CPU Comparison

The Intel Xeon 6736P and AMD EPYC 9135 are both 96th-percentile server processors, yet they achieve that status through fundamentally different strategies. The EPYC 9135 wins the majority of head-to-head tests (10 vs 7), but the Xeon 6736P dominates in several specialized workloads. The data reveals a clear split: the AMD part leads in general compute and single-threaded performance, while the Intel part excels in data handling, encryption, and prime-number workloads.

Where Each One Wins

The AMD EPYC 9135 takes the overall win count, securing victories in all six Cinebench tests (R15, R20, R23, both single and multi-core) plus PassMark's multithread, single-thread, and extended instructions tests. Its Cinebench leads are consistent at a 13.4% margin across every iteration, indicating a uniform architectural advantage rather than a workload-specific quirk. The single-thread PassMark result is particularly decisive: the EPYC scores 3672 versus 2024, a 44.9% gap that dwarfs any other difference between these two parts.

The Intel Xeon 6736P counters with seven wins, concentrated in data-centric and math-heavy workloads. Its biggest victory is in PassMark's find prime numbers test, where it scores 392 against 292, a 34.2% advantage. It also takes floating point math (149770 vs 126679, +18.2%), physics (6531 vs 5477, +19.2%), random string sorting (103723 vs 90064, +15.2%), and data encryption (46236 vs 40295, +14.7%). The Xeon's lead in data compression is narrower at 7.8% (796658 vs 739277), and its integer math edge is marginal at 1.9% (206833 vs 202962).

This split suggests that the EPYC 9135 is the better general-purpose compute engine, while the Xeon 6736P has specialized strengths in tasks involving data manipulation, cryptographic operations, and certain mathematical calculations. The Xeon's 36 cores and 72 threads give it a raw throughput advantage in parallel workloads that favor core count, but the EPYC's Zen 5 architecture delivers higher per-thread performance that translates to wins in most standard benchmarks.

The Verdict

For users running general server workloads, rendering, or compilation tasks, the AMD EPYC 9135 is the clear choice based on benchmark results. It wins every Cinebench test by 13.4%, including multi-core R23 where it scores 49136 versus 42561, and its PassMark multithread score of 57170 beats the Xeon's 50072 by 12.4%. The EPYC also offers substantially better single-thread performance, which matters for database queries, web serving, and any latency-sensitive application.

The Intel Xeon 6736P is the pick for workloads that match its specific strengths. If the application involves heavy encryption, data compression, random string sorting, or prime number calculations, the Xeon's advantages range from 7.8% to 34.2%. Its floating point math lead of 18.2% and physics score advantage of 19.2% also make it attractive for scientific computing and simulation tasks that rely on those operations.

The overall percentile ranking is identical at 96 for both parts, but the average benchmark scores tell a different story: the Xeon averages 87864, which is 5.9% higher than the EPYC's 82980. This discrepancy arises because the Xeon's wins in data-heavy tests produce large absolute scores (like 796658 in compression) that boost its average, even though it loses more individual tests. Users should therefore weight the specific benchmark suite that matches their intended workloads rather than relying on the aggregate score.

Head-to-Head Benchmarks

The most striking result is the PassMark single-thread test, where the EPYC 9135 scores 3672 against the Xeon's 2024, a 44.9% advantage that is the largest margin in any comparison. This reflects the EPYC's higher boost clock of 4.30 GHz versus 4.10 GHz, combined with its newer Zen 5 architecture. The Cinebench single-core tests confirm this pattern with a consistent 13.4% EPYC lead across R15 (699 vs 605), R20 (2913 vs 2523), and R23 (6936 vs 6008).

In multi-core workloads, the EPYC maintains its edge despite having 20 fewer cores. The Cinebench R23 multicore result shows 49136 for the EPYC versus 42561 for the Xeon, a 13.4% difference. The PassMark multithread test shows a narrower but still decisive 12.4% EPYC lead (57170 vs 50072). This is remarkable given the Xeon's 36 cores and 72 threads versus the EPYC's 16 cores and 32 threads, the EPYC's core efficiency more than compensates for its lower core count.

The Xeon's biggest win is in find prime numbers, where it scores 392 versus 292, a 34.2% advantage. This is followed by physics (6531 vs 5477, +19.2%) and floating point math (149770 vs 126679, +18.2%). The Xeon also leads in random string sorting by 15.2% (103723 vs 90064) and data encryption by 14.7% (46236 vs 40295). Its data compression win is more modest at 7.8% (796658 vs 739277), and integer math is nearly a tie at 1.9% (206833 vs 202962). The extended instructions test is essentially a dead heat, with the EPYC winning by just 0.5% (55822 vs 55563).

FAQ

Q: Which processor has higher single-thread performance?

A: The AMD EPYC 9135 wins decisively. PassMark single-thread shows 3672 versus 2024, a 44.9% advantage, and Cinebench R23 single-core shows 6936 versus 6008, a 13.4% lead.

Q: Does the Intel Xeon 6736P win any benchmark by a large margin?

A: Yes. Its largest win is in PassMark find prime numbers, scoring 392 versus 292, a 34.2% advantage. It also leads by 19.2% in physics and 18.2% in floating point math.

Q: How do the core counts compare?

A: The Intel Xeon 6736P has 36 cores and 72 threads, while the AMD EPYC 9135 has 16 cores and 32 threads. Despite having fewer than half the cores, the EPYC wins most multi-threaded benchmarks.

Q: What is the memory bandwidth difference?

A: The AMD EPYC 9135 offers 576.0 GB/s across a twelve-channel memory bus, while the Intel Xeon 6736P provides 409.6 GB/s over an eight-channel bus. The EPYC has a 40.6% bandwidth advantage.

Q: Which processor has a higher boost clock?

A: The AMD EPYC 9135 boosts to 4.30 GHz, compared to 4.10 GHz for the Intel Xeon 6736P. The EPYC also has a higher base clock at 3.65 GHz versus 2.00 GHz.

Q: Are there any benchmarks where the two processors are nearly identical?

A: PassMark extended instructions is the closest, with the AMD EPYC 9135 winning 55822 versus 55563, a 0.5% margin. Integer math is also close, with the Intel Xeon 6736P leading 206833 versus 202962, a 1.9% difference.

Architecture Differences

The two processors come from fundamentally different design philosophies. The Intel Xeon 6736P uses the Granite Rapids architecture on a 5 nm process node fabricated by Intel, with a die size of 598 mm². The AMD EPYC 9135 is built on Zen 5 (codenamed Turin) using TSMC's 4 nm process, with a die size of 2x 70.6 mm² and 16,630 million transistors. The EPYC's smaller, dual-die design with a more advanced process node contributes to its efficiency advantage.

Cache configurations differ significantly. The Xeon 6736P provides 112 KB of L1 per core, 2 MB of L2 per core, and a large 144 MB shared L3 cache. The EPYC 9135 has 80 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3, less than half the Xeon's L3 capacity. Despite this, the EPYC's higher clock speeds and architectural efficiency allow it to outperform the Xeon in most tests.

Both processors support DDR5 memory and ECC, with no integrated graphics. The Xeon has 88 PCIe Gen 5 lanes, while the EPYC offers 128 PCIe Gen 5 lanes, a 45% advantage in expansion capacity. The EPYC's twelve-channel memory bus versus the Xeon's eight-channel provides 576.0 GB/s versus 409.6 GB/s in bandwidth. Both are active production parts, with the Xeon releasing on 2025-02-23 and the EPYC on 2024-10-09.

Specification Differences

| Specification | Intel Xeon 6736P | AMD EPYC 9135 |

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

| Cores | 36 | 16 |

| Threads | 72 | 32 |

| Base Clock | 2.00 GHz | 3.65 GHz |

| Boost Clock | 4.10 GHz | 4.30 GHz |

| TDP | 205 W | 200 W |

| Socket | Intel Socket 4710 | AMD Socket SP5 |

| Process Node | 5 nm | 4 nm |

| Foundry | Intel | TSMC |

| Die Size | 598 mm² | 2x 70.6 mm² |

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

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

| L3 Cache | 144 MB (shared) | 64 MB (shared) |

| Memory Bus | Eight-channel | Twelve-channel |

| Memory Bandwidth | 409.6 GB/s | 576.0 GB/s |

| PCIe | Gen 5, 88 Lanes | Gen 5, 128 Lanes |

| Launch MSRP | $3351 | $1214 |

| Part Number | SRVNW | 100-000001150 |

The EPYC 9135 has a higher base clock by 1.65 GHz and a higher boost clock by 0.20 GHz, while consuming 5 W less power. Its transistor count of 16,630 million and smaller die area indicate a more dense and efficient design. The Xeon 6736P compensates with a much larger L3 cache (144 MB vs 64 MB) and double the core count. The launch MSRP of the EPYC is $1214, while the Xeon is $3351.