AMD EPYC 9115 vs Intel Xeon 6724P Comparison

Where Each One Wins

The benchmark data splits these two server processors into distinct roles. The Intel Xeon 6724P wins 13 of the 17 recorded head-to-head tests, while the AMD EPYC 9115 takes 4. That is a lopsided overall count, but the AMD victories cluster in specific workloads that matter for certain server deployments.

Intel's wins are broad and consistent across rendering, physics simulation, encryption, and instruction-heavy tasks. The Xeon 6724P leads every Cinebench test by exactly 3.9%, regardless of whether the workload is single-core or multi-core. That uniformity suggests a clock-speed advantage that scales across the entire chip. The Intel part also dominates in floating-point math with an 18.9% lead, extended instructions with a 19% lead, prime number finding with a 28.7% lead, and physics with a 19.5% lead. These are large margins in compute-heavy server workloads.

The AMD EPYC 9115 wins in integer math with a 5.3% lead, random string sorting with a 2.4% lead, and single-thread performance with a 2.4% lead. The single-thread result is notable because Intel wins every Cinebench single-core test by the same 3.9% margin. The PassMark single-thread test tells a different story, giving AMD the edge. This is a reminder that single-threaded performance is workload-dependent, not a fixed property of the chip.

For system builders, the practical split is clear. The Intel Xeon 6724P is the choice for rendering farms, physics simulation, encryption-heavy services, and any workload that exercises extended instruction sets or floating-point math. The AMD EPYC 9115 fits better in integer-heavy database workloads, string processing pipelines, and certain single-threaded service applications. Both parts sit at the 94th percentile of all CPUs in the database, so either is a high-end server option.

FAQ

Q: Which processor has the higher average benchmark score?

A: The Intel Xeon 6724P has an average benchmark score of 72,396, while the AMD EPYC 9115 averages 69,288. That places Intel roughly 4.3% higher on average across the recorded tests.

Q: How does the Intel Xeon 6724P compare to its closest rivals?

A: The database lists the Intel Xeon 6517P as its nearest rival with an average score of 72,350, just 0.1% behind. The Intel Core Ultra 7 265KF is 0.7% behind, the AMD Ryzen 7 8840HX is 0.8% behind, and the Intel Xeon Platinum 8270 trails by 1.4%.

Q: What are the closest competitors to the AMD EPYC 9115?

A: The AMD EPYC 9115's nearest rival is the Intel Core i7-14700K with an average score of 69,355, which is 0.1% ahead of the EPYC. The AMD Ryzen 9 7950X is 0.3% ahead, the AMD Ryzen 9 7940HX is 0.8% ahead, and the AMD Ryzen 7 9700F is 1% ahead.

Q: Which processor wins in integer math?

A: The AMD EPYC 9115 wins PassMark integer math with a score of 181,807 compared to Intel's 172,216, a 5.3% advantage. This is one of only four tests where AMD comes out ahead.

Q: Which processor has the larger cache?

A: The Intel Xeon 6724P has 72 MB of shared L3 cache and 2 MB of L2 cache per core. The AMD EPYC 9115 has 64 MB of shared L3 cache and 1 MB of L2 cache per core. Intel also has a larger L1 cache at 112 KB per core versus 80 KB per core for AMD.

Q: How do the memory channels compare?

A: The AMD EPYC 9115 supports twelve memory channels with 576.0 GB/s of memory bandwidth, while the Intel Xeon 6724P supports eight channels with 409.6 GB/s. Both support DDR5 with ECC memory.

Head-to-Head Benchmarks

The head-to-head results show Intel leading by a consistent 3.9% across all six Cinebench tests. In Cinebench R23 multi-core, Intel scores 43,643 versus AMD's 42,003. In R23 single-core, Intel scores 6,161 versus 5,929. The same 3.9% delta appears in Cinebench R20 multi-core (18,330 versus 17,641), R20 single-core (2,587 versus 2,490), R15 multi-core (4,399 versus 4,233), and R15 single-core (620 versus 597). This consistency indicates the Intel advantage is architectural or clock-driven rather than workload-specific.

The PassMark suite reveals where AMD fights back. The largest Intel win in the entire comparison is PassMark find prime numbers, where Intel scores 372 versus AMD's 289, a 28.7% margin. This is a remarkable gap for two 16-core processors at the same market tier. PassMark physics also favors Intel heavily at 5,004 versus 4,188, a 19.5% lead. Floating-point math follows with 135,404 versus 113,853, an 18.9% gap. Extended instructions show a 19% lead for Intel with 54,101 versus 45,477.

AMD's wins are narrower. Integer math gives AMD 181,807 versus Intel's 172,216, a 5.3% advantage. Random string sorting goes to AMD with 70,151 versus 68,477, a 2.4% edge. PassMark single-thread and singlethread (the same test recorded twice) both give AMD 3,360 versus Intel's 3,279, also a 2.4% margin.

The remaining PassMark tests favor Intel by single-digit margins. Data compression goes to Intel with 627,185 versus 600,099, a 4.5% lead. Multithread goes to Intel with 51,345 versus 48,936, a 4.9% edge. Data encryption gives Intel 34,332 versus 33,489, a 2.5% lead.

The overall pattern suggests Intel wins more tests and wins by larger margins on average, but AMD's wins are concentrated in exactly the areas where modern server software spends the most time on integer operations. Integer-heavy analytics workloads will see modest AMD advantages, while everything else sees Intel pulling ahead by anywhere from 2.5% to 28.7%.

Specification Differences

The two processors share the same core count at 16 cores and 16 threads, but nearly everything else differs. Intel runs at a 3.60 GHz base clock and 4.30 GHz boost clock, while AMD runs at 12.60 GHz base and 4.10 GHz boost. Intel's higher clocks explain its dominance in clock-sensitive rendering tests, where it leads by 3.9% across the board.

Power draw differs substantially. Intel is rated at 3.60 GHz with a TDP of 210 watts, while AMD is rated at 2.60 GHz with a TDP of 125 watts. The Intel part uses more power to achieve its higher clock speeds and benchmark scores. Memory bandwidth also diverges sharply: AMD offers twelve-channel memory with 576.0 GB/s, while Intel offers eight-channel with 409.6 GB/s. The AMD part has a 40.6% memory bandwidth advantage despite losing most compute tests.

PCIe connectivity favors AMD as well. The EPYC 9115 provides Gen 5 with 128 lanes (CPU only), while the Xeon 6724P provides Gen 5 with 88 lanes (CPU only). That is a 45.5% lane count difference for systems needing many expansion cards or storage devices.

Both processors are unlocked multipliers set to false, target the Server/Workstation market segment, support DDR5 with ECC, have no integrated graphics, and are currently Active in production. The AMD part has a launch MSRP of $726, while the Intel part has a launch MSRP of $3622. The AMD EPYC 9115 was released on 2024-10-09, while the Intel Xeon 6724P was released later on 2025-02-23. The Intel part number is SRVUA, and the AMD part number is 100-000001552. AMD belongs to the EPYC 9005 series, while Intel has no listed series designation.

Architecture Differences

The two chips come from fundamentally different designs. Intel uses Granite Rapids architecture on a 5 nm process fabricated by Intel itself. AMD uses Zen 5 architecture on a 4 nm process fabricated by TSMC. The AMD process node is one step smaller, which contributes to its lower 125 watt TDP. AMD lists 16,630 million transistors across a die size of 2x 70.6 mm², while the Intel part does not have transistor count or die size recorded in the database.

Cache hierarchies differ in both size and organization. Intel provides 112 KB of L1 cache per core, 2 MB of L2 per core, and 72 MB of shared L3 cache. AMD provides 80 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3 cache. Intel leads in every cache tier, which contributes to its wins in latency-sensitive workloads like prime number finding and physics simulation. AMD's smaller per-core caches are offset by its higher memory bandwidth and lower power draw.

The socket platforms are entirely incompatible. Intel uses Socket 4710, while AMD uses AMD Socket SP5. Intel's Granite Rapids generation is labeled "Xeon 6 (Granite Rapids-SP)", while AMD's generation is labeled "EPYC (Zen 5 (Turin))". Neither processor has 3D V-Cache.

The architectural split is clear: Intel trades power efficiency and memory bandwidth for higher clocks and larger caches, while AMD trades clocks and cache size for lower power and twice the memory channels. Intel's 5 nm node from its own foundry cannot match the density of TSMC's 4 nm process, but Intel compensates with a 3.60 GHz base clock that is a full 1 GHz higher than AMD's 2.60 GHz base.

The Verdict

The data points to the Intel Xeon 6724P for compute-heavy workloads that depend on raw throughput. It wins every Cinebench test by 3.9%, leads in floating-point math by 18.9%, extended instructions by 19%, physics by 19.5%, and prime number finding by 28.7%. It wins 13 of 17 head-to-head tests and has a higher average benchmark score of 72,396 versus 69,288. Its higher clock speeds and larger caches give it a clear edge in rendering, simulation, and instruction-heavy server tasks.

The AMD EPYC 9115 is the better fit for memory-bandwidth-sensitive and integer-heavy workloads. It wins integer math by 5.3% and random string sorting by 2.4%. It offers twelve-channel memory at 576.0 GB/s versus Intel's eight-channel 409.6 GB/s, and it provides 128 PCIe Gen 5 lanes versus Intel's 88. Its 125 watt TDP is substantially lower than Intel's 210 watts, and its launch MSRP is $726. The AMD part also holds a small single-thread advantage in PassMark testing, scoring 3,360 versus Intel's 3,279.

For system integrators building general-purpose servers with mixed workloads, the Intel Xeon 6724P offers the higher average performance and wins more tests. For deployments that prioritize memory throughput, expansion capacity, or integer processing, the AMD EPYC 9115 provides those specific strengths. The choice depends on whether the workload leans toward Intel's compute dominance or AMD's memory and integer advantages.