AMD EPYC 4565P vs AMD Ryzen 9 PRO 9955 Comparison

Where Each One Wins

The benchmark split between these two AMD parts is lopsided, but it tells a clear story. The AMD EPYC 4565P takes 9 of the 11 head-to-head comparisons in the database, while the AMD Ryzen 9 PRO 9955 wins only 2. That alone suggests the EPYC is the stronger all-rounder for most workloads, but the Ryzen's victories point to specific strengths that matter in certain scenarios.

The Ryzen 9 PRO 9955 wins in passmark_find_prime_numbers by a massive 56.8% margin, scoring 461 against the EPYC's 294. This is a single-threaded, latency-sensitive workload that rewards high per-core efficiency and low-level integer recursion. The Ryzen also wins passmark_physics, scoring 3332 versus 2976, a 12% lead. Physics simulation often favors cores with higher sustained single-thread performance and lower thread-scheduling overhead, so this win aligns with the prime number result.

Every other benchmark in the head-to-head set favors the EPYC 4565P. The biggest EPYC wins come in passmark_integer_math, where it scores 250683 against 182312, a 27.3% advantage. Data encryption shows a 30.1% lead for the EPYC, with 48268 versus 33754. Data compression follows at 20.5% ahead, 860786 versus 684470. Floating point math, extended instructions, random string sorting, and multithread all show the EPYC ahead by margins between 11.5% and 20.1%. Even single-thread performance, where the Ryzen might be expected to compete, goes to the EPYC by 2.4%, with 4712 versus 4597.

If the workload is heavily integer-based, encryption-heavy, or requires high aggregate throughput across many cores, the EPYC 4565P is the clear choice. If the task is more about per-core integer recursion or physics simulation with modest thread counts, the Ryzen 9 PRO 9955 has a niche advantage. The data does not show the Ryzen winning any throughput-oriented benchmark, so its role is narrower.

Architecture Differences

Both processors share the same fundamental silicon: TSMC 4 nm process, 16,630 million transistors, and a dual-die design with each die measuring 70.6 mm². Both use Zen 5 architecture, with the Ryzen codenamed Granite Ridge and the EPYC codenamed Grado. The core count differs: the Ryzen has 12 cores and 24 threads, while the EPYC has 16 cores and 32 threads. That 4-core, 8-thread gap is the primary driver of the EPYC's multithreaded wins.

Cache layouts are nearly identical on a per-core basis. Both have 80 KB of L1 per core and 1 MB of L2 per core. The L3 is 64 MB in both cases, though the database lists the Ryzen's L3 as simply "64 MB" while the EPYC's is marked as "64 MB (shared)". In practice, both are shared across the chiplet design, so the effective difference is negligible.

Clock speeds favor the EPYC. Its base clock is 4.30 GHz versus 3.40 GHz on the Ryzen, and its boost clock reaches 5.70 GHz versus 5.40 GHz. The EPYC also has a higher TDP at 170 watts versus 120 watts, which explains its ability to sustain those higher clocks. Both use dual-channel DDR5 memory with identical 89.6 GB/s bandwidth, and both support ECC memory. PCIe connectivity is the same: Gen 5 with 24 lanes from the CPU.

Both have Radeon integrated graphics, both are on AMD Socket AM5, and both are locked (multiplier not unlocked). The release dates differ, with the EPYC launching in May 2025 and the Ryzen in June 2026. The EPYC carries a launch MSRP of $589; the Ryzen has no listed launch MSRP. The market segment is identical: Server/Workstation. Neither is a mainstream desktop part, despite the AM5 socket. The Ryzen's part number is 100-000001971, while the EPYC's is 100-000001559.

FAQ

Q: Which CPU has more cores and threads?

A: The AMD EPYC 4565P has 16 cores and 32 threads, while the AMD Ryzen 9 PRO 9955 has 12 cores and 24 threads.

Q: Why does the Ryzen 9 PRO 9955 win in physics and prime number tests despite losing most other benchmarks?

A: Those two tests are passmark_find_prime_numbers (461 versus 294, a 56.8% lead) and passmark_physics (3332 versus 2976, a 12% lead). Both are more sensitive to per-core efficiency and scheduling than to raw core count, and the Ryzen's lower TDP suggests it may maintain tighter per-core performance in those specific workloads.

Q: Is the single-thread performance gap significant?

A: The EPYC leads by 2.4% in passmark_single_thread, with 4712 versus 4597. That is a small margin, but it is consistent across the two identical single-thread entries in the database.

Q: Do both CPUs support ECC memory and integrated graphics?

A: Yes. Both have ECC memory support and both include Radeon integrated graphics.

Q: Are these CPUs on the same socket?

A: Yes, both use AMD Socket AM5 and have Gen 5 PCIe with 24 lanes.

Q: Which CPU has the higher boost clock?

A: The EPYC 4565P boosts to 5.70 GHz, while the Ryzen 9 PRO 9955 boosts to 5.40 GHz.

Specification Differences

| Specification | AMD Ryzen 9 PRO 9955 | AMD EPYC 4565P |

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

| Cores | 12 | 16 |

| Threads | 24 | 32 |

| Base Clock | 3.40 GHz | 4.30 GHz |

| Boost Clock | 5.40 GHz | 5.70 GHz |

| TDP | 120 W | 170 W |

| Codename | Granite Ridge | Grado |

| L3 Cache | 64 MB | 64 MB (shared) |

| Release Date | 2026-06-29 | 2025-05-12 |

| Launch MSRP | None listed | $589 |

| Part Number | 100-000001971 | 100-000001559 |

All other specifications are identical: 4 nm TSMC process, 16,630 million transistors, 2x 70.6 mm² die size, 80 KB L1 per core, 1 MB L2 per core, DDR5 dual-channel memory, 89.6 GB/s bandwidth, ECC support, Gen 5 PCIe with 24 lanes, Radeon integrated graphics, AM5 socket, locked multiplier, and Server/Workstation market segment.

Head-to-Head Benchmarks

The largest single win for the EPYC 4565P comes in passmark_integer_math, where it scores 250683 against 182312, a 27.3% advantage. This is a pure throughput test, and the EPYC's 16 cores versus 12 provide a straightforward edge. Data encryption follows closely, with the EPYC at 48268 versus 33754, a 30.1% lead. Encryption workloads scale well with core count and benefit from the EPYC's higher base clock.

Data compression shows a 20.5% EPYC lead, 860786 versus 684470. This test involves both memory bandwidth and integer throughput, and the EPYC's higher clocks plus extra cores carry it. Floating point math goes to the EPYC by 20.1%, 152003 versus 121509. Extended instructions show a 14.7% EPYC lead, 64345 versus 54903. Random string sorting is 11.5% in the EPYC's favor, 81318 versus 71928. The multithread test, which is the most representative of general heavy parallel work, shows the EPYC at 63474 versus 54866, a 13.6% lead.

The Ryzen 9 PRO 9955's wins are narrower in scope but dramatic in one case. The prime number test is a 56.8% blowout, 461 versus 294. Physics is a 12% win, 3332 versus 2976. Looking at the percentile rankings, the Ryzen sits at the 97th percentile of all CPUs, while the EPYC sits at the 96th. The Ryzen's average benchmark score is 110612, while the EPYC's is 95764. This is a notable inversion: the EPYC wins most head-to-head tests, yet the Ryzen has a higher average score across all benchmarks in the database. That is likely because the Ryzen's prime number and physics wins are so lopsided that they pull its average up, and because the EPYC has several Cinebench results that may drag its average down.

The nearest rivals in the database put both chips in similar company. The Ryzen's nearest competitor is the Intel Xeon w7-2595X, which scores 108663 against the Ryzen's 110612, a 1.8% delta. The EPYC's nearest rival is the AMD EPYC 9175F, scoring 95615 against 95764, a 0.2% delta. Both chips are positioned among high-end workstation and server parts, with the Ryzen slightly ahead in aggregate benchmark score despite losing the head-to-head.

The Verdict

Choose the AMD EPYC 4565P if your workload is parallel, throughput-bound, or involves encryption, compression, integer math, or floating point math. It wins 9 of 11 head-to-head benchmarks, has 16 cores versus 12, clocks higher at 4.30 GHz base and 5.70 GHz boost, and carries a 170 W TDP to sustain that performance. Its single-thread score is also slightly higher, so there is no scenario in the data where the Ryzen is faster except for the two specific tests it wins.

Choose the AMD Ryzen 9 PRO 9955 if your workload is dominated by prime number finding or physics simulation. The 56.8% lead in prime numbers and the 12% lead in physics are significant, and the chip's lower 120 W TDP may make it easier to cool in a workstation chassis. Its 97th percentile ranking and higher average benchmark score (110612 versus 95764) also indicate that it excels in a broader set of benchmarks beyond the head-to-head list, even if those wins are not all captured in the direct comparison.

For a general-purpose server or workstation CPU, the EPYC 4565P is the safer recommendation based on the head-to-head data. For a specialized compute node focused on integer recursion or physics, the Ryzen 9 PRO 9955 has a clear, data-backed edge. Both are active production parts on the same AM5 socket, so platform compatibility is not a differentiator. The EPYC's earlier release date and listed launch MSRP of $589 give it a practical advantage in availability, while the Ryzen's later release suggests it may be the newer design despite sharing the same silicon.