AMD EPYC 9135 vs Intel Core Ultra 9 290K Plus Comparison

The Intel Core Ultra 9 290K Plus and AMD EPYC 9135 occupy different corners of the processor market, yet their benchmark scores place them within 1.2% of each other in average performance. The data shows a desktop enthusiast part trading blows with a server processor, with each holding decisive advantages in specific workloads. Despite the EPYC’s higher thread count and the Core Ultra’s higher clock speeds, the overall average benchmark scores are remarkably close: 84,003 for the Intel part versus 82,980 for the AMD part. Both sit at the 96th percentile among all CPUs, indicating top-tier standing, but their paths to that standing diverge sharply.

Head-to-Head Benchmarks

The Intel Core Ultra 9 290K Plus wins 12 of the 17 head-to-head benchmark comparisons, but the margins are not uniform. The largest Intel victory comes in PassMark’s find prime numbers test, where it scores 503 against the EPYC’s 292, a 72.3% advantage. Floating point math follows closely, with Intel scoring 214,760 versus 126,679, a 69.5% lead. These are not marginal differences; they represent a fundamental advantage in specific computational patterns.

Single-threaded performance heavily favors Intel. In PassMark single thread, the Core Ultra scores 4,823 versus the EPYC’s 3,672, a 31.3% gap. The same margin appears in data encryption, where Intel scores 52,563 against 40,295, a 30.4% lead. Across all Cinebench tests—R15, R20, and R23—Intel wins both single-core and multi-core by a consistent 5.3%. For instance, Cinebench R23 multi-core shows Intel at 51,731 versus AMD at 49,136, while single-core shows 7,303 against 6,936.

The AMD EPYC 9135 wins the remaining five benchmarks, and its victories are equally decisive in their own domains. The largest is PassMark physics, where the EPYC scores 5,477 against Intel’s 3,315, a 39.5% advantage. Integer math goes to AMD at 202,962 versus 166,194, an 18.1% lead. Data compression favors AMD at 739,277 against 698,346, a 5.5% margin. Random string sorting shows AMD ahead at 90,064 versus 79,744, an 11.5% gap. Extended instructions also favor AMD, scoring 55,822 against 52,338, a 6.2% edge.

The overall multithread score slightly favors Intel, with the Core Ultra at 60,860 versus the EPYC’s 57,170, a 6.5% difference. This is notable because the EPYC has 32 threads against Intel’s 24, yet Intel still leads in the aggregate multithread metric.

Where Each One Wins

The Intel Core Ultra 9 290K Plus is the clear choice for workloads that depend on single-core speed and specific math operations. Its 31.3% lead in PassMark single thread and 5.3% across all Cinebench single-core tests indicates superior responsiveness in lightly threaded applications. The 72.3% advantage in prime number finding and 69.5% lead in floating point math suggest strong performance in scientific computing, financial modeling, or any task heavily reliant on floating-point calculations. The 30.4% edge in data encryption points to advantages in security-related workloads and cryptographic operations.

For multi-core rendering, Intel also holds a consistent 5.3% lead across all three Cinebench versions. The Core Ultra’s 24 cores, despite lacking hyper-threading, outperform the EPYC’s 16 cores with 32 threads in these specific rendering tasks. This suggests that Intel’s core architecture is more efficient per thread in Cinebench workloads.

The AMD EPYC 9135 wins where its server-oriented design shines. The 39.5% advantage in PassMark physics is substantial, indicating superior performance in physics simulations common in engineering and scientific applications. The 18.1% lead in integer math makes it better suited for general-purpose computing tasks, database operations, and integer-heavy algorithms. Data compression favor AMD by 5.5%, making it a better choice for file compression, backup systems, and data archiving. Random string sorting, with an 11.5% advantage, points to strengths in sorting algorithms and text processing. Extended instructions favor AMD by 6.2%, suggesting better performance in workloads using SIMD or specialized instruction sets.

The EPYC’s higher thread count (32 versus 24) does not translate into universal multi-threaded dominance, but it does help in physics and integer workloads where thread-level parallelism matters more than raw clock speed.

Architecture Differences

The two processors come from fundamentally different design philosophies. The Intel Core Ultra 9 290K Plus uses the Arrow Lake Refresh architecture, built on a 3 nm process at TSMC. It packs 24 cores and 24 threads—no hyper-threading—with a base clock of 3.70 GHz and a boost clock of 5.80 GHz. The EPYC 9135 uses the Zen 5 architecture, codenamed Turin, on a 4 nm process, also at TSMC. It has 16 cores and 32 threads, with a base clock of 3.65 GHz and a boost clock of 4.30 GHz.

Transistor counts are similar: Intel lists 17,800 million transistors on a 243 mm² die, while AMD lists 16,630 million across a dual-die design of 2x 70.6 mm². The cache structures differ notably. Intel provides 192 KB of L1 per core, 3 MB of L2 per core, and 36 MB of shared L3. AMD provides 80 KB of L1 per core, 1 MB of L2 per core, and a larger 64 MB of shared L3. The larger L3 cache on the EPYC likely contributes to its advantage in data compression and integer math, where larger working sets can reside on-chip.

Memory architecture is a major differentiator. Both support DDR5, but Intel uses a dual-channel memory bus with 115.2 GB/s bandwidth, while AMD uses a twelve-channel bus with 576.0 GB/s bandwidth—exactly five times the theoretical memory bandwidth. Both support ECC memory. This massive bandwidth advantage for AMD is central to its server positioning, though it does not translate into wins in all memory-sensitive benchmarks.

PCIe connectivity also differs drastically. Intel provides Gen 5 with 20 lanes (CPU only), while AMD provides Gen 5 with 128 lanes (CPU only). This makes the EPYC far more suitable for systems with many GPUs, NVMe drives, or network cards. The Intel part includes integrated graphics (Arc Xe-LPG Graphics 64EU), while the EPYC has no integrated graphics at all. Power envelopes differ, with Intel rated at 125 W TDP and AMD at 200 W TDP, a 75 W difference that reflects the EPYC’s server orientation. Intel’s multiplier is unlocked for overclocking; AMD’s is locked.

The Verdict

From the data, the Intel Core Ultra 9 290K Plus is the better processor for single-threaded performance, floating-point math, encryption, and Cinebench rendering workloads. Its 31.3% lead in single-thread performance and 69.5% lead in floating point make it the superior choice for desktop users running engineering software, financial models, or any application where per-core speed matters. Its lower TDP (125 W versus 200 W) and integrated graphics add practical advantages for a desktop build.

The AMD EPYC 9135 is the better processor for integer-heavy workloads, physics simulations, data compression, and sorting tasks. Its 39.5% lead in physics and 18.1% lead in integer math make it a stronger candidate for scientific computing, database servers, and data processing pipelines. The twelve-channel memory bus providing 576.0 GB/s bandwidth and 128 PCIe Gen 5 lanes make it the clear choice for server environments requiring massive I/O and memory throughput.

For a desktop user who prioritizes raw single-core speed and rendering performance, the Intel Core Ultra 9 290K Plus is the data-backed pick. For a server or workstation user who needs integer throughput, physics simulation, and extensive I/O capabilities, the AMD EPYC 9135 is the justified selection. The EPYC’s launch MSRP is $1214.

FAQ

Q: Which processor has a higher average benchmark score?

A: The Intel Core Ultra 9 290K Plus has an average benchmark score of 84,003, while the AMD EPYC 9135 has 82,980, a difference of 1.2%.

Q: How much faster is Intel in single-threaded performance?

A: In PassMark single-thread testing, Intel scores 4,823 versus AMD’s 3,672, representing a 31.3% advantage.

Q: In which benchmark does the AMD EPYC 9135 have its largest win?

A: The EPYC’s largest win is in PassMark physics, where it scores 5,477 against Intel’s 3,315, a 39.5% lead.

Q: Does the EPYC’s higher thread count lead to better multi-threaded performance?

A: No. Despite having 32 threads versus Intel’s 24, the EPYC scores lower in PassMark multithread (57,170 versus 60,860, a 6.5% deficit) and in all three Cinebench multi-core tests.

Q: What are the memory bandwidth specifications for each processor?

A: The Intel Core Ultra 9 290K Plus has a dual-channel memory bus with 115.2 GB/s bandwidth. The AMD EPYC 9135 has a twelve-channel memory bus with 576.0 GB/s bandwidth.

Q: Which processor has integrated graphics?

A: The Intel Core Ultra 9 290K Plus includes Arc Xe-LPG Graphics 64EU. The AMD EPYC 9135 has no integrated graphics.

Specification Differences

| Specification | Intel Core Ultra 9 290K Plus | AMD EPYC 9135 |

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

| Cores | 24 | 16 |

| Threads | 24 | 32 |

| Base Clock | 3.70 GHz | 3.65 GHz |

| Boost Clock | 5.80 GHz | 4.30 GHz |

| TDP | 125 W | 200 W |

| Socket | Intel Socket 1851 | AMD Socket SP5 |

| Architecture | Arrow Lake Refresh | Zen 5 |

| Codename | Arrow Lake Refresh | Turin |

| Process Node | 3 nm | 4 nm |

| Transistors | 17,800 million | 16,630 million |

| Die Size | 243 mm² | 2x 70.6 mm² |

| L1 Cache | 192 KB (per core) | 80 KB (per core) |

| L2 Cache | 3 MB (per core) | 1 MB (per core) |

| L3 Cache | 36 MB (shared) | 64 MB (shared) |

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

| Memory Bandwidth | 115.2 GB/s | 576.0 GB/s |

| PCIe | Gen 5, 20 Lanes (CPU only) | Gen 5, 128 Lanes (CPU only) |

| Integrated Graphics | Arc Xe-LPG Graphics 64EU | N/A |

| Multiplier Unlocked | Yes | No |

| Market Segment | Desktop | Server/Workstation |