AMD EPYC 9115 vs Intel Core Ultra 7 265K Comparison

Head-to-Head Benchmarks

The benchmark data presents a clear split between these two processors, with the Intel Core Ultra 7 265K winning 13 of the 17 recorded head-to-head tests, and the AMD EPYC 9115 taking 4. The magnitude of those wins varies wildly depending on the workload type.

The Intel chip dominates the Cinebench R15 and R20 suites. In Cinebench R15 multicore, the 265K scores 5020 versus 4233 for the EPYC, a consistent 18.6% advantage. The same 18.6% delta appears in Cinebench R15 singlecore (708 vs 597), Cinebench R20 multicore (20918 vs 17641), and Cinebench R20 singlecore (2953 vs 2490). This uniformity suggests the Intel architecture holds a steady per-clock efficiency edge in these older rendering tests.

The picture flips dramatically in Cinebench R23. Here the EPYC 9115 posts 42003 multicore against the Intel’s 35850, a 14.6% swing in AMD’s favor. Even more extreme, the EPYC’s singlecore score of 5929 crushes the Intel’s 2020, a 65.9% margin. This is the single largest performance gap in the entire comparison, and it indicates the Cinebench R23 single-threaded test responds very differently to these two designs than earlier Cinebench versions.

In the Passmark suite, the Intel part wins most categories but often by varying amounts. The biggest Intel victories come in passmark_floating_point_math (189629 vs 113853, a 66.6% lead) and passmark_find_prime_numbers (491 vs 289, a 69.9% lead). Passmark_single_thread shows a 46.7% advantage for Intel (4928 vs 3360), and passmark_data_encryption shows a 44.1% lead (48246 vs 33489). Smaller but still solid Intel wins appear in passmark_extended_instructions (19.5%), passmark_multithread (19.7%), passmark_random_string_sorting (13.7%), and passmark_data_compression (10.9%).

The EPYC’s other two wins are narrower. Passmark_integer_math goes to AMD at 181807 versus 143242, a 21.2% margin. Passmark_physics favors AMD at 4188 versus 3731, a 10.9% lead. Notably, the EPYC’s multicore Cinebench R23 win is its only rendering victory, and it comes despite the Intel chip winning the two earlier multicore Cinebench tests by identical 18.6% margins.

Overall average benchmark scores sit close: the Intel part records 70879 versus 69288 for the EPYC, a difference of about 1.6% in Intel’s favor. Both processors land in the 94th percentile of all CPUs in the database.

Where Each One Wins

The Intel Core Ultra 7 265K is the stronger choice for single-threaded desktop responsiveness and legacy rendering workloads. Its Passmark single-thread score of 4928 versus 3360 shows a 46.7% advantage, which translates to snappier everyday application behavior. The consistent 18.6% wins across Cinebench R15 and R20, both multicore and singlecore, point to a mature architecture that handles older render engines efficiently. The Intel part also dominates encryption work with a 44.1% lead in passmark_data_encryption, plus strong showings in compression (10.9%) and extended instruction throughput (19.5%). For floating-point math, prime number finding, and random string sorting, the Intel chip leads by 66.6%, 69.9%, and 13.7% respectively.

The AMD EPYC 9115 excels in workloads that stress sustained multithreaded throughput with high memory bandwidth. Its Cinebench R23 multicore score of 42003 tops the Intel’s 35850 by 14.6%, and its singlecore R23 result of 5929 is a massive 65.9% higher. The integer math win (181807 vs 143242, a 21.2% lead) suggests the EPYC handles integer-heavy calculations better. The physics test also favors AMD, at 4188 versus 3731, a 10.9% margin. The EPYC’s twelve-channel DDR5 memory bus, rated at 576.0 GB/s versus the Intel’s dual-channel 102.4 GB/s, likely explains why memory-bandwidth-sensitive workloads such as Cinebench R23 and integer math see AMD pull ahead.

The Verdict

The data supports a clear use-case split. The Intel Core Ultra 7 265K is the better desktop processor for users who prioritize single-threaded speed, legacy application compatibility, and encryption or floating-point workloads. Its 20 cores and 20 threads handle modern multitasking well, and its 5.50 GHz boost clock against the EPYC’s 4.10 GHz helps explain the consistent single-thread wins. The Intel chip also carries integrated Arc Xe-LPG Graphics 64EU, making it a self-contained desktop solution.

The AMD EPYC 9115 is the better choice for server or workstation deployments where Cinebench R23-class rendering, integer math, and physics simulations matter more than legacy benchmarks. Its 16 cores with 32 threads provide twice the thread count of the Intel part, and the 64 MB shared L3 cache plus twelve-channel memory bandwidth give it a structural advantage in sustained parallel workloads. The 65.9% singlecore R23 lead is unusual and suggests the EPYC’s Zen 5 architecture extracts exceptional performance from that specific test.

For a desktop PC, the Intel part wins on versatility and raw single-thread performance. For a rack server handling modern render farms or data processing, the EPYC’s R23 and integer math results make it the safer bet. Neither processor is universally faster; the database shows a 13-4 win split, but the EPYC’s wins occur in exactly the categories where server buyers typically focus.

FAQ

Q: Which processor has the higher average benchmark score?

A: The Intel Core Ultra 7 265K records an average benchmark score of 70879, while the AMD EPYC 9115 records 69288, giving Intel a 1.6% overall lead.

Q: Why does the AMD EPYC win Cinebench R23 multicore despite losing Cinebench R15 and R20 multicore?

A: The EPYC scores 42003 in R23 multicore versus 35850 for Intel, a 14.6% win, but Intel wins R15 and R20 multicore by 18.6% each. The R23 test appears to favor the EPYC’s 32 threads and 64 MB L3 cache, while the older tests respond better to the Intel’s higher clock speeds.

Q: What is the largest single benchmark gap between these two processors?

A: The biggest difference is in Cinebench R23 singlecore, where the AMD EPYC scores 5929 against Intel’s 2020, a 65.9% margin in AMD’s favor.

Q: Does the Intel processor have integrated graphics?

A: Yes, the Intel Core Ultra 7 265K includes Arc Xe-LPG Graphics 64EU. The AMD EPYC 9115 has no integrated graphics (listed as N/A).

Q: How do the memory bandwidth specifications compare?

A: The Intel chip uses dual-channel DDR5 with 102.4 GB/s bandwidth. The AMD EPYC uses twelve-channel DDR5 with 576.0 GB/s bandwidth, over five times the memory throughput.

Q: Which processor has more PCIe lanes?

A: The AMD EPYC 9115 offers 128 PCIe Gen 5 lanes from the CPU, while the Intel Core Ultra 7 265K offers 20 PCIe Gen 5 lanes from the CPU.

Architecture Differences

The two chips come from fundamentally different design philosophies. The Intel Core Ultra 7 265K uses the Arrow Lake architecture on a 3 nm process, built by TSMC, with 20 cores and 20 threads. It lacks hyperthreading, which explains why thread count equals core count. The Intel chip has 17,800 million transistors on a 243 mm² die, and its cache hierarchy uses 192 KB L1 per core, 3 MB L2 per core, and 30 MB shared L3. It supports DDR5 memory in dual-channel mode, includes integrated Arc Xe-LPG Graphics 64EU, and has an unlocked multiplier for overclocking.

The AMD EPYC 9115 uses the Zen 5 (Turin) architecture on a 4 nm process, also from TSMC, with 16 cores and 32 threads. SMT gives it double the thread count of the Intel part. The EPYC packs 16,630 million transistors across two chiplets, each 70.6 mm², for a combined die area of roughly 141.2 mm². Its cache layout uses 80 KB L1 per core, 1 MB L2 per core, and a much larger 64 MB shared L3. The EPYC supports twelve-channel DDR5 memory, has no integrated graphics, and its multiplier is locked. Both processors support ECC memory, and both launched in October 2024, with the EPYC releasing on October 9 and the Intel on October 23.

Specification Differences

| Specification | Intel Core Ultra 7 265K | AMD EPYC 9115 |

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

| Cores | 20 | 16 |

| Threads | 20 | 32 |

| Base clock | 3.90 GHz | 2.60 GHz |

| Boost clock | 5.50 GHz | 4.10 GHz |

| 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 | 30 MB (shared) | 64 MB (shared) |

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

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

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

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

| Socket | Intel Socket 1851 | AMD Socket SP5 |

| Multiplier unlocked | Yes | No |

| Launch MSRP | $394 | $726 |

The Intel chip has a higher base clock (3.90 vs 2.60 GHz) and boost clock (5.50 vs 4.10 GHz), more cores (20 vs 16), but fewer threads (20 vs 32). The EPYC counters with double the L3 cache (64 MB vs 30 MB), over five times the memory bandwidth (576.0 vs 102.4 GB/s), and six times the PCIe lanes (128 vs 20). The Intel part integrates graphics and supports overclocking; the EPYC does neither. The EPYC’s launch MSRP of $726 is nearly double the Intel’s $394, though pricing analysis is outside the scope of this database entry. Both processors are active in production, both use TSMC fabrication, and both support ECC memory.