AMD EPYC 9355P vs Intel Xeon 6745P Comparison

Where Each One Wins

The AMD EPYC 9355P is the dominant part in this matchup, winning 14 of the 16 head-to-head benchmarks. Its wins are not narrow either; the average delta across its victories is substantial, with several workloads showing double-digit percentage leads. The Intel Xeon 6745P secures only two wins, and both are marginal, suggesting a very specific niche rather than a general-purpose advantage.

The EPYC 9355P’s territory is multi-threaded throughput and integer-heavy work. In the Cinebench suite (R15, R20, R23), it wins both multi-core and single-core tests by a uniform 15.5% margin. That consistency indicates a fundamental per-thread performance advantage rather than a scaling quirk. The PassMark integer math test shows a 22.3% lead, and random string sorting is 32.3% ahead. The physics test is the most lopsided result: the AMD part scores 13,515 versus 6,144, a 120% advantage, which points to a massive difference in how the two architectures handle that particular workload’s instruction pattern.

The Xeon 6745P’s two wins are in floating-point math and extended instructions. The floating-point win is 4% (267,438 versus 256,635), and the extended instructions win is a razor-thin 0.6% (108,326 versus 107,622). These are not workloads where the Intel part dominates; they are cases where it edges ahead. For most buyers, these margins are within noise for a single benchmark run, though they do hint that the Granite Rapids core has a slight advantage in certain math-heavy, vectorized code paths.

The data compression test is closer than most, with the AMD part winning by 5.7%. Encryption is a clear AMD win at 21.4%. In multi-threaded PassMark, the AMD part leads by 14.7%, and in single-thread it leads by 8.6%. The overall picture is unambiguous: the EPYC 9355P is the better processor for nearly every workload category tested, with the Xeon 6745P only showing a pulse in floating-point and SIMD-style extended instruction tasks.

FAQ

Q: Which CPU has a higher average benchmark score?

A: The AMD EPYC 9355P has an average benchmark score of 160,358, while the Intel Xeon 6745P scores 154,858. The AMD part is 3.6% ahead based on the nearestRivals deltaPct.

Q: How does the AMD EPYC 9355P compare to the Intel Xeon 6745P in multi-core rendering?

A: In Cinebench R23 multi-core, the AMD EPYC 9355P scores 82,666 versus 71,578 for the Intel Xeon 6745P, a 15.5% advantage. The same 15.5% delta appears in Cinebench R15 and R20 multi-core tests.

Q: Does the Intel Xeon 6745P win any benchmarks?

A: Yes, it wins two tests: PassMark floating-point math (267,438 versus 256,635, a 4% lead) and PassMark extended instructions (108,326 versus 107,622, a 0.6% lead).

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

A: The largest delta is in PassMark physics, where the AMD EPYC 9355P scores 13,515 versus 6,144 for the Intel Xeon 6745P, giving AMD a 120% lead.

Q: Which CPU has better single-thread performance?

A: The AMD EPYC 9355P leads in single-thread tests. In Cinebench R23 single-core, it scores 11,670 versus 4,244 for the Intel part (the Xeon R23 single-core score is not listed, so the R20 comparison applies). In PassMark single-thread, AMD leads 3,747 versus 3,450, an 8.6% margin.

Q: Are both CPUs in the same performance percentile?

A: Yes, both the AMD EPYC 9355P and the Intel Xeon 6745P sit at the 98th percentile versus all CPUs.

Head-to-Head Benchmarks

The Cinebench suite delivers the most consistent story. Across R15, R20, and R23, in both single-core and multi-core, the AMD EPYC 9355P beats the Intel Xeon 6745P by exactly 15.5%. For multi-core, that translates to 8,332 versus 7,214 in R15, 34,719 versus 30,062 in R20, and 82,666 versus 71,578 in R23. The single-core deltas are equally uniform: 1,176 versus 1,018 in R15, 4,901 versus 4,244 in R20. This uniformity suggests the AMD Zen 5 core is simply faster per clock and at the same time runs at a higher boost clock (4.40 GHz versus 4.30 GHz), producing a predictable lead.

The PassMark integer math test is another decisive AMD win. The EPYC 9355P scores 412,067, while the Xeon 6745P manages 336,926, a 22.3% gap. Random string sorting follows a similar pattern: 176,697 versus 133,528, a 32.3% lead. The data encryption test shows a 21.4% advantage for AMD (80,961 versus 66,665). Data compression is closer but still favors AMD: 1,429,976 versus 1,352,801, a 5.7% win.

The physics test is the outlier in its magnitude. The AMD part scores 13,515; the Intel part scores 6,144. That is a 120% delta, meaning the AMD CPU is more than twice as fast in this specific workload. The PassMark multi-thread test shows a 14.7% AMD lead (96,603 versus 84,210), and single-thread is an 8.6% AMD lead (3,747 versus 3,450).

The Intel Xeon 6745P’s wins are real but small. In floating-point math, it scores 267,438 versus 256,635, a 4% advantage. In extended instructions, it scores 108,326 versus 107,622, a 0.6% edge. Neither result flips the overall picture; the AMD part wins the remaining 14 benchmarks, including all five Cinebench tests and the majority of PassMark workloads.

Specification Differences

The two CPUs share the same core and thread count: 32 cores and 64 threads. Both support DDR5 memory and ECC, and neither has integrated graphics. Both are locked (multiplier unlocked: false) and target the server/workstation segment.

The AMD EPYC 9355P has a base clock of 3.55 GHz and a boost clock of 4.40 GHz. The Intel Xeon 6745P runs lower at 3.10 GHz base and 4.30 GHz boost. TDP differs: AMD is rated at 280 watts, Intel at 300 watts. The AMD part uses AMD Socket SP5; the Intel part uses Intel Socket 4710.

Memory channels and bandwidth are a key differentiator. The AMD EPYC 9355P has a twelve-channel memory bus with 576.0 GB/s bandwidth. The Intel Xeon 6745P has an eight-channel bus with 409.6 GB/s bandwidth. That is a 40.6% bandwidth advantage for AMD, which directly supports its leads in memory-sensitive workloads like compression and random string sorting.

PCIe lane counts also differ. The AMD part provides 128 Gen 5 lanes (CPU only), while the Intel part provides 88 Gen 5 lanes (CPU only). The AMD part has a larger L3 cache at 256 MB shared, versus 336 MB for Intel. The L1 and L2 caches differ per core: AMD has 80 KB L1 and 1 MB L2 per core; Intel has 112 KB L1 and 2 MB L2 per core. The AMD part is built on a 4 nm process at TSMC, while Intel uses a 5 nm process at its own foundry.

Architecture Differences

The AMD EPYC 9355P is built on the Zen 5 architecture, codenamed Turin, and belongs to the EPYC 9005 series. It uses a chiplet design with eight dies, each measuring 70.6 mm², totaling 66,520 million transistors. The Intel Xeon 6745P uses the Granite Rapids architecture, part of the Xeon 6 (Granite Rapids-SP) generation. Its die size is listed as two dies of 598 mm² each; no transistor count is provided.

The process node difference is meaningful: AMD uses TSMC’s 4 nm process, while Intel uses its own 5 nm process. This gives AMD a density advantage, which helps explain how it fits 32 Zen 5 cores with 256 MB of L3 cache into a lower TDP (280 watts versus 300 watts) while running at higher clocks.

The cache hierarchies differ in philosophy. AMD uses a large shared L3 (256 MB) with smaller per-core L1 and L2 (80 KB and 1 MB). Intel uses even larger per-core L1 and L2 (112 KB and 2 MB) and a larger shared L3 (336 MB). The Intel part’s bigger caches do not translate into benchmark wins, except in the two narrow cases noted. The memory bandwidth gap (576.0 GB/s versus 409.6 GB/s) is likely a larger factor in the AMD part’s overall lead.

The release dates are also different. The AMD EPYC 9355P launched on 2024-10-09, while the Intel Xeon 6745P launched on 2025-02-23. The AMD part’s launch MSRP is $2998; the Intel part’s launch MSRP is $5250.

The Verdict

The data points to one conclusion: the AMD EPYC 9355P is the superior processor for almost all workloads measured. It wins 14 of 16 benchmarks, including all five Cinebench tests, with a uniform 15.5% margin. It leads in integer math by 22.3%, in encryption by 21.4%, in multithread by 14.7%, and in single-thread by 8.6%. The physics test is a blowout at 120% ahead.

The Intel Xeon 6745P’s two wins are in floating-point math (4% lead) and extended instructions (0.6% lead). These are real but narrow, and they do not compensate for the broader pattern of losses. The Intel part also has a higher TDP (300 watts versus 280 watts), lower boost clock (4.30 GHz versus 4.40 GHz), fewer memory channels (eight versus twelve), less memory bandwidth (409.6 GB/s versus 576.0 GB/s), and fewer PCIe lanes (88 versus 128). Its only hardware advantages are larger caches (L1, L2, and L3) and a lower base clock, which do not appear to help in the benchmark results.

For buyers choosing between these two, the AMD EPYC 9355P is the clear pick for general server workloads, rendering, encryption, compression, and integer-heavy tasks. The Intel Xeon 6745P is only worth considering if the workload is dominated by floating-point math or extended instruction sets, where it shows a small edge. Given the AMD part’s lower TDP, higher bandwidth, and better benchmark showing across the board, the EPYC 9355P is the data-driven recommendation in this comparison.