AMD EPYC 4344P vs Intel Core Ultra 5 235 Comparison

Where Each One Wins

The recorded benchmark data splits this comparison into two clearly defined territories. The AMD EPYC 4344P dominates rendering workloads, winning every Cinebench test across R15, R20, and R23 in both single-core and multi-core configurations. Its margin is consistent at 48.4% ahead of the Intel Core Ultra 5 235 across all six Cinebench tests, a remarkably uniform advantage that points to a fundamental architectural edge in sustained compute throughput.

The Intel Core Ultra 5 235, however, takes the majority of the PassMark suite, winning eight of the eleven PassMark tests. Its wins are concentrated in single-threaded responsiveness, physics simulation, floating-point math, and integer-heavy operations. The most striking victory comes in prime number finding, where Intel scores 124.8% higher than the EPYC. Floating-point math also shows a substantial 86.3% advantage, while single-thread PassMark scores favor Intel by 28.1%. The multithread PassMark test, which often favors higher core counts, also goes to Intel by 13.5%, suggesting that the Intel part's thread scheduling and per-core efficiency translate well into this particular workload mix.

The AMD EPYC 4344P counters with wins in integer math, data compression, and random string sorting, though these margins are narrower. The data compression result is close at 3% in favor of AMD, and random string sorting is effectively a tie at 0% delta. Integer math is a more decisive 16.9% win for AMD. In terms of raw win counts, AMD edges out Intel nine wins to eight across all recorded benchmarks, but the magnitude of Intel's wins in several PassMark tests exceeds the magnitude of AMD's Cinebench dominance in percentage terms.

Architecture Differences

The two processors represent fundamentally different design philosophies. The Intel Core Ultra 5 235 is built on Arrow Lake architecture, fabricated on a 3 nm process at TSMC, with 14 cores and 14 threads. It uses a flat core topology without simultaneous multithreading, which explains the equal core and thread counts. The die contains 17,800 million transistors on a 243 mm² die, with 192 KB of L1 cache per core, 3 MB of L2 cache per core, and 24 MB of shared L3 cache.

The AMD EPYC 4344P uses Zen 4 architecture, specifically the Raphael codename, on a 5 nm TSMC process. It has 8 cores and 16 threads, enabled by simultaneous multithreading. The die is substantially smaller at 71 mm² with 6,570 million transistors. Cache configuration differs notably: 64 KB of L1 per core, 1 MB of L2 per core, and a larger 32 MB of shared L3 cache. The smaller per-core L2 cache is offset by the larger L3 pool, which can benefit workloads with shared data access patterns.

Process node differences are significant: Intel uses a 3 nm process versus AMD's 5 nm process, both from TSMC. This gives Intel a transistor density advantage on paper, though the AMD design achieves competitive results with fewer, larger cores. Clock speeds favor AMD: the EPYC 4344P has a 3.80 GHz base clock and 5.30 GHz boost clock, while the Intel part runs at 3.40 GHz base and 5.00 GHz boost. Both have a 65 W TDP.

Memory support diverges in one critical aspect. Both support DDR5 and dual-channel memory buses, but memory bandwidth differs: Intel records 102.4 GB/s versus AMD's 83.2 GB/s. The AMD EPYC 4344P supports ECC memory, while the Intel Core Ultra 5 235 does not. PCIe capabilities also differ, with AMD offering Gen 5 across 28 CPU lanes versus Intel's Gen 5 across 20 CPU lanes. The Intel part includes integrated Arc Xe-LPG Graphics with 24 execution units, while AMD includes Radeon Graphics.

Head-to-Head Benchmarks

The Cinebench results reveal a consistent and dramatic gap. In Cinebench R23 multi-core, the AMD EPYC 4344P scores 28,636 against Intel's 14,769, a 48.4% advantage. The same percentage appears in R23 single-core: 4,042 versus 2,085. This pattern repeats in Cinebench R20 multi-core (12,027 versus 6,202) and single-core (1,697 versus 875), and in R15 multi-core (2,886 versus 1,488) and single-core (407 versus 210). The uniformity of the 48.4% delta across every Cinebench test is remarkable, indicating that the advantage scales identically regardless of thread count or workload complexity within that benchmark family.

PassMark results tell a different story. In find prime numbers, Intel scores 371 versus AMD's 165, a 124.8% advantage. Floating-point math shows Intel at 117,951 versus 63,309, an 86.3% lead. Physics simulation favors Intel 2,570 to 1,987, a 29.3% margin. Single-thread performance in PassMark gives Intel 4,516 versus 3,526, a 28.1% lead, and the multithread test gives Intel 37,816 versus 33,325, a 13.5% advantage. Data encryption favors Intel by 19.7% (29,293 versus 24,476), and extended instructions go to Intel by 10.7% (32,752 versus 29,582).

AMD's PassMark wins include integer math at 105,779 versus 87,948, a 16.9% margin. Data compression is close: 402,701 versus 390,711, only 3% apart. Random string sorting is essentially tied at 49,001 versus 48,980, a 0% delta. These results suggest that AMD's architecture excels at sustained integer throughput and data manipulation, while Intel's design pulls ahead in floating-point, encryption, and latency-sensitive single-thread tasks.

Specification Differences

The table below lists only the fields where the two processors differ in the recorded data.

| Specification | Intel Core Ultra 5 235 | AMD EPYC 4344P |

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

| Cores | 14 | 8 |

| Threads | 14 | 16 |

| Base clock | 3.40 GHz | 3.80 GHz |

| Boost clock | 5.00 GHz | 5.30 GHz |

| Socket | Intel Socket 1851 | AMD Socket AM5 |

| Architecture | Arrow Lake | Zen 4 |

| Codename | Arrow Lake-S | Raphael |

| Generation | Ultra 5 (Arrow Lake) | EPYC (Zen 4 (Raphael)) |

| Process node | 3 nm | 5 nm |

| Transistors | 17,800 million | 6,570 million |

| Die size | 243 mm² | 71 mm² |

| L1 cache | 192 KB (per core) | 64 KB (per core) |

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

| L3 cache | 24 MB (shared) | 32 MB (shared) |

| Memory bandwidth | 102.4 GB/s | 83.2 GB/s |

| ECC memory | false | true |

| PCIe | Gen 5, 20 Lanes (CPU only) | Gen 5, 28 Lanes (CPU only) |

| Integrated graphics | Arc Xe-LPG Graphics 24EU | Radeon Graphics |

| Market segment | Desktop | Server/Workstation |

| Release date | 2025-01-06 | 2024-05-20 |

| Launch MSRP | $257 | $329 |

| Part number | SRQAS | 100-000001479 |

Both processors are active in production, have locked multipliers, and fall in the 89th percentile of all CPUs in the database. The Intel part has a higher average benchmark score of 46,062 versus 45,122 for AMD, though both sit close to their nearest rivals. Intel's nearest competitor is the AMD Ryzen AI 9 HX 375 at 46,030 (0.1% delta), while AMD's nearest rival is the Intel Core i9-12900KS at 45,094 (0.1% delta).

FAQ

Q: Which processor is faster in Cinebench rendering tests?

A: The AMD EPYC 4344P wins every Cinebench test by exactly 48.4%, including R23 multi-core where it scores 28,636 versus Intel's 14,769.

Q: Does the Intel Core Ultra 5 235 win any benchmarks?

A: Yes, it wins eight of the seventeen head-to-head tests, including floating-point math by 86.3%, prime number finding by 124.8%, single-thread PassMark by 28.1%, and data encryption by 19.7%.

Q: What is the core and thread configuration of each processor?

A: The Intel Core Ultra 5 235 has 14 cores and 14 threads, while the AMD EPYC 4344P has 8 cores and 16 threads. Intel does not use simultaneous multithreading, while AMD does.

Q: Which processor supports ECC memory?

A: The AMD EPYC 4344P supports ECC memory. The Intel Core Ultra 5 235 does not.

Q: How do the clock speeds compare?

A: The AMD EPYC 4344P has a base clock of 3.80 GHz and a boost clock of 5.30 GHz. The Intel Core Ultra 5 235 has a base clock of 3.40 GHz and a boost clock of 5.00 GHz.

Q: Which processor has a higher average benchmark score?

A: The Intel Core Ultra 5 235 has an average benchmark score of 46,062, compared to 45,122 for the AMD EPYC 4344P. Both are in the 89th percentile of all CPUs.