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

The AMD EPYC 9135 and Intel Core Ultra 9 290HX Plus occupy opposite ends of the computing spectrum: the EPYC is a 16-core, 32-thread server/workstation processor built on Zen 5 (Turin) with a 200 W TDP, while the Core Ultra 9 290HX Plus is a 24-core, 24-thread mobile chip from Arrow Lake-HX Refresh with a 55 W TDP. In the database, the EPYC 9135 posts a higher average benchmark score (82,980 vs. 79,574) and sits at the 96th percentile of all CPUs, one point above the Intel part's 95th percentile. Yet the head-to-head record is nearly even: Intel wins 9 of 17 tests, AMD wins 8. The results reveal two very different performance profiles, with each chip dominating in distinct workload categories.

Head-to-Head Benchmarks

The largest single-test margin belongs to AMD in Cinebench R23 single-core, where the EPYC 9135 scores 6,936 against Intel's 2,356, a 194.4% advantage. That is not a small gap; it indicates a fundamental difference in how the two chips handle lightly threaded rendering tasks. AMD also leads Cinebench R15 single-core by 105.6% (699 vs. 340) and Cinebench R23 multi-core by 23.8% (49,136 vs. 39,684). In PassMark integer math, AMD is 23.1% ahead (202,962 vs. 164,839), and in PassMark physics it leads by 61.7% (5,477 vs. 3,387). AMD also wins data compression (739,277 vs. 658,724, +12.2%), extended instructions (55,822 vs. 51,290, +8.8%), and random string sorting (90,064 vs. 80,327, +12.1%).

Intel's biggest win is in Cinebench R15 multi-core, where it scores 5,981 against AMD's 4,952, a 17.2% margin. It also takes PassMark floating-point math by 37.2% (201,773 vs. 126,679) and PassMark find prime numbers by 43.7% (519 vs. 292). In PassMark single-thread, Intel leads by 25.8% (4,951 vs. 3,672), and in data encryption it is 19.4% ahead (50,008 vs. 40,295). The remaining Intel wins are narrower: Cinebench R20 multi-core (21,198 vs. 20,637, +2.6%), Cinebench R20 single-core (2,992 vs. 2,913, +2.6%), and PassMark multithread (59,439 vs. 57,170, +3.8%). Notably, Intel's single-thread advantage in PassMark contrasts sharply with AMD's massive lead in Cinebench single-core, showing that benchmark methodology matters.

Where Each One Wins

AMD's wins cluster around integer-heavy, physics, and compression workloads. The EPYC 9135 is the clear choice for Cinebench R23 multi-core and single-core, which are common proxies for 3D rendering and general productivity. Its 23.1% lead in integer math and 61.7% lead in physics suggest strong performance in scientific simulation and engineering tasks. Data compression and random string sorting also favor AMD, pointing to database and archival workloads. The 12.2% advantage in data compression and 12.1% in random string sorting are consistent with a server chip designed for high-throughput data handling.

Intel's wins are more varied. The Core Ultra 9 290HX Plus excels in floating-point math (37.2% ahead), prime number finding (43.7% ahead), and data encryption (19.4% ahead). These are compute-heavy, often vectorized tasks. Its PassMark single-thread score is 25.8% higher, which may benefit lightly threaded applications that rely on high clock speeds. The Intel chip also wins the older Cinebench R15 multi-core test by 17.2%, though it loses the newer R23 multi-core by 23.8%, suggesting that the EPYC's architecture scales better with the newer benchmark's workload. In multithreaded PassMark, Intel is only 3.8% ahead, a narrow margin given its 24 cores versus AMD's 16 cores and 32 threads.

Architecture Differences

The two processors are built on different process nodes and microarchitectures. AMD uses TSMC's 4 nm process with Zen 5 (Turin) architecture, while Intel uses TSMC's 3 nm process with Arrow Lake-HX Refresh. AMD's die is composed of two 70.6 mm² chiplets, totaling 16,630 million transistors; Intel uses a single 243 mm² die with 17,800 million transistors. Core counts differ: AMD has 16 cores and 32 threads, Intel has 24 cores and 24 threads (no hyperthreading). Cache hierarchies are also distinct: AMD provides 80 KB L1 per core, 1 MB L2 per core, and 64 MB shared L3; Intel provides 192 KB L1 per core, 3 MB L2 per core, and 36 MB shared L3. The EPYC's larger L3 (64 MB vs. 36 MB) likely contributes to its wins in compression and integer math.

Memory support is a major differentiator. AMD uses a twelve-channel DDR5 memory bus with 576.0 GB/s bandwidth, while Intel uses a dual-channel bus with 102.4 GB/s. This 5.6x bandwidth advantage is critical for server workloads. PCIe connectivity also diverges: AMD offers 128 Gen 5 lanes (CPU only), Intel offers 20 Gen 5 lanes. The EPYC has no integrated graphics, while Intel includes Arc Xe-LPG Graphics with 64 execution units. Both support ECC memory, but the EPYC's market segment is server/workstation, while Intel's is mobile. The EPYC uses AMD Socket SP5, Intel uses BGA 2114. Clock speeds differ: AMD base 3.65 GHz, boost 4.30 GHz; Intel base 2.70 GHz, boost 5.50 GHz. Intel's multiplier is unlocked, AMD's is not. Release dates are far apart: AMD launched on 2024-10-09, Intel on 2026-03-16.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Core Ultra 9 290HX Plus has 24 cores and 24 threads, while the AMD EPYC 9135 has 16 cores and 32 threads. Intel has more physical cores, but AMD has more threads due to simultaneous multithreading.

Q: Which chip has higher boost clock speed?

A: Intel boosts to 5.50 GHz, while AMD boosts to 4.30 GHz. Intel's base clock is 2.70 GHz, AMD's is 3.65 GHz.

Q: How do memory bandwidth and PCIe lanes compare?

A: AMD provides 576.0 GB/s over a twelve-channel DDR5 bus and 128 Gen 5 PCIe lanes. Intel provides 102.4 GB/s over a dual-channel bus and 20 Gen 5 lanes. AMD's memory bandwidth is roughly 5.6 times higher.

Q: Which processor has a larger L3 cache?

A: AMD has 64 MB shared L3, Intel has 36 MB shared L3. AMD also has smaller per-core L1 and L2 caches (80 KB and 1 MB) compared to Intel's 192 KB and 3 MB.

Q: Which chip includes integrated graphics?

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

Q: What are the TDP and market segment differences?

A: AMD has a 200 W TDP and targets server/workstation; Intel has a 55 W TDP and targets mobile. Intel's lower TDP reflects its mobile design, while AMD's higher TDP supports its server role.

The Verdict

The data points to a clear split: the AMD EPYC 9135 is the stronger choice for server and workstation workloads that demand high memory bandwidth, many PCIe lanes, and superior performance in integer math, physics, compression, and Cinebench R23 rendering. Its 194.4% lead in Cinebench R23 single-core and 23.8% lead in multi-core are decisive for content creation and scientific computing. The 12.2% advantage in data compression and 12.1% in random string sorting reinforce its suitability for data-heavy server tasks. The EPYC also holds a higher average benchmark score (82,980 vs. 79,574) and a higher percentile ranking (96th vs. 95th).

The Intel Core Ultra 9 290HX Plus is the better fit for mobile or power-constrained environments where its 55 W TDP and integrated graphics are essential. It wins in floating-point math, encryption, prime number finding, and PassMark single-thread, making it strong for certain computational workloads and lightly threaded applications. Its 24 cores and 5.50 GHz boost clock give it an edge in some multithreaded tests, though the margin over AMD is often small (e.g., 3.8% in PassMark multithread). For users who need a compact, low-power processor with built-in graphics and high single-thread performance in PassMark, Intel is the logical pick. For anyone building a server or workstation where memory bandwidth, PCIe expansion, and Cinebench rendering performance are paramount, the EPYC 9135 is the clear winner.

Specification Differences

| Field | AMD EPYC 9135 | Intel Core Ultra 9 290HX Plus |

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

| Cores | 16 | 24 |

| Threads | 32 | 24 |

| Base clock | 3.65 GHz | 2.70 GHz |

| Boost clock | 4.30 GHz | 5.50 GHz |

| TDP | 200 W | 55 W |

| Socket | AMD Socket SP5 | Intel BGA 2114 |

| Architecture | Zen 5 | (not specified) |

| Codename | Turin | Arrow Lake-HX Refresh |

| Generation | EPYC (Zen 5 (Turin)) | Ultra 9 (Arrow Lake-HX) |

| Process node | 4 nm | 3 nm |

| Foundry | TSMC | TSMC |

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

| Die size | 2x 70.6 mm² | 243 mm² |

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

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

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

| Memory bus | Twelve-channel | Dual-channel |

| Memory bandwidth | 576.0 GB/s | 102.4 GB/s |

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

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

| Market segment | Server/Workstation | Mobile |

| Release date | 2024-10-09 | 2026-03-16 |

| Launch MSRP | $1214 | (not recorded) |

| Multiplier unlocked | No | Yes |

| Part number | 100-000001150 | SADSS |