AMD EPYC 9454P vs Intel Core Ultra 7 165U Comparison

The AMD EPYC 9454P and Intel Core Ultra 7 165U represent opposite extremes of the processor spectrum: a 48-core server behemoth against a 12-core mobile chip. The benchmark data is unequivocal in its head-to-head results, yet the two parts occupy such different design philosophies that their specification sheets tell a more nuanced story than their raw scores alone. The EPYC 9454P’s 74th percentile ranking and the Core Ultra 7 165U’s identical 74th percentile ranking suggest they are statistically comparable in overall average benchmark scores, but the manner in which they achieve those numbers could not be more different.

Head-to-Head Benchmarks

The AMD EPYC 9454P dominates every single benchmark in the head-to-head comparison, winning all 8 tests. The most dramatic disparity appears in Cinebench R23 multi-core, where the EPYC scores 81,492 against the Core Ultra 7’s 10,226, a staggering 696.9% advantage. This is not a marginal lead; it is an order-of-magnitude difference that reflects the EPYC’s 48 cores and 96 threads being leveraged across all available execution units. The Cinebench R20 multi-core test shows a similarly lopsided result, with the EPYC scoring 34,226 versus 5,943, a 475.9% delta.

Single-core results narrow the gap considerably, though the EPYC still wins every time. In Geekbench single-core, the EPYC scores 2,011 against the Core Ultra 7’s 1,450, a 38.7% difference. This is the closest margin in the entire comparison, suggesting that Intel’s Meteor Lake architecture has respectable per-thread performance even when facing AMD’s Zen 4. The Cinebench R15 single-core test shows a wider 401.7% gap (1,159 vs 231), but this older benchmark appears to favor the EPYC’s higher base clock of 2.75 GHz over the Ultra 7’s 1,700 MHz base frequency.

Multi-core Geekbench tells a different story: the EPYC’s 19,941 versus 7,329 represents a 172.1% lead. While substantial, this is notably smaller than the Cinebench multi-core deltas, indicating that Geekbench’s workload scaling does not fully harness the EPYC’s massive thread count as efficiently as Cinebench does. The Cinebench R23 single-core result of 11,504 versus 1,662 (592.2% delta) is puzzling given the closer Geekbench single-core margin, but it highlights how different benchmark suites stress different aspects of the memory hierarchy and instruction pipeline.

Architecture Differences

The foundational divergence lies in their process nodes and physical design. The EPYC 9454P uses TSMC’s 5 nm process with a die size of 8x 72 mm², packing 52,560 million transistors. Intel’s Core Ultra 7 165U uses Intel’s 7 nm process, with no transistor count or die size listed in the data. This node advantage likely contributes to the EPYC’s ability to fit 48 Zen 4 cores while maintaining a 290 W TDP, compared to the Ultra 7’s 15 W TDP.

Cache hierarchies could not be more different. The EPYC provides 64 KB L1 per core, 1 MB L2 per core, and a massive 256 MB shared L3 cache. The Ultra 7 offers 112 KB L1 per core, 2 MB L2 per core, but only 12 MB shared L3. The EPYC’s L3 cache is over 21 times larger, which directly supports its server workload profile where working sets frequently exceed smaller cache capacities. Interestingly, the Ultra 7 has larger per-core L1 and L2 caches, suggesting Intel optimized for latency-sensitive single-thread workloads, while AMD prioritized aggregate capacity.

Memory subsystems reinforce the separation. The EPYC uses twelve-channel DDR5 with 460.8 GB/s bandwidth and supports ECC memory. The Ultra 7 uses dual-channel DDR5 with 89.6 GB/s bandwidth and no ECC support. The 5.1x bandwidth advantage for the EPYC is critical for multi-threaded math and database workloads where memory throughput becomes the bottleneck. PCIe connectivity also differs: the EPYC offers Gen 5 with 128 lanes (CPU only), while the Ultra 7 provides Gen 4 with 12 lanes. The Ultra 7 integrates Arc Xe-LPG 64EU graphics, whereas the EPYC has no integrated graphics at all, reflecting its server role where discrete GPUs are assumed.

Where Each One Wins

The EPYC 9454P wins in every measured category, but the practical implications depend on workload type. For heavily parallel tasks like Cinebench multi-core rendering, the EPYC’s 696.9% lead over the Ultra 7 is the defining characteristic. The R20 multi-core delta of 475.9% and R15 multi-core delta of 411.8% confirm that scaling across 96 threads is where this chip excels. The EPYC also dominates memory-bandwidth-sensitive applications, given its twelve-channel configuration and 460.8 GB/s throughput. Its 128 PCIe Gen 5 lanes enable massive I/O expansion, making it suitable for storage arrays, GPU clusters, or network accelerators.

The Core Ultra 7 165U’s wins are more contextual than direct. Its 15 W TDP versus the EPYC’s 290 W means it can operate in fanless or passively cooled mobile devices, a physical impossibility for the EPYC. Its integrated Arc Xe-LPG 64EU graphics provides display output and basic acceleration without a discrete GPU, which the EPYC cannot do at all. The Ultra 7’s 4.90 GHz boost clock exceeds the EPYC’s 3.80 GHz peak, and its larger per-core L1 and L2 caches (112 KB vs 64 KB L1, 2 MB vs 1 MB L2) suggest better latency characteristics for bursty, low-thread-count workloads. The data shows the Ultra 7 achieves a 74th percentile ranking despite its modest 12 cores, implying that its efficiency and single-thread performance per watt are competitive.

FAQ

Q: Which processor has higher multi-core performance?

A: The AMD EPYC 9454P wins all multi-core benchmarks. In Cinebench R23 multi-core, it scores 81,492 versus the Intel Core Ultra 7 165U’s 10,226, a 696.9% advantage. Geekbench multi-core shows a 172.1% lead (19,941 vs 7,329).

Q: How do single-core scores compare between the two?

A: The EPYC 9454P wins every single-core test. The closest margin is Geekbench single-core at 38.7% (2,011 vs 1,450). The largest single-core gap is Cinebench R23 at 592.2% (11,504 vs 1,662).

Q: What memory bandwidth do these processors support?

A: The EPYC 9454P supports twelve-channel DDR5 with 460.8 GB/s bandwidth. The Core Ultra 7 165U supports dual-channel DDR5 with 89.6 GB/s bandwidth, making the EPYC approximately 5.1 times faster in memory throughput.

Q: Do both processors have integrated graphics?

A: No. The Intel Core Ultra 7 165U includes Arc Xe-LPG 64EU integrated graphics. The AMD EPYC 9454P has no integrated graphics, requiring a separate GPU for display output.

Q: What are the power requirements for each chip?

A: The EPYC 9454P has a TDP of 290 W, while the Core Ultra 7 165U has a TDP of 15 W. This 275 W difference makes the Ultra 7 suitable for mobile devices, whereas the EPYC requires substantial cooling and power delivery.

Q: How do their cache sizes differ?

A: The EPYC 9454P has 64 KB L1 per core, 1 MB L2 per core, and 256 MB shared L3. The Core Ultra 7 165U has 112 KB L1 per core, 2 MB L2 per core, and 12 MB shared L3. The EPYC’s L3 is over 21 times larger.

The Verdict

The data supports selecting the AMD EPYC 9454P for any workload where raw compute throughput is the primary requirement. Its 8-0 benchmark sweep, with wins ranging from 38.7% to 696.9%, demonstrates overwhelming superiority in both single-thread and multi-thread scenarios. The EPYC’s 96 threads, 256 MB L3, and 460.8 GB/s memory bandwidth make it the obvious choice for server virtualization, scientific simulation, database serving, or content rendering pipelines. Its 74th percentile ranking among all CPUs, despite being a server part, indicates it competes favorably even against newer consumer chips.

The Intel Core Ultra 7 165U is the appropriate selection when physical constraints dominate. Its 15 W TDP and integrated Arc graphics enable ultraportable laptop designs where the EPYC’s 290 W requirement is impossible. The Ultra 7 also wins on boost clock (4.90 GHz vs 3.80 GHz) and per-core cache sizes, though these advantages do not translate into benchmark wins. For mobile productivity, light development, or media consumption, the Ultra 7 provides adequate performance (74th percentile) while fitting in a slim chassis. The data does not show any benchmark where the Ultra 7 beats the EPYC, so the choice is purely about deployment context rather than performance superiority.

Specification Differences

| Specification | AMD EPYC 9454P | Intel Core Ultra 7 165U |

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

| Cores | 48 | 12 |

| Threads | 96 | 14 |

| Base Clock | 2.75 GHz | 1700.00 MHz |

| Boost Clock | 3.80 GHz | 4.90 GHz |

| TDP | 290 W | 15 W |

| Socket | AMD Socket SP5 | Intel BGA 2049 |

| Architecture | Zen 4 | Meteor Lake |

| Codename | Genoa | Meteor Lake |

| Process Node | 5 nm | 7 nm |

| Foundry | TSMC | Intel |

| Transistors | 52,560 million | Not listed |

| Die Size | 8x 72 mm² | Not listed |

| L1 Cache | 64 KB (per core) | 112 KB (per core) |

| L2 Cache | 1 MB (per core) | 2 MB (per core) |

| L3 Cache | 256 MB (shared) | 12 MB (shared) |

| Memory Support | DDR5 | DDR5 (depends on motherboard) |

| Memory Bus | Twelve-channel | Dual-channel |

| Memory Bandwidth | 460.8 GB/s | 89.6 GB/s |

| ECC Memory | Yes | No |

| PCIe | Gen 5, 128 Lanes (CPU only) | Gen 4, 12 Lanes (CPU only) |

| Integrated Graphics | None | Arc Xe-LPG 64EU |

| Market Segment | Server/Workstation | Mobile |

| Release Date | 2022-11-09 | 2023-12-13 |

| Launch MSRP | $4598 | $448 |

| Part Number | 100-100000873 | SRN6D |