AMD EPYC 7F72 vs Intel Xeon 654 Comparison

Head-to-Head Benchmarks

The benchmark comparison between the Intel Xeon 654 and the AMD EPYC 7F72 reveals a clear split across different workload types. The Intel Xeon 654 wins 13 of the 17 recorded head-to-head tests, while the AMD EPYC 7F72 takes 4 wins. The average benchmark score for the Intel part is 90,717, placing it 6.6% above the AMD EPYC 7F72's 85,072 average. Both processors sit at the 96th percentile among all CPUs in the database.

The most striking margin comes in single-threaded performance. The Intel Xeon 654 scores 3,778 in the PassMark single-thread test, which is 58.5% higher than the EPYC 7F72's 2,384. This gap repeats across the Cinebench suite: the Xeon 654 leads by 16.5% in Cinebench R15 single-core (742 vs. 637), by 16.4% in R20 single-core (3,092 vs. 2,657), and by 16.3% in R23 single-core (7,362 vs. 6,328). These consistent margins indicate that the Intel architecture holds a substantial per-thread advantage.

Multi-core results follow a similar but slightly narrower pattern. In Cinebench R23 multi-core, the Xeon 654 scores 52,150 against 44,829 for the EPYC 7F72, a 16.3% advantage. The R20 multi-core test shows 21,903 vs. 18,828, again a 16.3% lead. The R15 multi-core result is 5,256 vs. 4,518, a 16.3% difference. The PassMark multi-thread score of 61,353 for Intel versus 52,740 for AMD confirms the same 16.3% gap. It is notable that the EPYC 7F72 has 24 cores and 48 threads, while the Xeon 654 has 18 cores and 36 threads, yet the Intel chip still delivers higher aggregate multi-threaded throughput.

Floating-point math shows the largest Intel win. The Xeon 654 records 163,093 in PassMark floating-point math, which is 50.4% ahead of the EPYC 7F72's 108,437. Extended instruction performance also strongly favors Intel: 63,539 vs. 46,936, a 35.4% advantage. Integer math goes to Intel by 14.7% (207,745 vs. 181,103), and data compression is a narrow Intel win at 818,902 vs. 808,795, just 1.2% apart.

The AMD EPYC 7F72 takes its wins in more specialized areas. Data encryption is its best result: 56,261 vs. 40,675, a 27.7% advantage for AMD. Random string sorting goes to AMD at 102,436 vs. 82,828, a 19.1% margin. Prime number finding favors AMD at 498 vs. 390, a 21.7% lead, and the physics test goes to AMD at 6,459 vs. 5,596, a 13.4% edge. These four wins suggest that the EPYC 7F72 has specific strengths in integer-heavy and memory-intensive operations, even though its overall average is lower.

Where Each One Wins

The Intel Xeon 654 is the clear choice for single-threaded responsiveness and general compute throughput. Its 58.5% lead in PassMark single-thread and 16.3% to 16.5% leads across all three Cinebench versions mean that any workload dominated by per-core performance will favor Intel. The 50.4% advantage in floating-point math makes it well-suited for scientific computing, simulations, and rendering tasks that rely on floating-point throughput. The 35.4% lead in extended instructions indicates strong SIMD and vector processing capabilities. Integer math, data compression, and multi-threaded workloads also go to Intel, making it a versatile all-rounder.

The AMD EPYC 7F72 wins in four targeted areas. Data encryption performance is 27.7% higher, which points to stronger cryptographic acceleration. Random string sorting is 19.1% higher, suggesting that memory-bound sorting workloads perform better on the AMD platform. Prime number finding is 21.7% higher, a classic integer-heavy test that favors AMD's core count. The physics test shows a 13.4% AMD advantage. These wins cluster around tasks that benefit from the EPYC 7F72's larger L3 cache and higher core count, even though its per-core performance is lower.

For a server or workstation that runs mixed workloads, the Intel Xeon 654 offers broader dominance across most benchmark categories. The AMD EPYC 7F72 would be preferable in environments where encryption, sorting, or physics simulation is the primary task, given its consistent wins in those specific tests.

Architecture Differences

The Intel Xeon 654 is built on Granite Rapids architecture using a 5 nm process at Intel's own foundry. It belongs to the Xeon 600 (Granite Rapids-WS) generation and uses Intel Socket 4710. The die is composed of 2x 598 mm² tiles. The AMD EPYC 7F72 uses Zen 2 architecture, codenamed Rome, manufactured on a 7 nm process by TSMC. It fits into AMD Socket SP3 and uses a 74 mm² die. The EPYC 7F72 contains 3,800 million transistors, while the transistor count for the Xeon 654 is not recorded in the database.

Cache hierarchies differ substantially. The Xeon 654 allocates 112 KB of L1 cache per core and 2 MB of L2 per core, with 72 MB of shared L3. The EPYC 7F72 has 96 KB of L1 per core, 512 KB of L2 per core, and 192 MB of shared L3. The AMD part has significantly more L3 cache, which helps explain its wins in memory-sensitive tasks like random string sorting. The Intel part has larger per-core L1 and L2 caches, which supports its single-threaded advantage.

The memory subsystems also diverge. The Xeon 654 supports DDR5 memory with an eight-channel bus and 409.6 GB/s of bandwidth. The EPYC 7F72 uses DDR4 with an eight-channel bus and 204.8 GB/s of bandwidth. This is a key architectural difference: Intel has twice the memory bandwidth, which likely contributes to its dominant floating-point and multi-threaded results. Both platforms support ECC memory.

PCIe connectivity differs as well. The Xeon 654 provides Gen 5 with 128 lanes (CPU only), while the EPYC 7F72 offers Gen 4 without a lane count specified in the database. The Intel part has an unlocked multiplier, whereas the EPYC 7F72 does not. Neither processor includes integrated graphics.

Specification Differences

The two processors differ across nearly every core specification. The Xeon 654 has 18 cores and 36 threads, while the EPYC 7F72 has 24 cores and 48 threads. Base clocks are close: the Xeon 654 runs at 3.10 GHz, and the EPYC 7F72 at 3.20 GHz. Boost clocks differ more significantly, with the Xeon 654 reaching 4.80 GHz versus 3.70 GHz for the EPYC 7F72. This 1.10 GHz boost advantage for Intel aligns with its single-thread performance lead.

Thermal design power is higher on the AMD side. The EPYC 7F72 is rated at 240 W, while the Xeon 654 is rated at 200 W. The Intel part has a launch MSRP of $1,199, while no launch MSRP is recorded for the AMD part. Process nodes differ as noted: 5 nm for Intel versus 7 nm for AMD. Memory support is DDR5 for Intel and DDR4 for AMD, with the bandwidth difference already noted. The EPYC 7F72 was released in April 2020, while the Xeon 654 has a release date of February 2026. Both are marked as Active in production status and target the server/workstation market segment.

FAQ

Q: Which processor has the higher average benchmark score?

A: The Intel Xeon 654 has an average benchmark score of 90,717, which is 6.6% higher than the AMD EPYC 7F72's 85,072.

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

A: In the PassMark single-thread test, the Intel Xeon 654 scores 3,778 versus 2,384 for the AMD EPYC 7F72, a 58.5% advantage for Intel.

Q: Does the AMD EPYC 7F72 win any benchmark categories?

A: Yes, it wins 4 of 17 head-to-head tests: data encryption (56,261 vs. 40,675, a 27.7% lead), random string sorting (102,436 vs. 82,828, a 19.1% lead), find prime numbers (498 vs. 390, a 21.7% lead), and physics (6,459 vs. 5,596, a 13.4% lead).

Q: How do the core counts compare?

A: The AMD EPYC 7F72 has 24 cores and 48 threads, while the Intel Xeon 654 has 18 cores and 36 threads. Despite fewer cores, the Intel part wins multi-threaded benchmarks by about 16.3%.

Q: What is the memory bandwidth difference?

A: The Intel Xeon 654 supports DDR5 with 409.6 GB/s of bandwidth, while the AMD EPYC 7F72 uses DDR4 with 204.8 GB/s. Both use an eight-channel memory bus.

Q: Which processor has a higher boost clock?

A: The Intel Xeon 654 boosts to 4.80 GHz, compared to 3.70 GHz for the AMD EPYC 7F72. The base clocks are 3.10 GHz and 3.20 GHz, respectively.