AMD EPYC 4464P vs Intel Xeon 6515P Comparison

Head-to-Head Benchmarks

The benchmark data presents a clear split between the Intel Xeon 6515P and the AMD EPYC 4464P, with the AMD part winning 12 of the 17 recorded comparisons while the Intel part takes 5. The average benchmark score for the Xeon 6515P is 67006, placing it 0.1% ahead of the AMD EPYC 4465P and 0.2% ahead of the Intel Core Ultra 5 250K Plus, while trailing the Intel Core Ultra 9 275HX by 0.7% and the Intel Xeon w5-3525 by 1%. The EPYC 4464P records an average benchmark score of 64823, which puts it 0.3% ahead of the Intel Core Ultra 7 265, 0.6% ahead of the Intel Core Ultra 7 265F, and 1% ahead of the AMD EPYC 7343, while the AMD EPYC 9124 leads it by 0.4%.

The single-threaded results heavily favor the AMD EPYC 4464P. In Passmark single-thread, the AMD scores 4146 against the Intel's 2855, a 31.1% advantage. Cinebench R15 single-core shows the AMD at 572 versus 548, a 4.2% lead, and the same 4.2% delta appears in Cinebench R20 single-core with 2384 against 2285. Cinebench R23 single-core repeats the pattern at 5677 versus 5442, again a 4.1% gap. These consistent margins across all three Cinebench versions indicate a fundamental single-thread performance advantage for the AMD design.

Multi-threaded results tell a more nuanced story. The AMD EPYC 4464P leads in Cinebench R15 multicore with 4053 against 3885, a 4.1% margin, and in R20 multicore with 16890 versus 16189, a 4.2% lead. Cinebench R23 multicore shows 40215 for the AMD against 38547 for the Intel, a 4.1% advantage. Passmark multithread also goes to the AMD at 47514 versus 45350, a 4.6% lead. The Intel Xeon 6515P, despite having more cores and threads, cannot overcome the AMD's per-thread efficiency in these workloads.

The Intel Xeon 6515P dominates in several specialized Passmark workloads. Floating point math shows the Intel at 128954 against 93090 for the AMD, a 38.5% advantage, the largest win recorded for either processor. Extended instructions favor the Intel by 35.6% with 53383 versus 39359. Physics results give the Intel a 33.3% lead at 3829 versus 2873. Prime number finding goes to the Intel at 385 versus 343, a 12.2% margin, and data compression favors the Intel at 592645 versus 574304, a 3.2% advantage.

The AMD EPYC 4464P counters with wins in data encryption, integer math, and random string sorting. Data encryption shows the AMD at 35816 versus 30476, a 14.9% lead. Integer math favors the AMD at 160410 versus 147047, an 8.3% margin, and random string sorting goes to the AMD at 70200 versus 64425, an 8.2% advantage. These results suggest the AMD processor handles encryption workloads and integer-heavy tasks more efficiently, while the Intel part excels at floating-point calculations and extended instruction sets.

Architecture Differences

The two processors come from fundamentally different design philosophies. The Intel Xeon 6515P uses the Granite Rapids architecture, built on a 5 nm process at Intel's own foundry. It packs 16 cores and 32 threads, with a base clock of 2.30 GHz and a boost clock of 3.80 GHz. The AMD EPYC 4464P belongs to the EPYC 4004 series, uses the Zen 4 architecture codenamed Raphael, and is manufactured by TSMC on a 5 nm node. It provides 12 cores and 24 threads, with a base clock of 3.70 GHz and a boost clock of 5.40 GHz.

Cache hierarchies differ substantially. The Intel part allocates 112 KB of L1 cache per core, 2 MB of L2 cache per core, and 72 MB of shared L3 cache. The AMD part uses 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3 cache. The Intel processor therefore offers more cache per core at every level, which contributes to its strengths in floating-point and physics workloads. The AMD processor counters with higher clock speeds, both at base and boost, which explains its single-thread dominance.

Memory support presents another major divergence. The Intel Xeon 6515P uses eight-channel DDR5 memory with a theoretical bandwidth of 409.6 GB/s, while the AMD EPYC 4464P uses dual-channel DDR5 with 83.2 GB/s. This fivefold difference in memory bandwidth is substantial, though the benchmark results show the Intel part winning only data compression among memory-sensitive workloads. Both processors support ECC memory. The Intel part supplies 88 PCIe Gen 5 lanes from the CPU, while the AMD part provides 28 PCIe Gen 5 lanes. The Intel processor has no integrated graphics, whereas the AMD part includes Radeon Graphics.

The transistor count and die size for the AMD EPYC 4464P are recorded as 13,140 million transistors across 2x 71 mm² dies. No transistor count or die size is listed for the Intel Xeon 6515P. The Intel part has a TDP of 150 watts compared to 65 watts for the AMD part. The Intel processor uses Intel Socket 4710, while the AMD processor uses AMD Socket AM5. Neither processor has an unlocked multiplier. The Intel Xeon 6515P was released on 2025-02-23 with a launch MSRP of $740. The AMD EPYC 4464P was released on 2024-05-20 with a launch MSRP of $429. Both are currently active in production, and both target the server and workstation market segment.

Where Each One Wins

The AMD EPYC 4464P is the clear choice for single-threaded and lightly threaded workloads. Its 31.1% lead in Passmark single-thread and consistent 4.1% to 4.2% advantages across all Cinebench single-core tests make it the stronger option for applications that depend on per-core performance. The data also shows the AMD part winning in integer math, random string sorting, and data encryption, indicating strength in general-purpose compute, text processing, and cryptographic workloads. The multithreaded Cinebench results, despite the Intel part having 16 cores against 12, also go to the AMD, which suggests the higher clock speeds and architectural efficiency carry more weight than raw core count in these rendering workloads.

The Intel Xeon 6515P claims the specialized compute workloads. The 38.5% lead in floating-point math is decisive, and the 35.6% advantage in extended instructions points to strength in vectorized and advanced instruction set workloads. The 33.3% lead in physics tests further reinforces this pattern. Data compression goes to the Intel part by a modest 3.2%, and prime number finding sees a 12.2% Intel advantage. For users running scientific simulations, physics engines, compression pipelines, or heavy floating-point numerical code, the Intel processor provides measurably better results.

The two processors also occupy different power envelopes, with the Intel part rated at 150 watts TDP against 65 watts for the AMD part. The Intel processor offers eight-channel memory with 409.6 GB/s bandwidth and 88 PCIe Gen 5 lanes, making it the stronger platform for memory-hungry and I/O-intensive server configurations. The AMD processor provides 28 PCIe Gen 5 lanes and dual-channel memory at 83.2 GB/s. Neither processor is multiplier unlocked, so overclocking is not part of the equation.

FAQ

Q: Which processor has the higher single-thread performance?

A: The AMD EPYC 4464P leads in every single-thread benchmark. Passmark single-thread shows 4146 versus 2855, a 31.1% advantage. Cinebench R15, R20, and R23 single-core tests all show the AMD ahead by 4.1% to 4.2%.

Q: How do the two processors compare in multi-threaded workloads?

A: The AMD EPYC 4464P wins Cinebench R15 multicore with 4053 against 3885, R20 multicore with 16890 against 16189, R23 multicore with 40215 against 38547, and Passmark multithread with 47514 against 45350. The margins range from 4.1% to 4.6%.

Q: Where does the Intel Xeon 6515P have its biggest advantages?

A: The Intel part leads in floating-point math by 38.5% (128954 versus 93090), extended instructions by 35.6% (53383 versus 39359), physics by 33.3% (3829 versus 2873), prime numbers by 12.2% (385 versus 343), and data compression by 3.2% (592645 versus 574304).

Q: What are the core and thread counts for each processor?

A: The Intel Xeon 6515P has 16 cores and 32 threads. The AMD EPYC 4464P has 12 cores and 24 threads.

Q: How does memory bandwidth differ between the two?

A: The Intel Xeon 6515P uses eight-channel DDR5 with 409.6 GB/s bandwidth. The AMD EPYC 4464P uses dual-channel DDR5 with 83.2 GB/s bandwidth.

Q: What is the cache configuration on each processor?

A: The Intel Xeon 6515P has 112 KB L1 per core, 2 MB L2 per core, and 72 MB shared L3. The AMD EPYC 4464P has 64 KB L1 per core, 1 MB L2 per core, and 64 MB shared L3.

The Verdict

The benchmark data supports a straightforward recommendation. The AMD EPYC 4464P should be the default choice for users whose workloads are dominated by single-threaded performance, general integer compute, encryption, or the Cinebench-style rendering workloads. It wins 12 of 17 head-to-head comparisons, including all Cinebench tests and Passmark multithread, and does so while operating at 65 watts TDP compared to 150 watts for the Intel part. Its 5.40 GHz boost clock and 3.70 GHz base clock give it a substantial clock advantage that shows up clearly in the 31.1% single-thread lead.

The Intel Xeon 6515P is the pick for specialized compute environments. The 38.5% floating-point advantage and 35.6% extended instructions lead are decisive for numerical workloads, and the 33.3% physics lead matters for simulation and similar applications. The 16-core, 32-thread configuration combined with eight-channel memory at 409.6 GB/s and 88 PCIe Gen 5 lanes makes it the stronger platform for memory-bandwidth-intensive server deployments, even though the benchmark averages put it only 0.1% ahead of the AMD EPYC 4465P and 1% behind the Intel Xeon w5-3525.

Both processors sit at the 93rd percentile against all CPUs, meaning either choice places the user in the top tier of available processors. The AMD EPYC 4464P offers the better all-around benchmark profile with more wins and a lower TDP. The Intel Xeon 6515P offers specific workload advantages that may outweigh its losses for users with the right application mix. The data does not support a single universal winner; it supports matching the processor to the workload.