CPU Comparison

The Intel Xeon 696X and AMD EPYC 9684X are both 99th-percentile server processors, but the benchmark data draws a stark line between them. The AMD EPYC 9684X wins 12 of the 14 head-to-head comparisons, often by wide margins, while the Intel Xeon 696X secures a decisive win in single-threaded performance. This is a matchup between Intel’s highest-clocking Granite Rapids part and AMD’s massive-cache Genoa-X flagship, and the data shows that the EPYC 9684X is the dominant multi-core performer while the Xeon 696X is the clear choice for lightly threaded workloads. The average benchmark scores reinforce this: the Xeon 696X posts an average of 286,102, while the EPYC 9684X averages 266,914, yet that aggregate number hides the EPYC’s crushing wins in nearly every compute-heavy test.

Where Each One Wins

The AMD EPYC 9684X wins decisively in every multi-threaded and throughput-oriented benchmark listed. In Cinebench R15, R20, and R23 multi-core tests, the EPYC 9684X scores 10,418, 43,409, and 103,355 respectively, versus the Xeon 696X’s 8,994, 37,475, and 89,227. That is a consistent 13.7% advantage across all three rendering workloads. The EPYC 9684X also dominates in PassMark’s integer math (844,145 vs 572,072, a 32.2% lead), data encryption (178,453 vs 112,529, a 36.9% lead), and random string sorting (349,126 vs 180,392, a 48.3% lead). The largest single win comes in PassMark physics, where the EPYC 9684X scores 24,686 against the Xeon 696X’s 3,382, an 86.3% gap that indicates a fundamental advantage in simulation workloads.

The Intel Xeon 696X wins only in single-threaded performance. Its PassMark single-thread score of 3,742 beats the EPYC 9684X’s 2,891 by 29.4%. This is the sole bright spot for Intel in the head-to-head data, but it is a meaningful one for workloads that scale poorly across cores. The Xeon 696X also carries a higher boost clock of 4.80 GHz versus the EPYC 9684X’s 3.70 GHz, which explains the single-thread edge. For database queries, legacy application code, or any serialized task, the Xeon 696X is the faster processor. However, the data shows that nearly every modern server workload, rendering, encryption, compression, integer math, physics simulation, belongs to the EPYC 9684X.

Architecture Differences

The two processors take fundamentally different approaches to achieving performance. The Intel Xeon 696X uses the Granite Rapids architecture on a 5 nm process fabricated by Intel, with a die size of 2x 598 mm². It packs 64 cores and 128 threads, with a base clock of 2.40 GHz and a boost clock of 4.80 GHz. Its cache hierarchy includes 112 KB of L1 per core, 2 MB of L2 per core, and 336 MB of shared L3 cache. Memory support is DDR5 over an eight-channel bus with 409.6 GB/s of bandwidth. The chip is unlocked, which is unusual for a server part, and it uses Intel Socket 4710.

The AMD EPYC 9684X uses the Zen 4 architecture on a 5 nm process fabricated by TSMC, with 135,240 million transistors spread across 12 dies of 72 mm² each. It offers 96 cores and 192 threads, with a base clock of 2.55 GHz and a boost clock of 3.70 GHz. The defining feature is its cache: 1,152 MB of shared L3 cache, more than three times the Xeon 696X’s 336 MB. This massive L3 is the hallmark of the Genoa-X line, and it explains the EPYC’s dominance in data-heavy workloads. The EPYC 9684X also has a wider memory bus: twelve channels of DDR5 delivering 460.8 GB/s, compared to Intel’s eight channels at 409.6 GB/s. Both CPUs support ECC memory and provide Gen 5 PCIe with 128 lanes from the CPU, but the EPYC 9684X is locked, while the Xeon 696X allows multiplier adjustments.

The core count difference is stark: 96 vs 64 cores, and 192 vs 128 threads. That 50% core advantage is the primary driver of the EPYC’s multi-core wins. The Xeon 696X compensates with a 1.10 GHz higher boost clock, but the data shows that clock speed alone cannot overcome the EPYC’s combination of more cores, more cache, and higher memory bandwidth.

Head-to-Head Benchmarks

The biggest wins for the AMD EPYC 9684X come in specialized workloads. PassMark physics shows an 86.3% advantage (24,686 vs 3,382), which is the largest delta in the entire comparison. This suggests the EPYC’s large L3 cache is exceptionally effective for physics simulation, where working sets often exceed smaller caches. PassMark find prime numbers follows with a 57.8% lead (2,020 vs 853), indicating a strong advantage in integer-heavy, branch-predictable loops. Random string sorting shows a 48.3% gap (349,126 vs 180,392), and data encryption shows a 36.9% gap (178,453 vs 112,529). These are not marginal victories, they represent workloads where the EPYC 9684X is in a different performance class.

The Intel Xeon 696X’s sole victory is PassMark single-thread, where it scores 3,742 against the EPYC 9684X’s 2,891, a 29.4% advantage. This is the largest win for either processor in percentage terms, and it underscores the Xeon’s superior per-core clock speed. The Xeon 696X also comes closer in PassMark extended instructions (172,975 vs 177,956, a 2.8% deficit) and floating point math (450,164 vs 472,174, a 4.7% deficit), showing that the gap narrows in vectorized and floating-point workloads. However, in integer math the EPYC extends its lead to 32.2% (844,145 vs 572,072), and in data compression it leads by 16.1% (2,698,807 vs 2,264,907).

Cinebench results are uniform across all three versions: R15, R20, and R23 each show the EPYC 9684X ahead by exactly 13.7%. This consistency suggests that the performance difference scales linearly with multi-core rendering workloads, likely reflecting the core count advantage rather than any architectural quirk. PassMark multithread confirms the pattern with a 13.7% lead (121,595 vs 104,974).

FAQ

Q: Which processor has more cores and threads?

A: The AMD EPYC 9684X has 96 cores and 192 threads, while the Intel Xeon 696X has 64 cores and 128 threads.

Q: Which processor has the higher boost clock?

A: The Intel Xeon 696X boosts to 4.80 GHz, while the AMD EPYC 9684X boosts to 3.70 GHz.

Q: What is the L3 cache difference between the two?

A: The AMD EPYC 9684X has 1,152 MB of shared L3 cache, compared to 336 MB on the Intel Xeon 696X.

Q: Which processor wins in single-threaded benchmarks?

A: The Intel Xeon 696X wins PassMark single-thread with a score of 3,742, beating the AMD EPYC 9684X’s 2,891 by 29.4%.

Q: Which processor wins in Cinebench multi-core tests?

A: The AMD EPYC 9684X wins all three Cinebench multi-core tests (R15, R20, R23) by 13.7% each.

Q: Do both processors support ECC memory?

A: Yes, both the Intel Xeon 696X and the AMD EPYC 9684X support ECC memory.

Specification Differences

| Specification | Intel Xeon 696X | AMD EPYC 9684X |

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

| Cores | 64 | 96 |

| Threads | 128 | 192 |

| Base Clock | 2.40 GHz | 2.55 GHz |

| Boost Clock | 4.80 GHz | 3.70 GHz |

| TDP | 350 W | 400 W |

| Socket | Intel Socket 4710 | AMD Socket SP5 |

| Architecture | Granite Rapids | Zen 4 |

| Foundry | Intel | TSMC |

| Transistors | Not specified | 135,240 million |

| Die Size | 2x 598 mm² | 12x 72 mm² |

| L1 Cache | 112 KB (per core) | 64 KB (per core) |

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

| L3 Cache | 336 MB (shared) | 1,152 MB (shared) |

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

| Memory Bandwidth | 409.6 GB/s | 460.8 GB/s |

| Multiplier Unlocked | Yes | No |

| Launch MSRP | $5599 | $14756 |

| Release Date | 2026-02-01 | 2023-06-12 |

The specification table shows that the EPYC 9684X leads in core count, cache, memory channels, and bandwidth, while the Xeon 696X counters with higher clocks, lower TDP, and an unlocked multiplier. The Xeon 696X also has a significantly lower launch MSRP of $5599 versus the EPYC 9684X’s $14756, but the performance data indicates the EPYC’s higher price does not translate to a higher average benchmark score, the Xeon 696X averages 286,102 versus 266,914 for the EPYC 9684X.

The Verdict

The data is unambiguous for multi-threaded server workloads: the AMD EPYC 9684X is the superior processor. It wins 12 of 14 head-to-head benchmarks, with margins ranging from 2.8% to 86.3%. Its 96 cores, 1,152 MB of L3 cache, and twelve-channel memory bus make it the choice for rendering, encryption, compression, physics simulation, and integer math. The 13.7% lead across all Cinebench versions indicates a consistent advantage in content creation and simulation tasks. If the workload is parallel and cache-hungry, the EPYC 9684X is the clear winner.

The Intel Xeon 696X is the pick only when single-threaded performance is the priority. Its 29.4% lead in PassMark single-thread and 4.80 GHz boost clock make it the faster processor for serialized tasks, legacy applications, or workloads with poor core scaling. It also runs at a lower TDP (350 W vs 400 W) and offers an unlocked multiplier, which may appeal to certain workstation users. But those advantages are narrow: the Xeon 696X loses to the EPYC 9684X even in floating-point math (450,164 vs 472,174) and extended instructions (172,975 vs 177,956), where its clock advantage might have been expected to help.

The average benchmark scores tell the full story. The Xeon 696X posts a higher average (286,102) than the EPYC 9684X (266,914), but this is misleading because the EPYC’s wins are concentrated in the most demanding workloads. The Xeon 696X’s nearest rivals include the AMD EPYC 9565 (deltaPct 0.2) and AMD EPYC 9555P (deltaPct -0.3), while the EPYC 9684X sits relative to the Intel Xeon 6980P (deltaPct 6.1) and AMD EPYC 9565 (deltaPct -6.5). In absolute terms, the EPYC 9684X is the more powerful chip for compute-heavy server deployments. The Xeon 696X is the better choice for single-thread-bound environments where clock speed matters more than core count.