AMD EPYC 9555P vs AMD EPYC 9655P Comparison

FAQ

Q: Which processor is faster in multi-threaded work, the AMD EPYC 9655P or the AMD EPYC 9555P?

A: The AMD EPYC 9655P wins every multi-threaded benchmark in the database. In Cinebench R23 multi-core, it scores 136354 against 115186 for the 9555P, an 18.4% advantage. The gap widens further in PassMark multi-thread, where the 9655P leads by 29.9% (160490 vs 123576).

Q: Is the AMD EPYC 9655P also faster in single-thread performance?

A: Yes, the 9655P holds the single-thread lead as well. In Cinebench R23 single-core, it scores 19250 versus 16261 for the 9555P, an 18.4% delta. PassMark single-thread shows a smaller but still clear 12.9% advantage (3849 vs 3410).

Q: What is the biggest performance gap between these two chips?

A: The largest delta appears in PassMark physics, where the 9655P beats the 9555P by 67% (25847 vs 15474). PassMark find prime numbers also shows a 58% gap (1686 vs 1067), indicating the 9655P scales exceptionally well in integer-heavy and simulation workloads.

Q: Do both processors use the same socket and memory configuration?

A: Yes, both use AMD Socket SP5, support DDR5 memory, and run on a twelve-channel memory bus with 576.0 GB/s of bandwidth. They also share the same PCIe configuration: Gen 5 with 128 lanes (CPU only).

Q: What is the difference in thermal design power (TDP)?

A: The AMD EPYC 9655P has a TDP of 400 watts, while the AMD EPYC 9555P is rated at 360 watts. The 40-watt difference reflects the 9655P's higher core count and slightly higher boost clock.

Q: How do these chips compare to their nearest rivals in the database?

A: The 9655P sits at the 100th percentile among all CPUs, with an average benchmark score of 397773. It trails the AMD Ryzen Threadripper PRO 9995WX by 3.5% but leads the AMD EPYC 9535 by 4.8% and the AMD EPYC 9655 by 6.5%. The 9555P ranks at the 99th percentile, scoring 287066 on average, and edges out the Intel Xeon 696X by 0.3% while leading the AMD EPYC 9565 by 0.6%.

Architecture Differences

Both the AMD EPYC 9655P and the AMD EPYC 9555P belong to the EPYC 9005 series and share the Zen 5 architecture with the codename Turin. Both are built on TSMC's 4 nm process node, and both are active production parts released on 2024-10-09.

The core configuration is where they diverge. The 9655P packs 96 cores and 192 threads, while the 9555P offers 64 cores and 128 threads. This 50% core-count advantage is the primary driver of the 9655P's performance lead in heavily parallel workloads. Transistor counts scale accordingly: the 9655P uses 99,780 million transistors across 12x 70.6 mm² dies, while the 9555P uses 66,520 million transistors across 8x 70.6 mm² dies.

Cache hierarchies also differ. Both chips provide 80 KB of L1 cache per core and 1 MB of L2 cache per core. The shared L3 cache, however, is 384 MB on the 9655P versus 256 MB on the 9555P. That extra 128 MB of L3 on the 9655P can be decisive for datasets that fit within the larger cache footprint, especially in database and virtualization workloads.

Base and boost clocks favor the smaller chip slightly. The 9555P runs a base clock of 3.20 GHz and boosts to 4.40 GHz, while the 9655P has a lower base of 2.60 GHz but a higher boost of 4.50 GHz. The 9655P's higher boost clock explains why it also wins single-thread tests despite having more cores competing for thermal headroom.

Both processors share identical memory and I/O capabilities: DDR5 with twelve-channel support, 576.0 GB/s of bandwidth, ECC memory support, and PCIe Gen 5 with 128 lanes (CPU only). Neither has integrated graphics, and both are locked (multiplier unlocked: false). The launch MSRP for the 9655P is $10811, while the 9555P is $7983.

Head-to-Head Benchmarks

The database records 17 head-to-head benchmark comparisons, and the AMD EPYC 9655P wins all 17. The 9555P does not record a single victory. Here is how the margins break down.

In Cinebench R15, R20, and R23, the 9655P leads by a consistent 18.4% in both multi-core and single-core tests. Whether the workload is fully threaded or lightly threaded, the 9655P maintains the same relative advantage, which is notable because single-core performance typically depends more on clock speed than core count. The 9655P's 4.50 GHz boost clock versus 4.40 GHz on the 9555P explains this narrow but consistent edge.

PassMark results show larger deltas in specific workloads. Data encryption favors the 9655P by 47.8% (220074 vs 148896), indicating strong scaling for cryptographic workloads. Floating-point math shows a 47.2% gap (715866 vs 486407), while integer math is even wider at 55.7% (1225251 vs 787106). Find prime numbers, a test sensitive to both core count and integer throughput, records a 58% delta (1686 vs 1067).

The single biggest margin is in PassMark physics at 67% (25847 vs 15474). This test simulates rigid-body physics and benefits heavily from raw core count. Random string sorting also shows a large 61.1% delta (451824 vs 280398), reflecting the 9655P's advantage in memory-bound sorting tasks. Data compression trails at 32.1% (3486158 vs 2639400), and extended instructions are 20.7% ahead (230609 vs 191082).

PassMark single-thread is the closest contest at 12.9% (3849 vs 3410). Even here, the 9655P wins, which reinforces that the higher boost clock on the 9655P is enough to keep it ahead in lightly threaded scenarios.

The Verdict

The data paints a clear picture: the AMD EPYC 9655P is the superior processor in every recorded benchmark. If raw performance is the only criterion, the choice is unambiguous. The 9655P delivers 18.4% higher multi-core scores in Cinebench, 29.9% higher PassMark multi-thread results, and margins exceeding 50% in integer math and physics workloads. For compute-heavy server deployments, database workloads, or virtualization hosts, the 9655P is the pick.

The AMD EPYC 9555P, however, is not a weak part. It sits at the 99th percentile of all CPUs and matches or slightly beats its nearest rivals in the database. It has a higher base clock (3.20 GHz vs 2.60 GHz) and a 40-watt lower TDP, which can simplify cooling and power delivery in dense chassis. For workloads that do not require 96 cores, the 9555P provides 64 cores with a faster base frequency and a lower launch MSRP.

Choose the 9655P when core count and L3 cache size drive performance, such as large-scale parallel processing, scientific computing, or heavy virtualization. Choose the 9555P when the workload is more modest, when power constraints matter, or when the lower launch MSRP aligns better with the project budget. The data does not show any scenario where the 9555P outperforms the 9655P, but it does show a more efficient, lower-power alternative for mid-range server builds.

Specification Differences

| Specification | AMD EPYC 9655P | AMD EPYC 9555P |

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

| Cores | 96 | 64 |

| Threads | 192 | 128 |

| Base Clock | 2.60 GHz | 3.20 GHz |

| Boost Clock | 4.50 GHz | 4.40 GHz |

| TDP | 400 W | 360 W |

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

| Die Size | 12x 70.6 mm² | 8x 70.6 mm² |

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

| Launch MSRP | $10811 | $7983 |

| Part Number | 100-000001522 | 100-000001523 |

Both chips share the same process node (4 nm TSMC), socket (SP5), memory support (DDR5, twelve-channel, 576.0 GB/s), PCIe configuration (Gen 5, 128 lanes), and lack of integrated graphics. Both are locked multipliers and belong to the EPYC 9005 series with Zen 5 (Turin) architecture.

The 9655P has a 50% core advantage, a 50% larger L3 cache, and a slightly higher boost clock. The 9555P counters with a higher base clock, lower TDP, and lower launch MSRP. The power-per-core ratio favors the 9555P: at 360 watts for 64 cores, it draws 5.6 watts per core, while the 9655P draws 4.2 watts per core at 400 watts for 96 cores.

Where Each One Wins

The AMD EPYC 9655P wins every benchmark category in the database, but its margins vary by workload type. The largest wins come in physics (67%), random string sorting (61.1%), and integer math (55.7%). These are workloads that scale almost linearly with core count and memory bandwidth. The 9655P is the clear choice for high-performance computing clusters, large-scale data analytics, and simulation workloads where every additional core translates directly into throughput.

The 9655P also excels in data encryption (47.8% ahead), making it suitable for secure cloud infrastructure, VPN gateways, and encrypted storage servers. Its floating-point advantage (47.2%) benefits scientific computing and machine learning inference. The 32.1% lead in data compression suits backup servers and content delivery networks that handle large volumes of compressed data.

The AMD EPYC 9555P does not win any benchmark, but it wins in operational characteristics. Its 360-watt TDP makes it easier to cool in a 1U or 2U server chassis. Its 3.20 GHz base clock provides snappy responsiveness for interactive workloads or real-time processing where sustained all-core frequency matters more than peak boost. The 64-core count is still substantial for most enterprise workloads, and the lower launch MSRP allows more budget for memory, storage, or additional nodes.

For a single-socket server handling mixed enterprise workloads, the 9555P offers a balanced profile: strong multi-threading with lower power draw. For a dedicated compute node where performance is the sole priority, the 9655P's 17-win sweep and 100th percentile ranking make it the definitive choice. The database shows no workload category where the 9555P closes the gap, so the decision rests on power, thermal, and cost constraints rather than performance.