Intel Xeon 6781P vs Intel Xeon 6980P Comparison

# Intel Xeon 6781P vs Intel Xeon 6980P

The Intel Xeon 6781P and Intel Xeon 6980P are both Granite Rapids server processors, yet the benchmark data reveals two sharply different performance identities. The 6781P wins 11 of 14 head-to-head tests, often by massive margins, while the 6980P counters in only three workloads. The 6781P is the clear compute leader in most measured scenarios, but the 6980P claims specific advantages in integer math, encryption, and extended instruction workloads.

Where Each One Wins

The Intel Xeon 6781P dominates multi-threaded rendering and general computation. In Cinebench R15, R20, and R23 multi-core tests, the 6781P scores 10105, 42106, and 100254 respectively, against the 6980P's 6367, 26533, and 63175. Each of these represents a 58.7% advantage for the 6781P. The PassMark multithread score repeats the same pattern: 117946 versus 74324, again a 58.7% lead.

The 6781P also excels in single-threaded performance. PassMark single-thread scores show 3152 versus 1681, a 87.5% advantage. This is a decisive gap, suggesting the 6781P's core design delivers substantially higher per-core throughput in this specific test suite.

The most extreme victories for the 6781P come in specialized workloads. PassMark physics shows 17753 versus 3350, a 429.9% advantage. PassMark find prime numbers records 1687 versus 555, a 204% lead. These are not marginal wins; they indicate fundamentally different computational characteristics.

The 6980P wins three tests. PassMark integer math favors it at 637476 versus 584834, an 8.3% lead. PassMark data encryption shows 125246 versus 119623, a 4.5% advantage. PassMark extended instructions records 214794 versus 199048, a 7.3% margin. These wins suggest the 6980P has an edge in specific instruction-heavy and integer-oriented workloads.

Architecture Differences

Both processors share the Granite Rapids architecture and are built on Intel's 5 nm process. However, the physical implementations diverge significantly. The 6781P uses two dies of 598 mm² each, while the 6980P uses three dies of the same size. This die count difference directly corresponds to core count: the 6781P has 80 cores and 160 threads, while the 6980P has 128 cores and 256 threads.

Cache configurations scale with core count. Both allocate 112 KB of L1 cache and 2 MB of L2 cache per core. However, the shared L3 cache differs: the 6781P has 336 MB shared, while the 6980P has 504 MB shared. The 6980P's larger L3 pool provides more aggregate cache capacity across its additional cores.

Memory subsystems differ as well. The 6781P uses eight-channel DDR5 memory with 409.6 GB/s bandwidth. The 6980P uses twelve-channel DDR5 with 614.4 GB/s bandwidth. The 6980P's wider memory interface provides 50% more theoretical bandwidth, which may explain its advantages in memory-intensive integer operations.

Clock speeds show subtle differences. Both have a 2.00 GHz base clock, but the boost clock differs: 3.80 GHz for the 6781P versus 3.90 GHz for the 6980P. The 6980P's higher boost clock is notable given its larger core count, though the 6781P's benchmark results suggest its lower core count allows higher sustained performance in many tests.

Socket and PCIe configurations also separate them. The 6781P uses Intel Socket 4710 with 136 PCIe Gen 5 lanes. The 6980P uses Intel Socket 7529 with 96 PCIe Gen 5 lanes. The 6781P offers more PCIe connectivity despite having fewer cores. The 6980P has a higher TDP of 500 watts versus 350 watts for the 6781P.

The Verdict

The data points decisively toward the Intel Xeon 6781P for most workloads. Its 58.7% lead across all three Cinebench multi-core tests and the PassMark multithread test establishes a clear pattern. The 87.5% single-thread advantage reinforces that the 6781P delivers stronger per-core performance, not just aggregate throughput. For rendering, physics simulation, prime number computation, and random string sorting, the 6781P is the superior choice.

The 6980P's wins are narrower but meaningful. Its 8.3% lead in integer math, 4.5% in encryption, and 7.3% in extended instructions indicate that workloads heavily dependent on integer operations or specific instruction sets may benefit from its 128-core configuration and larger L3 cache. The 6980P's twelve-channel memory interface with 614.4 GB/s bandwidth likely supports these workloads.

The 6781P also holds a significant advantage in average benchmark score: 315524 versus 251516, a 25.4% gap. Both processors sit in the 99th percentile of all CPUs, but the 6781P's average score places it closer to rivals like the AMD EPYC 9575F (311774, 1.2% behind) and AMD EPYC 9734 (310619, 1.6% behind). The 6980P's average score of 251516 places it 3% ahead of the AMD EPYC 9634 (244274) and 5.6% ahead of the Intel Xeon 6747P (238263), but 5.8% behind the AMD EPYC 9684X (266914).

For users prioritizing multi-core rendering, physics, and single-thread performance, the 6781P is the clear pick. For integer-heavy enterprise workloads, encryption, and extended instruction processing, the 6980P offers a compelling alternative despite its lower overall benchmark averages.

FAQ

Q: Which processor wins more benchmark comparisons?

A: The Intel Xeon 6781P wins 11 of 14 head-to-head tests, while the Intel Xeon 6980P wins 3. The 6781P's victories include all Cinebench multi-core tests, PassMark multithread, single-thread, physics, and find prime numbers.

Q: How large is the performance gap in multi-core rendering?

A: In Cinebench R15, R20, and R23 multi-core tests, the 6781P leads by 58.7% in each case. The 6781P scores 10105, 42106, and 100254 respectively, versus 6367, 26533, and 63175 for the 6980P.

Q: Where does the 6980P outperform the 6781P?

A: The 6980P wins in PassMark integer math (637476 versus 584834, an 8.3% lead), data encryption (125246 versus 119623, a 4.5% lead), and extended instructions (214794 versus 199048, a 7.3% lead).

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

A: The 6781P has 80 cores and 160 threads. The 6980P has 128 cores and 256 threads. Despite having 60% more cores, the 6980P loses most benchmark comparisons.

Q: How do the memory systems differ?

A: The 6781P uses eight-channel DDR5 with 409.6 GB/s bandwidth. The 6980P uses twelve-channel DDR5 with 614.4 GB/s bandwidth, offering 50% more theoretical memory bandwidth.

Q: What are the average benchmark scores?

A: The 6781P has an average benchmark score of 315524, while the 6980P averages 251516. The 6781P sits 1.2% ahead of the AMD EPYC 9575F (311774) and 1.6% ahead of the AMD EPYC 9734 (310619). The 6980P is 3% ahead of the AMD EPYC 9634 (244274) and 5.6% ahead of the Intel Xeon 6747P (238263).

Head-to-Head Benchmarks

The largest single victory belongs to the 6781P in PassMark physics. Its score of 17753 versus 3350 represents a 429.9% advantage. This extreme gap suggests the 6781P's core architecture is far more efficient at physics calculations, likely due to its higher per-core performance and lower core count avoiding scheduling overhead.

PassMark find prime numbers shows another dramatic gap. The 6781P scores 1687 versus 555 for the 6980P, a 204% lead. This workload is highly dependent on integer arithmetic and cache behavior, yet the 6781P's smaller core count still dominates.

The single-thread gap is equally striking. PassMark single-thread scores show 3152 versus 1681, an 87.5% advantage for the 6781P. This indicates the 6781P's cores are substantially faster individually, which explains its dominance in lightly threaded workloads.

The Cinebench multi-core results consistently show 58.7% advantages for the 6781P across all three versions (R15, R20, R23). This uniformity suggests a consistent architectural advantage in rendering workloads, not a benchmark-specific artifact. The PassMark multithread score confirms this pattern with the same 58.7% delta.

Random string sorting favors the 6781P by 11.5% (268573 versus 240792). Data compression shows a narrower 3.3% lead (2441690 versus 2364519). Floating point math is nearly tied at 507406 versus 501720, a 1.1% margin.

The 6980P's best win is in integer math at 637476 versus 584834, an 8.3% margin. This suggests its 128 cores and larger L3 cache help in workloads with high parallelism in integer operations. Data encryption shows a 4.5% advantage (125246 versus 119623), and extended instructions a 7.3% edge (214794 versus 199048).

Specification Differences

| Specification | Intel Xeon 6781P | Intel Xeon 6980P |

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

| Cores | 80 | 128 |

| Threads | 160 | 256 |

| Boost Clock | 3.80 GHz | 3.90 GHz |

| TDP | 350 W | 500 W |

| Socket | Intel Socket 4710 | Intel Socket 7529 |

| Die Size | 2x 598 mm² | 3x 598 mm² |

| L3 Cache | 336 MB (shared) | 504 MB (shared) |

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

| Memory Bandwidth | 409.6 GB/s | 614.4 GB/s |

| PCIe | Gen 5, 136 Lanes (CPU only) | Gen 5, 96 Lanes (CPU only) |

| Launch MSRP | $8960 | $12460 |

| Release Date | 2025-02-23 | 2024-09-23 |

Both processors share the same base clock of 2.00 GHz, L1 cache of 112 KB per core, L2 cache of 2 MB per core, DDR5 memory support, ECC memory support, 5 nm process node, and no integrated graphics. They also share the same Granite Rapids codename and architecture, with the 6781P belonging to the Granite Rapids-SP generation and the 6980P to Granite Rapids-AP.