AMD EPYC 9184X vs Intel Xeon 6515P Comparison

The AMD EPYC 9184X and Intel Xeon 6515P are both 16-core, 32-thread server processors, but the benchmark data shows they serve very different roles. The EPYC 9184X wins 13 of the 17 head-to-head comparisons, while the Xeon 6515P takes 4, with the AMD part leading by an average margin that spans from a narrow 1.2% deficit in single-threaded Passmark to a commanding 74.3% advantage in Passmark physics. The Xeon 6515P, however, counters with substantial wins in floating-point math and extended instructions, making the choice between them a matter of workload profile rather than outright superiority.

The Verdict

The data points to the AMD EPYC 9184X as the default choice for general-purpose server workloads, given its 13 wins across the benchmark suite. Its average benchmark score of 68202 places it in the 94th percentile of all CPUs, while the Intel Xeon 6515P’s 67006 average sits in the 93rd percentile—a small but consistent gap. The EPYC 9184X is particularly strong in memory-sensitive and integer-heavy tasks, as evidenced by its wins in Passmark data encryption (22.6% ahead), random string sorting (24% ahead), and find prime numbers (20.8% ahead). For buyers running database, compression, or encryption workloads, the AMD part is the clear winner from the data.

The Intel Xeon 6515P, by contrast, is the part to pick for floating-point or SIMD-heavy applications. Its Passmark floating-point math score of 128954 is 26% higher than the EPYC 9184X’s 95476, and its extended instructions score of 53383 beats the AMD’s 43562 by 18.4%. The Xeon also edges out the AMD in Passmark single-thread performance, scoring 2855 versus 2822, a 1.2% margin. If the workload is dominated by vector math, scientific computing, or similar floating-point operations, the Intel part has a measurable advantage. The Xeon 6515P’s launch MSRP is $740, which may inform procurement decisions, but the data alone does not quantify value.

Where Each One Wins

Breaking down the benchmark suite by category, the AMD EPYC 9184X dominates in multi-threaded and memory-bound tasks. In Cinebench R23 multicore, it scores 40515 versus the Xeon’s 38547, a 5.1% lead that repeats across all Cinebench versions—R15, R20, and R23—both single-core and multi-core, with the same 5.1% delta each time. The AMD part also wins Passmark multithread (47665 vs 45350, 5.1% ahead), integer math (157483 vs 147047, 7.1% ahead), and data compression (614873 vs 592645, 3.8% ahead). Its biggest win is in Passmark physics, where it scores 6674 against the Xeon’s 3829, a 74.3% margin that suggests a substantial advantage in physics simulation or constraint-based workloads. Data encryption also favors AMD heavily: 37376 vs 30476, a 22.6% lead.

The Intel Xeon 6515P wins are concentrated in two areas: floating-point math and extended instruction sets. The floating-point math score of 128954 is 26% higher than the AMD’s 95476, which is the largest single-benchmark win for Intel in absolute terms. The extended instructions score of 53383 represents an 18.4% advantage over the AMD’s 43562. Additionally, the Xeon takes Passmark single-thread (2855 vs 2822, 1.2% ahead) and the duplicate singlethread test with the same scores. These wins indicate that for workloads that can exploit AVX-512 or similar wide vector instructions, the Xeon 6515P is clearly superior, but those wins are narrow in single-thread and do not translate to overall multi-threaded dominance.

Architecture Differences

The two processors come from different architectural lineages. The AMD EPYC 9184X is built on the Zen 4 architecture with the codename Genoa-X, part of the EPYC 9004 series, and uses a 5 nm process at TSMC. It integrates 90,160 million transistors across a die size of 8x 72 mm², with a complex cache hierarchy: 64 KB of L1 per core, 1 MB of L2 per core, and a massive 768 MB of shared L3 cache. That 768 MB L3 is the standout feature, likely enabling the AMD’s strength in data compression and encryption, where large working sets can remain on-chip. The EPYC 9184X supports DDR5 memory over a twelve-channel bus, yielding a memory bandwidth of 460.8 GB/s, and provides 128 PCIe Gen 5 lanes. It was released on 2023-06-12 and remains in active production.

The Intel Xeon 6515P uses the Granite Rapids architecture, part of the Xeon 6 (Granite Rapids-SP) generation, also on a 5 nm process but fabricated by Intel. Its cache layout differs: 112 KB of L1 per core, 2 MB of L2 per core, and only 72 MB of shared L3—a tenth of the AMD’s L3 capacity. The Xeon supports DDR5 over an eight-channel bus with a memory bandwidth of 409.6 GB/s, which is 51.2 GB/s lower than the AMD part. It provides 88 PCIe Gen 5 lanes, fewer than the AMD’s 128. The Xeon 6515P was released on 2025-02-23, significantly later than the AMD, and also carries active production status. Intel lists its integrated graphics as N/A, while the AMD part has no integrated graphics listed. The Xeon’s part number is SRVU6.

FAQ

Q: Which processor has a higher average benchmark score?

A: The AMD EPYC 9184X has an average benchmark score of 68202, which is higher than the Intel Xeon 6515P’s 67006. The AMD part sits in the 94th percentile of all CPUs, while the Intel part sits in the 93rd percentile.

Q: How do the two compare in Cinebench R23 multicore?

A: The AMD EPYC 9184X scores 40515 in Cinebench R23 multicore, which is 5.1% higher than the Intel Xeon 6515P’s 38547. This 5.1% delta is consistent across all Cinebench R15, R20, and R23 tests, both single-core and multi-core.

Q: What is the largest benchmark win for the Intel Xeon 6515P?

A: The largest Intel win is in Passmark floating-point math, where the Xeon scores 128954 versus the AMD EPYC 9184X’s 95476, a 26% advantage. The Xeon also leads in extended instructions (53383 vs 43562, 18.4% ahead) and Passmark single-thread (2855 vs 2822, 1.2% ahead).

Q: What is the largest benchmark win for the AMD EPYC 9184X?

A: The largest AMD win is in Passmark physics, where the EPYC 9184X scores 6674 against the Xeon’s 3829, a 74.3% margin. The AMD also wins Passmark data encryption by 22.6% and random string sorting by 24%.

Q: How does memory bandwidth differ between the two?

A: The AMD EPYC 9184X has a memory bandwidth of 460.8 GB/s over a twelve-channel DDR5 bus, while the Intel Xeon 6515P has 409.6 GB/s over an eight-channel DDR5 bus. The AMD part also has 128 PCIe Gen 5 lanes compared to the Intel’s 88.

Q: Which processor has more L3 cache?

A: The AMD EPYC 9184X has 768 MB of shared L3 cache, which is far larger than the Intel Xeon 6515P’s 72 MB of shared L3. The AMD part also has 1 MB of L2 per core versus the Intel’s 2 MB per core, and 64 KB of L1 per core versus the Intel’s 112 KB per core.

Head-to-Head Benchmarks

The Cinebench results are uniform: the AMD EPYC 9184X wins all six Cinebench tests by exactly 5.1%. In R15 multicore, the AMD scores 4083 against 3885; in R15 single-core, it is 576 against 548. The R20 tests show 17016 versus 16189 for multicore and 2401 versus 2285 for single-core. The R23 tests follow the same pattern—40515 versus 38547 multicore, and 5719 versus 5442 single-core. This consistency suggests a per-clock or per-thread advantage rather than a workload-specific one.

In Passmark, the AMD EPYC 9184X wins 9 of the 11 distinct tests. Data compression goes to AMD, 614873 versus 592645, a 3.8% edge. Data encryption is a larger win for AMD: 37376 versus 30476, a 22.6% margin. Find prime numbers favors AMD 465 to 385, a 20.8% lead. Integer math sees AMD at 157483 versus 147047, a 7.1% advantage. Multithread goes to AMD, 47665 versus 45350, a 5.1% gain. Physics is the standout: AMD’s 6674 beats the Xeon’s 3829 by 74.3%. Random string sorting also goes AMD’s way, 79913 versus 64425, a 24% margin.

The Intel Xeon 6515P’s Passmark wins are in floating-point math (128954 vs 95476, 26% ahead), extended instructions (53383 vs 43562, 18.4% ahead), and both single-thread and singlethread tests, where it scores 2855 versus 2822, a 1.2% margin. The extended instructions result is particularly notable because it shows the Intel part can outperform the AMD in vectorized code, even though it loses in most other Passmark subtests. The overall win count is 13 for AMD and 4 for Intel, but those 4 Intel wins are concentrated in areas that matter for specific high-performance computing niches.

Specification Differences

The core and thread counts are identical: both have 16 cores and 32 threads. The base clock differs significantly, with the AMD EPYC 9184X running at 3.55 GHz versus the Intel Xeon 6515P’s 2.30 GHz, and the boost clock also favors AMD, 4.20 GHz versus 3.80 GHz. The thermal design power is another major split: the AMD part is rated at 320 W, while the Intel part is rated at 150 W, meaning the Intel processor draws less than half the power budget of the AMD. The sockets are different—AMD uses Socket SP5, while Intel uses Socket 4710. The process nodes are both 5 nm, but the foundries differ: TSMC for AMD, Intel for the Xeon. The AMD part has 90,160 million transistors and a die size of 8x 72 mm², while the Intel part does not list transistor count or die size in the data. The L1, L2, and L3 caches all differ, as detailed above. Memory support is DDR5 for both, but the AMD part uses a twelve-channel bus versus the Intel’s eight-channel, with corresponding bandwidth differences. PCIe lane counts also differ: 128 for AMD versus 88 for Intel. The EPYC 9184X has a launch MSRP of $4928, while the Xeon 6515P has a launch MSRP of $740. The release dates are different, with the AMD launching in June 2023 and the Intel in February 2025. Both are in active production and do not have unlocked multipliers.