EPYC 9755 and AMD EPYC 9655P are both top-tier server processors in the EPYC 9005 series, sharing the Zen 5 architecture and 4 nm process node. However, the data shows they are positioned for different workloads: the 9755 is the undisputed multi-core champion with a 27.2% higher average benchmark score than the 9655P, while the 9655P counters with a 9% advantage in single-threaded PassMark performance. This guide breaks down the benchmark results, architectural differences, and use cases to help determine which processor fits specific server roles.

FAQ

Q: Which processor has the higher average benchmark score?

A: The AMD EPYC 9755 leads with an average benchmark score of 505,778, while the AMD EPYC 9655P scores 397,773. The 9755 is 27.2% ahead of the 9655P, according to the nearestRivals deltaPct data.

Q: What is the core and thread count difference?

A: The EPYC 9755 features 128 cores and 256 threads, while the EPYC 9655P has 96 cores and 192 threads. The 9755 offers 32 additional cores and 64 additional threads.

Q: How do the two compare in single-threaded PassMark performance?

A: The EPYC 9655P wins the passmark_single_thread test with a score of 3849, defeating the EPYC 9755's 3503. This represents a 9% advantage for the 9655P.

Q: Does the EPYC 9755 win all Cinebench tests?

A: Yes, the EPYC 9755 wins all six Cinebench tests (R15, R20, R23, both multi-core and single-core variants), with a consistent 3.7% deltaPct advantage over the EPYC 9655P in each test.

Q: What are the TDP ratings for these processors?

A: The EPYC 9755 has a TDP of 500 watts, while the EPYC 9655P has a TDP of 400 watts. The 9755 consumes more power to drive its higher core count.

Q: Which processor has more L3 cache?

A: The EPYC 9755 has 512 MB of shared L3 cache, while the EPYC 9655P has 384 MB of shared L3 cache. The 9755 offers 128 MB more L3 cache.

Architecture Differences

Both processors are built on AMD's Zen 5 architecture, codenamed Turin, and are manufactured by TSMC on a 4 nm process node. They share the same AMD Socket SP5, support DDR5 memory with a twelve-channel memory bus, and offer 576.0 GB/s of memory bandwidth. Both also feature 128 PCIe Gen 5 lanes (CPU only) and have no integrated graphics. The core architecture is identical: each core has 80 KB of L1 cache and 1 MB of L2 cache.

The primary architectural difference lies in the number of compute chiplets and the associated transistor count. The EPYC 9755 uses 16x 70.6 mm² dies with 133,040 million transistors, whereas the EPYC 9655P uses 12x 70.6 mm² dies with 99,780 million transistors. This translates to the 9755 having 128 cores versus the 9655P's 96 cores, and a corresponding 512 MB versus 384 MB of shared L3 cache. The 9755's additional four chiplets deliver 33,260 million more transistors, which directly accounts for its higher core count and larger cache pool.

Clock speeds also differ, though not in the direction one might expect from the core counts. The EPYC 9755 has a base clock of 2.70 GHz and a boost clock of 4.10 GHz. The EPYC 9655P has a lower base clock of 2.60 GHz but a higher boost clock of 4.50 GHz. This higher boost clock on the 9655P explains its single-threaded PassMark advantage, as fewer cores allow for more aggressive frequency headroom. Both processors are not multiplier-unlocked, and both have active production status with identical release dates of 2024-10-09.

Head-to-Head Benchmarks

The benchmark data shows a clear split between multi-threaded dominance and single-threaded efficiency. Across 17 head-to-head tests, the EPYC 9755 wins 15, while the EPYC 9655P wins only 2.

In Cinebench tests, the EPYC 9755 maintains a consistent 3.7% lead across all six tests. The scores are: R15 multi-core 14250 vs 13744, R15 single-core 2011 vs 1940, R20 multi-core 59378 vs 57268, R20 single-core 8382 vs 8085, R23 multi-core 141378 vs 136354, and R23 single-core 19959 vs 19250. This uniformity suggests the 9755's advantage scales evenly across rendering workloads, regardless of thread count.

The PassMark suite reveals where the 9755 pulls decisively ahead. The largest margin is in passmark_extended_instructions, where the 9755 scores 303321 versus 230609, a 31.5% advantage. Data compression shows a 29.6% lead (4517407 vs 3486158), and data encryption follows closely at 29.5% (284927 vs 220074). Floating point math favors the 9755 by 28.9% (922900 vs 715866), while integer math shows a 26.5% edge (1549946 vs 1225251). Random string sorting is 26.4% faster on the 9755 (571185 vs 451824), and prime number finding is 21.4% ahead (2047 vs 1686).

In passmark_multithread, the 9755 leads by a smaller 3.6% (166328 vs 160490), and passmark_physics shows a 7.6% advantage (27806 vs 25847). The only tests where the 9655P wins are passmark_single_thread and passmark_singlethread, both scoring 3849 versus 3503 for the 9755, a 9% delta. These results show the 9755's 32 additional cores provide massive gains in parallelizable workloads, while the 9655P's higher boost clock yields a meaningful single-thread edge.

The Verdict

The data strongly favors the AMD EPYC 9755 for most server workloads. Its 27.2% higher average benchmark score, combined with wins in 15 of 17 tests, makes it the clear choice for compute-intensive applications that scale with core count. The 9755's 128 cores and 512 MB L3 cache deliver 29-31% improvements in extended instructions, data compression, and encryption, which are critical for database, virtualization, and scientific computing tasks. Its consistent 3.7% Cinebench lead indicates solid rendering performance across all thread scales.

The AMD EPYC 9655P is not without merit, however. Its 9% single-thread PassMark victory and higher 4.50 GHz boost clock make it suitable for workloads that prioritize per-core performance over raw throughput. The 9655P's lower TDP of 400 watts versus 500 watts also suggests better power efficiency per core, though the total power draw is lower. For users with workloads that are heavily single-threaded or latency-sensitive, the 9655P's frequency advantage is a tangible benefit.

The choice ultimately depends on whether the workload scales with core count. The 9755 is the definitive multi-core leader, while the 9655P offers a more balanced profile with superior single-thread speed. The 9755's 18.7% average score advantage over the EPYC 9745 and its 100th percentile ranking among all CPUs underscores its flagship status. The 9655P also ranks at the 100th percentile, but its average score of 397773 places it 3.5% behind the AMD Ryzen Threadripper PRO 9995WX, showing it is still elite but not the absolute top performer.

Specification Differences

The following table highlights only the fields where the two processors differ:

| Specification | AMD EPYC 9755 | AMD EPYC 9655P |

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

| Cores | 128 | 96 |

| Threads | 256 | 192 |

| Base Clock | 2.70 GHz | 2.60 GHz |

| Boost Clock | 4.10 GHz | 4.50 GHz |

| TDP | 500 W | 400 W |

| Transistors | 133,040 million | 99,780 million |

| Die Size | 16x 70.6 mm² | 12x 70.6 mm² |

| L3 Cache | 512 MB (shared) | 384 MB (shared) |

| Launch MSRP | $12984 | $10811 |

| Part Number | 100-000001443 | 100-000001522 |

Where Each One Wins

The AMD EPYC 9755 wins decisively in multi-threaded and data-intensive workloads. Its 29.6% advantage in data compression and 29.5% edge in data encryption make it ideal for storage servers, database engines, and security applications that process large volumes of data. The 31.5% lead in extended instructions (SIMD operations) is crucial for scientific simulations, financial modeling, and machine learning inference. The 28.9% floating point and 26.5% integer math gains further cement its position for high-performance computing and rendering tasks. For any workload that can utilize more than 96 cores, the 9755's 128-core configuration delivers 3.6-7.6% improvements in multithreaded and physics benchmarks, making it the superior choice for heavy virtualization and cloud infrastructure.

The AMD EPYC 9655P wins in single-threaded scenarios. Its 9% passmark_single_thread advantage and 4.50 GHz boost clock make it the better pick for applications with low thread counts but high per-core frequency requirements, such as certain legacy enterprise software, real-time processing tasks, or latency-critical web serving workloads. The 9655P's lower TDP of 400 watts also makes it more suitable for dense server deployments where power and cooling are constrained, even if its overall performance is lower. The 9655P also offers a more accessible entry point with a lower launch MSRP of $10811 compared to the 9755's $12984, though both are premium server processors. For users whose primary workloads are single-threaded, the 9655P's frequency advantage provides a tangible performance benefit that the 9755 cannot match.