AMD EPYC 9684X vs Intel Xeon 6980P Comparison

The AMD EPYC 9684X and Intel Xeon 6980P represent two distinct approaches to high-core-count server computing. The data shows a decisive victory for the AMD part in most workloads, with the Intel chip winning only in specific, narrow scenarios. Benchmark results indicate that the EPYC 9684X is the dominant performer for general compute, while the Xeon 6980P offers a different set of trade-offs, particularly in memory bandwidth and specific math operations.

Head-to-Head Benchmarks

The most striking pattern in the head-to-head data is the sheer scale of AMD’s victory in Cinebench tests. Across all six Cinebench R15, R20, and R23 runs, the EPYC 9684X wins by a consistent 63.6% margin, regardless of single-core or multi-core load. For instance, in Cinebench R23 multi-core, the AMD chip scores 103,355 against the Intel’s 63,175. The single-core scores tell the same story: 14,591 versus 8,918, a 63.6% gap. This consistency suggests a fundamental architectural advantage in rendering and CPU-bound tasks, not a workload-specific quirk.

The PassMark suite reveals a more nuanced picture. In PassMark multi-thread, the AMD EPYC 9684X again wins by 63.6%, scoring 121,595 versus 74,324. The AMD part also dominates in integer math, posting 844,145 versus 637,476, a 32.4% lead. Even more lopsided is the find prime numbers test, where the EPYC 9684X scores 2,020 against the Xeon’s 555, a 264% advantage. The physics test is the largest single gap: 24,686 versus 3,350, a 636.9% difference. In data encryption, the AMD chip leads 178,453 to 125,246, a 42.5% margin.

Intel’s two wins are worth examining. The Xeon 6980P takes PassMark extended instructions with a score of 214,794 versus 177,956, a 17.2% lead. This indicates a more capable SIMD or specialized instruction set implementation. The Xeon also wins in floating-point math, scoring 501,720 versus 472,174, a 5.9% margin. These wins are narrow compared to AMD’s, but they point to areas where Intel’s design holds a specific edge. Overall, the AMD EPYC 9684X wins 15 of the 17 head-to-head benchmarks.

FAQ

Q: Which CPU has more cores and threads?

A: The Intel Xeon 6980P has 128 cores and 256 threads, while the AMD EPYC 9684X has 96 cores and 192 threads. Despite Intel’s 33% core advantage, the AMD chip wins most multi-threaded benchmarks.

Q: Does the AMD EPYC 9684X have a higher boost clock?

A: No. The Intel Xeon 6980P boosts to 3.90 GHz, while the AMD EPYC 9684X boosts to 3.70 GHz. The AMD chip’s base clock is higher at 2.55 GHz versus 2.00 GHz for Intel.

Q: How much L3 cache does each processor have?

A: The AMD EPYC 9684X has 1152 MB of shared L3 cache. The Intel Xeon 6980P has 504 MB of shared L3 cache. AMD’s cache is more than double Intel’s.

Q: What is the memory bandwidth difference?

A: The Intel Xeon 6980P supports 614.4 GB/s of memory bandwidth, whereas the AMD EPYC 9684X supports 460.8 GB/s. Both use DDR5 memory with a twelve-channel bus.

Q: Which CPU has a higher average benchmark score?

A: The AMD EPYC 9684X has an average benchmark score of 266,914, which is 6.1% higher than the Intel Xeon 6980P’s average of 251,516.

Q: Are these CPUs unlocked for overclocking?

A: No. Both the AMD EPYC 9684X and the Intel Xeon 6980P have a locked multiplier, meaning they are not unlocked for overclocking.

Architecture Differences

The two processors are built on fundamentally different architectures, despite both using a 5 nm process node. The AMD EPYC 9684X uses the Zen 4 architecture with the “Genoa-X” codename, manufactured by TSMC. It features 135,240 million transistors across a die size of 12x 72 mm². The Intel Xeon 6980P uses the Granite Rapids architecture, manufactured by Intel itself, with a die size of 3x 598 mm². Intel does not list transistor count in the data.

Cache hierarchies diverge sharply. AMD’s L1 cache is 64 KB per core, and L2 is 1 MB per core. Intel’s L1 is 112 KB per core, and L2 is 2 MB per core. The major difference is L3: AMD provides 1152 MB shared, while Intel provides 504 MB shared. This 648 MB difference likely explains AMD’s dominance in data-heavy workloads like encryption and integer math. Intel’s smaller per-core L1 and L2 caches are offset by a higher per-core size, but the aggregate L3 advantage for AMD is substantial.

Memory and I/O also differ. Both use DDR5 with twelve-channel memory, but Intel’s bandwidth is higher at 614.4 GB/s versus 460.8 GB/s. PCIe lanes are another split: AMD offers Gen 5 with 128 lanes (CPU only), while Intel offers Gen 5 with 96 lanes (CPU only). The Intel part lists integrated graphics as “N/A,” while AMD does not list any integrated graphics. The sockets are incompatible: AMD uses Socket SP5, and Intel uses Socket 7529.

The Verdict

The data is unambiguous: the AMD EPYC 9684X is the superior processor for almost every measured workload. It wins 15 out of 17 head-to-head benchmarks, including all Cinebench tests, all but one PassMark math test, and the multi-thread and single-thread tests. The 63.6% margin in Cinebench and multi-thread tests is not a small edge; it is a generational gap. For users running rendering, physics simulations, or integer-heavy tasks, the EPYC 9684X is the clear choice.

The Intel Xeon 6980P wins only in PassMark extended instructions and floating-point math. These are niche wins. The 17.2% lead in extended instructions suggests better handling of specialized vector or cryptographic instruction sets. The 5.9% lead in floating-point math is marginal. For any workload that relies on those specific operations, Intel has a slight edge. However, the Xeon’s average benchmark score is 5.8% lower than the EPYC 9684X, and it loses to the AMD chip in its own nearestRivals list by that margin.

The Intel Xeon 6980P does offer more cores (128 versus 96) and higher memory bandwidth (614.4 GB/s versus 460.8 GB/s), but these specifications do not translate into benchmark wins. The data suggests that AMD’s larger L3 cache and higher base clock are more impactful than Intel’s core count and memory throughput. For a server or workstation builder prioritizing raw compute performance, the EPYC 9684X is the data-backed recommendation.

Specification Differences

The two CPUs differ in nearly every major specification. The AMD EPYC 9684X has 96 cores and 192 threads, while the Intel Xeon 6980P has 128 cores and 256 threads. AMD’s base clock is 2.55 GHz, higher than Intel’s 2.00 GHz. Intel’s boost clock is 3.90 GHz, higher than AMD’s 3.70 GHz. Thermal design power (TDP) is 400 watts for AMD and 500 watts for Intel.

Cache sizes differ significantly. AMD has 64 KB L1 per core and 1 MB L2 per core, while Intel has 112 KB L1 per core and 2 MB L2 per core. The L3 cache is the largest differentiator: 1152 MB for AMD versus 504 MB for Intel. Memory support is DDR5 for both, with a twelve-channel bus, but bandwidth is 460.8 GB/s for AMD and 614.4 GB/s for Intel. PCIe lanes are 128 for AMD and 96 for Intel, both Gen 5.

The sockets are different: AMD uses Socket SP5, Intel uses Socket 7529. Process node is 5 nm for both, but the foundry differs: TSMC for AMD, Intel for Intel. The die size is 12x 72 mm² for AMD and 3x 598 mm² for Intel. Release dates differ by over a year: AMD launched on 2023-06-12, Intel on 2024-09-23. The AMD part has a launch MSRP of $14756; the Intel part has a launch MSRP of $12460.

Where Each One Wins

The AMD EPYC 9684X is the winner for rendering and CPU-bound productivity. The Cinebench R23 multi-core score of 103,355 versus 63,175 is a 63.6% lead that directly translates to faster video rendering, 3D modeling, and any software that scales with multi-core performance. The single-core lead of 63.6% also means snappier response in lightly-threaded tasks. For integer math, the 32.4% advantage makes it the pick for database operations, financial modeling, and general server logic. The physics test win by 636.9% is extreme, indicating a massive edge in simulation workloads.

The AMD chip also wins decisively in data compression (14.1% lead), encryption (42.5% lead), and random string sorting (45% lead). These are common in file servers, database workloads, and security applications. The multi-thread PassMark score of 121,595 versus 74,324 confirms AMD’s overall throughput advantage. For any workload that is not specifically floating-point or extended-instruction heavy, the EPYC 9684X is the data-backed choice.

The Intel Xeon 6980P wins in two specific areas. The 17.2% lead in PassMark extended instructions suggests it is better suited for workloads using AVX-512 or similar vector extensions, which are common in scientific computing and AI inference. The 5.9% lead in floating-point math is smaller but still relevant for certain engineering simulations and financial risk calculations. Additionally, Intel’s higher memory bandwidth of 614.4 GB/s versus 460.8 GB/s, and its greater core count of 128 versus 96, may provide advantages in memory-bound workloads that are not captured in these benchmark scores. For users running primarily floating-point or extended-instruction code, the Xeon 6980P is the better fit.