AMD EPYC 9535 vs Intel Xeon 6960P Comparison

FAQ

Q: How does the AMD EPYC 9535 compare to the Intel Xeon 6960P in average benchmark score?

A: The AMD EPYC 9535 has an average benchmark score of 379,408, while the Intel Xeon 6960P scores 365,194. The EPYC 9535 leads by 3.9% in this aggregate metric.

Q: Which processor has more cores and threads?

A: The Intel Xeon 6960P has 72 cores and 144 threads, while the AMD EPYC 9535 has 64 cores and 128 threads. The Xeon offers 8 more cores and 16 more threads.

Q: What is the single-thread performance difference?

A: The AMD EPYC 9535 scores 3,720 in PassMark single-thread testing, which is 13.2% ahead of the Xeon 6960P's 3,287. This is one of the few benchmarks where AMD wins.

Q: How large is the gap in physics processing?

A: The Intel Xeon 6960P achieves a PassMark physics score of 24,937, compared to the EPYC 9535's 3,834. This represents an 84.6% advantage for Intel, the largest delta in any head-to-head test.

Q: What are the launch MSRPs of these two processors?

A: The AMD EPYC 9535 has a launch MSRP of $8,992, and the Intel Xeon 6960P has a launch MSRP of $9,625.

Q: Which processor has higher memory bandwidth?

A: The Intel Xeon 6960P supports 614.4 GB/s of memory bandwidth, while the AMD EPYC 9535 supports 576.0 GB/s. Both use DDR5 with a twelve-channel memory bus.

Architecture Differences

The AMD EPYC 9535 is built on the Zen 5 architecture under the Turin codename, part of the EPYC 9005 series. It uses a 4 nm process node fabricated by TSMC, with a die size of 8x 70.6 mm² and 66,520 million transistors. The Intel Xeon 6960P belongs to the Granite Rapids family under the Xeon 6 (Granite Rapids-AP) generation. It uses a 5 nm process node from Intel's own foundry, with a larger die footprint of 3x 598 mm²; transistor count is not listed in the data.

Cache structures differ substantially. The EPYC 9535 provides 80 KB of L1 cache per core, 1 MB of L2 per core, and 256 MB of shared L3 cache. The Xeon 6960P offers larger per-core caches: 112 KB of L1 per core, 2 MB of L2 per core, and a much larger 432 MB of shared L3. This gives Intel a 68.75% advantage in total L3 capacity, which likely contributes to its strong performance in data-heavy workloads.

Socket and platform requirements diverge completely. The EPYC 9535 fits AMD Socket SP5, whereas the Xeon 6960P requires Intel Socket 7529. Both processors support DDR5 memory with twelve-channel buses and ECC memory, but the Xeon 6960P has higher peak memory bandwidth at 614.4 GB/s versus 576.0 GB/s for the EPYC. PCIe connectivity also differs: the EPYC 9535 offers 128 Gen 5 lanes (CPU only), while the Xeon 6960P provides 96 Gen 5 lanes. Neither processor includes integrated graphics.

Clock speeds favor AMD in boost but Intel in base. The EPYC 9535 runs at a 2.40 GHz base clock and boosts to 4.30 GHz, whereas the Xeon 6960P has a higher 2.70 GHz base but a lower 3.90 GHz boost. Thermal design power is notably different, with the EPYC 9535 rated at 300 W and the Xeon 6960P at 500 W. Release dates are close, with Intel shipping on 2024-09-23 and AMD on 2024-10-09.

The Verdict

The data presents a clear split between two distinct workload profiles. For users prioritizing raw multi-threaded throughput, the Intel Xeon 6960P is the stronger choice. It wins 8 of 11 head-to-head benchmarks, including all major data-processing tests, with particularly dominant margins in physics processing (84.6% ahead) and prime number finding (41.1% ahead). Its larger core count (72 vs 64), higher base clock (2.70 GHz vs 2.40 GHz), and substantially larger L3 cache (432 MB vs 256 MB) align with these results.

For workloads that depend on single-thread responsiveness or integer math, the AMD EPYC 9535 holds the advantage. It leads by 13.2% in single-thread performance and edges out the Xeon by a slim 0.3% in integer math. The EPYC also achieves a higher average benchmark score overall (379,408 vs 365,194), driven by its boost clock of 4.30 GHz and the efficiency of the Zen 5 architecture on a smaller 4 nm process.

The Xeon 6960P's higher TDP (500 W vs 300 W) and larger die area (3x 598 mm² vs 8x 70.6 mm²) suggest it trades power efficiency for raw compute capacity. The EPYC 9535, with lower TDP and higher boost clock, appears better suited for power-constrained environments or workloads that favor higher per-core frequency. Neither processor is universally superior; the choice depends entirely on whether the workload is dominated by parallel data processing or single-threaded latency.

Specification Differences

| Field | AMD EPYC 9535 | Intel Xeon 6960P |

|---|---|---|

| Cores | 64 | 72 |

| Threads | 128 | 144 |

| Base Clock | 2.40 GHz | 2.70 GHz |

| Boost Clock | 4.30 GHz | 3.90 GHz |

| TDP | 300 W | 500 W |

| Socket | AMD Socket SP5 | Intel Socket 7529 |

| Architecture | Zen 5 | Granite Rapids |

| Codename | Turin | Granite Rapids |

| Process Node | 4 nm | 5 nm |

| Foundry | TSMC | Intel |

| Die Size | 8x 70.6 mm² | 3x 598 mm² |

| Transistors | 66,520 million | Not listed |

| L1 Cache | 80 KB (per core) | 112 KB (per core) |

| L2 Cache | 1 MB (per core) | 2 MB (per core) |

| L3 Cache | 256 MB (shared) | 432 MB (shared) |

| Memory Bandwidth | 576.0 GB/s | 614.4 GB/s |

| PCIe Lanes | 128 (Gen 5) | 96 (Gen 5) |

| Launch MSRP | $8,992 | $9,625 |

| Part Number | 100-000001147 | SRPKX |

Identical features include DDR5 memory support, twelve-channel memory bus, ECC memory support, no integrated graphics, server/workstation market segment, active production status, and locked multipliers.

Head-to-Head Benchmarks

The Intel Xeon 6960P dominates the benchmark suite with decisive wins in data-heavy workloads. In PassMark data compression, Intel scores 2,797,724 against AMD's 2,308,822, a 17.5% advantage. Data encryption shows a similar pattern: Intel reaches 162,013 versus AMD's 127,372, a 21.4% gap. The Xeon also leads in extended instructions (193,404 vs 175,784, 9.1% ahead) and floating-point math (527,473 vs 457,047, 13.4% ahead).

The largest Intel victory comes in physics processing, where the Xeon 6960P scores 24,937 compared to the EPYC 9535's 3,834 — an 84.6% margin. Prime number finding also heavily favors Intel, with scores of 1,484 versus 874, a 41.1% difference. Random string sorting shows Intel ahead by 33.4% (371,795 vs 247,506), and multithread performance gives Intel a 12.3% edge (130,659 vs 114,528).

The AMD EPYC 9535 secures three wins, though only one is substantial. Its most notable victory is in single-thread performance, scoring 3,720 versus Intel's 3,287 — a 13.2% lead. In integer math, AMD narrowly edges out Intel with 730,281 versus 727,750, a mere 0.3% difference. The third win is a duplicate listing of the same single-thread result (3,720 vs 3,287, 13.2% ahead).

The aggregate picture is clear: Intel wins the majority of tests, but AMD's wins are concentrated in areas that matter for latency-sensitive or frequency-dependent workloads. The average benchmark scores reflect this mixed outcome, with AMD at 379,408 and Intel at 365,194.

Where Each One Wins

Intel Xeon 6960P wins in data processing and parallel compute. The data shows Intel ahead in compression (17.5%), encryption (21.4%), and random string sorting (33.4%). These are typical server workloads where the Xeon's 72 cores, 144 threads, and 432 MB L3 cache provide substantial throughput advantages. The physics benchmark (84.6% lead) and prime number finding (41.1% lead) further confirm Intel's strength in compute-heavy parallel tasks. For databases, analytics, scientific simulations, or any workload that can utilize all available cores and cache, the Xeon 6960P is the data-backed choice.

AMD EPYC 9535 wins in single-thread responsiveness and integer operations. The 13.2% single-thread advantage is significant for workloads that cannot be easily parallelized, such as certain database queries, legacy applications, or latency-sensitive services. The 0.3% integer math win, though marginal, adds another data point for integer-heavy workloads. The EPYC's higher boost clock (4.30 GHz vs 3.90 GHz) and lower TDP (300 W vs 500 W) make it attractive for environments where per-core performance and power efficiency are priorities. Its higher average benchmark score (379,408 vs 365,194) also indicates better overall balance across mixed workloads.

The practical split is by workload type. Choose the Intel Xeon 6960P for batch processing, data compression pipelines, encryption services, or any throughput-oriented deployment where the 8 additional cores and larger cache translate directly to faster completion times. Choose the AMD EPYC 9535 for mixed workloads that include substantial single-threaded components, or for deployments where the 200 W TDP difference matters for cooling and power infrastructure. The Xeon's higher launch MSRP ($9,625 vs $8,992) aligns with its larger core count, but the EPYC's superior single-thread and aggregate scores suggest it delivers competitive performance in a more power-efficient package.