AMD EPYC 9575F vs Intel Xeon 6747P Comparison

FAQ

Q: How much faster is the AMD EPYC 9575F in multi-threaded workloads?

A: The EPYC 9575F leads in every recorded multi-core test. In Cinebench R23 multi-core it scores 127739 versus 86432 for the Intel Xeon 6747P, a 47.8% advantage. The same delta appears in Cinebench R15 and R20 multi-core, both at 47.8% in favor of AMD.

Q: Which processor has the higher single-thread performance?

A: The EPYC 9575F wins single-thread tests by 29%. Its PassMark single-thread score is 4173, while the Xeon 6747P records 3236. The AMD chip also boosts to 5.00 GHz against 3.90 GHz for Intel, which explains the gap.

Q: How do the two compare in integer and floating-point math?

A: The AMD part dominates both. In PassMark integer math, the EPYC 9575F scores 891817 versus 468518 for the Xeon, a 90.3% lead. In floating-point math, AMD scores 526003 against 365904, a 43.8% margin. These are large, consistent wins.

Q: What about memory bandwidth and channels?

A: The AMD EPYC 9575F uses a twelve-channel memory bus with 576.0 GB/s bandwidth. The Intel Xeon 6747P uses eight channels with 409.6 GB/s. Both support DDR5 and ECC memory, but AMD has a raw bandwidth advantage of roughly 40% based on the recorded figures.

Q: Are both processors in the same market segment?

A: Yes, both are listed as Server/Workstation parts. Both are active production chips, use no integrated graphics, and have locked multipliers. Their sockets differ: AMD uses Socket SP5, Intel uses Socket 4710.

Q: Does the Intel part win any benchmark at all?

A: No. In the head-to-head data, the EPYC 9575F wins all 14 recorded benchmarks. The Xeon 6747P has zero wins. Even in the closest test, PassMark find prime numbers, AMD still leads by 5.6% (1215 versus 1151).

Architecture Differences

The AMD EPYC 9575F is built on the Zen 5 architecture, codenamed Turin, and belongs to the EPYC 9005 series. It uses a 4 nm process at TSMC, with a reported transistor count of 66,520 million spread across 8 chiplets of 70.6 mm² each. The Intel Xeon 6747P uses the Granite Rapids architecture from the Xeon 6 family, fabricated on Intel's 5 nm process, with a die size of 2x 598 mm². Intel does not report a transistor count in the database.

Core and cache layouts differ substantially. The AMD part has 64 cores and 128 threads, with 80 KB of L1 per core, 1 MB of L2 per core, and 256 MB of shared L3. The Intel part has 48 cores and 96 threads, with 112 KB of L1 per core, 2 MB of L2 per core, and 288 MB of shared L3. Although Intel has more L3 in absolute terms, AMD's larger core count and higher clocks change how that cache is used.

Clock speeds are a major differentiator. The EPYC 9575F runs at a 3.30 GHz base and boosts to 5.00 GHz. The Xeon 6747P sits at 2.70 GHz base and 3.90 GHz boost. AMD's boost advantage is 1.10 GHz, which shows up directly in single-thread and lightly threaded workloads. The TDP also differs: AMD is rated at 400 W, Intel at 330 W.

PCIe lane counts and memory channels are architecturally different. AMD provides Gen 5 with 128 lanes (CPU only), while Intel provides Gen 5 with 88 lanes (CPU only). AMD's twelve-channel memory bus delivers 576.0 GB/s, versus Intel's eight-channel 409.6 GB/s. Both support DDR5 and ECC. Neither has integrated graphics.

The release timing also differs. The EPYC 9575F launched on 2024-10-09, while the Xeon 6747P arrived later on 2025-02-23. Both are active production parts. The AMD part carries part number 100-000001554, and the Intel part is SRVEZ. Neither processor has an unlocked multiplier.

Head-to-Head Benchmarks

The recorded benchmark data shows a clean sweep for the AMD EPYC 9575F. Every single test, from Cinebench to PassMark workloads, lands in AMD's favor. The margins vary from modest to overwhelming, so it is worth breaking them down by workload type.

Starting with Cinebench, the multi-core results are identical in percentage terms. In Cinebench R15 multi-core, AMD scores 12876 versus 8712 for Intel, a 47.8% lead. The same 47.8% delta appears in R20 (53650 versus 36301) and R23 (127739 versus 86432). This consistency suggests the advantage is structural, tied to core count, clock speed, and memory bandwidth, rather than workload-specific quirks.

The PassMark integer math test is the single largest win for AMD. The EPYC 9575F scores 891817, while the Xeon 6747P scores 468518, a 90.3% delta. That is nearly double the throughput. Random string sorting also shows a huge gap: AMD scores 348500 versus 180382, a 93.2% lead. These two tests indicate that AMD's Zen 5 cores are dramatically more efficient at these specific processing patterns.

Data encryption and compression follow a similar pattern. In encryption, AMD scores 162818 against 90789, a 79.3% advantage. In data compression, AMD scores 2773634 versus 1833378, a 51.3% lead. These are common server workloads, and the margins here are decisive.

Physics and floating-point tests also favor AMD, but with smaller margins. PassMark physics shows AMD at 20652 versus 13398, a 54.1% lead. Floating-point math shows AMD at 526003 versus 365904, a 43.8% lead. Extended instructions are closer at 38.5% (197484 versus 142557). The smallest margin is in find prime numbers, where AMD scores 1215 versus 1151, a 5.6% edge. Even there, AMD does not lose.

Single-thread results are less dramatic but still clear. PassMark single-thread shows AMD at 4173 versus 3236, a 29% lead. This aligns with the boost clock difference of 5.00 GHz versus 3.90 GHz. The multithread test shows AMD at 147998 versus 101685, a 45.5% lead, which is consistent with the Cinebench multi-core results.

The overall picture is one of consistent dominance. Across 14 recorded benchmarks, the AMD EPYC 9575F wins every one. The Intel Xeon 6747P records zero wins. The average benchmark score reflects this: AMD sits at 311774, while Intel sits at 238263.

Specification Differences

The table below covers only the fields where the two processors differ.

| Specification | AMD EPYC 9575F | Intel Xeon 6747P |

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

| Cores | 64 | 48 |

| Threads | 128 | 96 |

| Base clock | 3.30 GHz | 2.70 GHz |

| Boost clock | 5.00 GHz | 3.90 GHz |

| TDP | 400 W | 330 W |

| Socket | AMD Socket SP5 | Intel Socket 4710 |

| Architecture | Zen 5 | Granite Rapids |

| Codename | Turin | Granite Rapids |

| Generation | EPYC (Zen 5, Turin) | Xeon 6 (Granite Rapids-SP) |

| Process node | 4 nm | 5 nm |

| Foundry | TSMC | Intel |

| Die size | 8x 70.6 mm² | 2x 598 mm² |

| Transistors | 66,520 million | Not reported |

| L1 cache | 80 KB per core | 112 KB per core |

| L2 cache | 1 MB per core | 2 MB per core |

| L3 cache | 256 MB shared | 288 MB shared |

| Memory bus | Twelve-channel | Eight-channel |

| Memory bandwidth | 576.0 GB/s | 409.6 GB/s |

| PCIe lanes | Gen 5, 128 lanes | Gen 5, 88 lanes |

| Release date | 2024-10-09 | 2025-02-23 |

| Launch MSRP | $11791 | $6497 |

| Part number | 100-000001554 | SRVEZ |

Both processors share DDR5 memory support, ECC memory, no integrated graphics, the Server/Workstation market segment, active production status, and locked multipliers.

The Verdict

The data is unambiguous. The AMD EPYC 9575F outperforms the Intel Xeon 6747P in every recorded benchmark. The smallest margin is 5.6% in find prime numbers, and the largest is 93.2% in random string sorting. In compute-heavy multi-core workloads, the AMD part is consistently 47.8% ahead in Cinebench and 45.5% ahead in PassMark multithread. The single-thread advantage is 29%, which matters for workloads that cannot scale across cores.

The Intel part does have a lower TDP at 330 W versus 400 W, and it has a lower launch MSRP at $6497 versus $11791. It also offers more L3 cache per core and a larger L2 cache per core. However, the benchmark results do not translate those specifications into performance wins. The Xeon 6747P is a capable server processor, but in this direct comparison, it is behind on every measurable workload.

The EPYC 9575F also carries a higher average benchmark score of 311774 against 238263 for Intel. Both parts sit in the 99th percentile of all CPUs in the database, meaning neither is a weak choice. But the AMD part is clearly the higher performer in this matchup.

Where Each One Wins

The AMD EPYC 9575F wins in every use case that relies on raw compute throughput. If the workload is multi-threaded rendering, database compression, encryption, integer math, floating-point math, or physics simulation, the data says AMD is faster by margins ranging from 38.5% to 93.2%. The Cinebench results, which are commonly used as a proxy for 3D rendering and content creation, show a 47.8% advantage across all three versions. For server environments that run mixed workloads, the EPYC 9575F is the safer bet.

The Intel Xeon 6747P does not have a single benchmark win, but it has structural advantages that may matter in specific deployments. Its TDP is 70 W lower than AMD's, which can reduce cooling requirements and power delivery complexity. Its 288 MB of L3 cache is larger than AMD's 256 MB, and its L2 cache is double per core at 2 MB versus 1 MB. For workloads that are cache-sensitive but not captured in the recorded benchmarks, the Intel part could behave differently. Its launch MSRP is also lower at $6497 versus $11791, though pricing discussion is limited to that single stated figure.

For single-threaded tasks, the AMD part wins by 29%, so even lightly threaded applications favor AMD. For memory-bandwidth-bound workloads, AMD's twelve-channel bus at 576.0 GB/s is substantially ahead of Intel's eight-channel 409.6 GB/s. The PCIe lane count also favors AMD at 128 lanes versus 88 lanes, which matters for systems with many GPUs or NVMe drives.

If the priority is maximum performance per socket, the AMD EPYC 9575F is the clear choice from this data. If the priority is lower power draw, lower launch MSRP, or larger per-core caches, the Intel Xeon 6747P has those specific attributes. But on raw benchmark scores, the EPYC 9575F wins all 14 recorded tests.