AMD EPYC 9124 vs Intel Core Ultra 7 265 Comparison

The AMD EPYC 9124 and Intel Core Ultra 7 265 represent two fundamentally different interpretations of high-performance computing, separated by market segment, architecture, and intended workload. The data shows a near-total statistical tie in aggregate performance—the EPYC 9124 averages 65,104 points against the Core Ultra 7 265’s 64,640, a mere 0.7% gap—yet the benchmark-by-benchmark breakdown reveals starkly divergent strengths. This analysis examines where each processor dominates, the architectural roots of those differences, and the specific numbers that separate them.

Where Each One Wins

The AMD EPYC 9124 is a server-class part, and its wins cluster around memory-bandwidth-hungry and integer-heavy workloads. It takes 7 of the 17 head-to-head tests, including decisive victories in multi-threaded cinebench r20 (15,652 vs 6,268, a 149.7% margin) and single-core cinebench r20 (2,209 vs 884, +149.9%). Its Passmark wins include data compression (599,417 vs 522,983, +14.6%), extended instructions (43,380 vs 41,478, +4.6%), integer math (148,785 vs 134,773, +10.4%), physics (3,662 vs 2,923, +25.3%), and random string sorting (74,177 vs 63,833, +16.2%). These are classic server patterns: compression, sorting, and integer throughput benefit from the EPYC’s massive L3 cache and twelve-channel memory bus.

The Intel Core Ultra 7 265 counters with 10 wins, concentrated in single-threaded responsiveness and floating-point throughput. It leads in all three cinebench R15 and R23 metrics, including multi-core R23 (42,216 vs 37,269, +11.7%) and single-core R23 (5,960 vs 5,261, +11.7%). Its Passmark advantages are equally pronounced: single-thread score (4,689 vs 2,719, +42%), floating-point math (172,776 vs 87,057, +49.6%), prime number finding (418 vs 256, +38.8%), data encryption (40,456 vs 36,078, +10.8%), and the multithread aggregate (49,682 vs 43,846, +11.7%). The pattern is clear: Intel wins where clock speed and per-core efficiency matter, AMD wins where memory bandwidth and cache capacity dominate.

Architecture Differences

The two chips diverge at every level of design. The EPYC 9124 is built on TSMC’s 5 nm process with 26,280 million transistors spread across four 72 mm² chiplets, using AMD’s Zen 4 architecture (codename Genoa). It is a 16-core, 32-thread part with a base clock of 3.00 GHz and boost of 3.70 GHz, drawing a 200 W TDP. Crucially, it features 64 MB of shared L3 cache, 1 MB L2 per core, and 64 KB L1 per core. Its memory subsystem is twelve-channel DDR5, yielding 460.8 GB/s of bandwidth, and it connects via 128 PCIe Gen 5 lanes. It is a server/workstation part on AMD Socket SP5, with ECC memory support and no integrated graphics.

The Core Ultra 7 265 is a desktop processor on Intel Socket 1851, built on TSMC’s 3 nm process with 17,800 million transistors on a 243 mm² die. It uses Arrow Lake-S architecture (Core Ultra Series 2) with 20 cores and 20 threads—no hyperthreading—running at 2.40 GHz base and 5.30 GHz boost, within a 65 W TDP. Its cache layout is larger per core but smaller overall: 192 KB L1 per core, 3 MB L2 per core, and only 30 MB shared L3. Memory is dual-channel DDR5 with 102.4 GB/s bandwidth, and it offers just 20 PCIe Gen 5 lanes. It includes Arc Xe-LPG Graphics 32EU integrated graphics, does not support ECC, and has a much higher single-core clock ceiling. The process node advantage (3 nm vs 5 nm) and the 1.6 GHz boost advantage explain much of Intel’s single-thread dominance, while AMD’s 4.5x memory bandwidth advantage and 2.1x larger L3 cache explain its server-side wins.

Head-to-Head Benchmarks

The largest single margin in the entire comparison is AMD’s 149.7% lead in cinebench r20 multi-core, followed by the identical 149.9% lead in r20 single-core. These results are anomalous compared to every other multi-core test, where Intel holds an 11.7% edge. The r20 numbers suggest a possible workload-specific optimization for AMD’s architecture in that particular benchmark version, but the data stands as recorded. In contrast, cinebench R15 and R23 both show Intel winning by a consistent 11.7% in both single- and multi-core, indicating that the Core Ultra 7 265’s higher boost clock (5.30 GHz vs 3.70 GHz) translates into a steady advantage across those rendering workloads.

The Passmark suite reveals the most instructive splits. Intel’s floating-point math score of 172,776 is 49.6% higher than AMD’s 87,057, a massive gap that likely reflects the Core Ultra’s superior per-core FPU throughput and higher clock. Similarly, Intel’s single-thread score of 4,689 beats AMD’s 2,719 by 42%, and prime number finding (418 vs 256) shows a 38.8% Intel lead. AMD’s wins are equally decisive in their domains: physics (3,662 vs 2,923, +25.3%), random string sorting (74,177 vs 63,833, +16.2%), and data compression (599,417 vs 522,983, +14.6%) all favor the EPYC. Integer math goes to AMD by 10.4% (148,785 vs 134,773), while encryption favors Intel by 10.8% (40,456 vs 36,078). The overall multithread Passmark score goes to Intel at 49,682 vs 43,846, an 11.7% margin that mirrors the cinebench multi-core results.

Specification Differences

The two processors differ in nearly every measurable specification. Core count: 16 (AMD) vs 20 (Intel). Thread count: 32 vs 20. Base clock: 3.00 GHz vs 2.40 GHz. Boost clock: 3.70 GHz vs 5.30 GHz. TDP: 200 W vs 65 W. Process node: 5 nm vs 3 nm. Transistor count: 26,280 million vs 17,800 million. Die size: 4x 72 mm² vs 243 mm². L1 cache: 64 KB per core vs 192 KB per core. L2 cache: 1 MB per core vs 3 MB per core. L3 cache: 64 MB shared vs 30 MB shared. Memory bus: twelve-channel vs dual-channel. Memory bandwidth: 460.8 GB/s vs 102.4 GB/s. PCIe lanes: 128 vs 20 (both Gen 5). ECC support: yes vs no. Integrated graphics: none vs Arc Xe-LPG Graphics 32EU. Socket: SP5 vs 1851. Market segment: Server/Workstation vs Desktop. Release date: 2022-11-09 vs 2025-01-06. Launch MSRP for the EPYC 9124 is $1083; launch MSRP for the Core Ultra 7 265 is $394.

FAQ

Q: Which processor has a higher average benchmark score?

A: The AMD EPYC 9124 averages 65,104 points versus the Intel Core Ultra 7 265’s 64,640, a 0.7% difference. Both sit at the 93rd percentile of all CPUs.

Q: Why does the EPYC 9124 win some multi-core tests but lose others?

A: The EPYC 9124 wins cinebench r20 multi-core by 149.7% (15,652 vs 6,268), but loses cinebench R15 multi-core (3,756 vs 4,255) and R23 multi-core (37,269 vs 42,216) by 11.7% each. The r20 result is an outlier, while the R15/R23 pattern aligns with Intel’s higher boost clock.

Q: How large is the single-thread performance gap?

A: Intel leads by 42% in Passmark single-thread (4,689 vs 2,719) and by 11.7% in cinebench R23 single-core (5,960 vs 5,261). The boost clock difference—5.30 GHz vs 3.70 GHz—is the primary factor.

Q: Which processor has more memory bandwidth?

A: The EPYC 9124 provides 460.8 GB/s via twelve-channel DDR5, which is 4.5 times the Core Ultra 7 265’s 102.4 GB/s from dual-channel DDR5. This explains AMD’s wins in data compression and random string sorting.

Q: Do both processors support ECC memory?

A: No. The AMD EPYC 9124 supports ECC memory, while the Intel Core Ultra 7 265 does not. This reflects their respective server and desktop market segments.

Q: What is the core and thread configuration difference?

A: The EPYC 9124 has 16 cores and 32 threads (with SMT), while the Core Ultra 7 265 has 20 cores and 20 threads (no SMT). Intel compensates with a 5.30 GHz boost clock versus AMD’s 3.70 GHz.