Intel Xeon 6747P vs Intel Xeon 6780E Comparison

FAQ

Q: Which processor has more physical cores, the Intel Xeon 6780E or the Intel Xeon 6747P?

A: The Intel Xeon 6780E has 144 cores and 144 threads, while the Intel Xeon 6747P has 48 cores and 96 threads. The 6780E offers three times the core count but does not support hyperthreading, whereas the 6747P uses two threads per core.

Q: How do the two chips compare in single-threaded performance?

A: The Intel Xeon 6747P is far ahead in single-threaded workloads. In PassMark single-thread testing, the 6747P scores 3236 versus 1923 for the 6780E, a 40.6% advantage. The 6747P also has a higher boost clock of 3.90 GHz compared to 3.00 GHz on the 6780E.

Q: Which processor wins more head-to-head benchmarks?

A: The Intel Xeon 6747P wins 9 of the 14 recorded head-to-head comparisons, while the Intel Xeon 6780E wins 5. The 6747P takes all three Cinebench multi-core tests plus several PassMark workloads, while the 6780E dominates data compression, encryption, integer math, and random string sorting.

Q: What is the difference in cache layout between the two Xeons?

A: The 6780E has 96 KB of L1 per core, 4 MB of L2 per module, and 108 MB of shared L3. The 6747P has 112 KB of L1 per core, 2 MB of L2 per core, and a much larger 288 MB of shared L3. The 6747P's L3 is nearly 2.7 times larger.

Q: Are both processors built on the same manufacturing process?

A: Yes, both use Intel's 5 nm process node. However, they are based on different architectures: the 6780E uses Sierra Forest (Sierra Forest-SP) while the 6747P uses Granite Rapids (Granite Rapids-SP). The die sizes differ, with the 6780E at 578 mm² and the 6747P at 2x 598 mm².

Q: Do both chips support the same memory configuration?

A: Both support DDR5 memory with an eight-channel memory bus and 409.6 GB/s of memory bandwidth. Both also support ECC memory and use the same Intel Socket 4710. PCIe support is identical as well: Gen 5 with 88 lanes from the CPU.

Architecture Differences

The two processors represent fundamentally different design philosophies within Intel's Xeon 6 lineup. The Intel Xeon 6780E is built on the Sierra Forest architecture, optimized for high core density and throughput in scale-out workloads. Its 144 cores are paired with 144 threads, meaning each core handles a single thread. The L2 cache is organized per module at 4 MB, and the shared L3 pool sits at 108 MB. This design favors massive parallelism over per-core sophistication.

The Intel Xeon 6747P uses the Granite Rapids architecture, which is a performance-optimized design. It has 48 cores but 96 threads, enabling simultaneous multithreading. Each core gets 2 MB of dedicated L2, and the shared L3 expands to 288 MB. The L1 cache is slightly larger per core at 112 KB versus 96 KB. This cache hierarchy suggests the 6747P is built for workloads that benefit from lower latency and larger on-chip data residency.

Clock speeds tell a similar story. The 6747P runs at 2.70 GHz base and 3.90 GHz boost, while the 6780E sits at 2.20 GHz base and 3.00 GHz boost. Both are locked multipliers, and both draw 330 W TDP. The 6780E uses a single 578 mm² die, while the 6747P uses two dies measuring 598 mm² each. Despite the architectural differences, both share the same socket, memory system, PCIe lanes, and ECC support.

The release dates differ significantly. The 6780E launched on 2024-06-02, while the 6747P followed on 2025-02-23. Both are active production parts in the Server/Workstation segment, and neither includes integrated graphics.

The Verdict

The data points to two distinct roles. The Intel Xeon 6780E is the choice for workloads that scale with raw core count and benefit from massive parallel throughput. Its 144 cores deliver overwhelming wins in data compression (39.5% ahead), encryption (112.6% ahead), integer math (37% ahead), and random string sorting (81.3% ahead). These are classic scale-out patterns found in database compression, storage pipelines, and bulk data processing.

The Intel Xeon 6747P is the pick for workloads that favor per-core performance, larger caches, and higher clock speeds. It wins all three Cinebench multi-core tests by 14.7%, plus PassMark multithread, physics, extended instructions, prime number finding, and single-thread tests. Its 288 MB L3 and 3.90 GHz boost make it stronger for scientific computing, simulation, and latency-sensitive transactional workloads.

The benchmark split is not subtle: the 6747P wins 9 of 14 comparisons, but the 6780E's 5 wins are often by massive margins. The 6780E's average benchmark score is 280438 versus 238263 for the 6747P, placing it in the 99th percentile of all CPUs, same as the 6747P. The nearest rivals for the 6780E are AMD Ryzen Threadripper 9970X (0.2% higher), AMD EPYC 9565 (1.8% higher), Intel Xeon 696X (2% higher), and AMD EPYC 9555P (2.3% higher). For the 6747P, rivals include AMD EPYC 9634 (2.5% higher), Intel Xeon 6980P (5.3% higher), AMD EPYC 9455P (9.4% lower), and Intel Xeon w9-3595X (13.5% lower).

Choose the 6780E for encryption, compression, and integer-heavy pipelines. Choose the 6747P for single-thread responsiveness, floating-point simulation, and workloads that can exploit 288 MB of shared cache.

Specification Differences

| Specification | Intel Xeon 6780E | Intel Xeon 6747P |

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

| Cores | 144 | 48 |

| Threads | 144 | 96 |

| Base Clock | 2.20 GHz | 2.70 GHz |

| Boost Clock | 3.00 GHz | 3.90 GHz |

| Architecture | Sierra Forest | Granite Rapids |

| Codename | Sierra Forest | Granite Rapids |

| Generation | Xeon 6 (Sierra Forest-SP) | Xeon 6 (Granite Rapids-SP) |

| Die Size | 578 mm² | 2x 598 mm² |

| L1 Cache | 96 KB (per core) | 112 KB (per core) |

| L2 Cache | 4 MB (per module) | 2 MB (per core) |

| L3 Cache | 108 MB (shared) | 288 MB (shared) |

| Release Date | 2024-06-02 | 2025-02-23 |

| Launch MSRP | $11350 | $6497 |

| Part Number | SRPG3 | SRVEZ |

Both chips share identical TDP (330 W), socket (Intel Socket 4710), process node (5 nm), memory type (DDR5), memory bus (eight-channel), memory bandwidth (409.6 GB/s), ECC support (true), PCIe (Gen 5, 88 lanes CPU only), integrated graphics (N/A), market segment (Server/Workstation), production status (Active), and multiplier lock (false).

Head-to-Head Benchmarks

The largest win for the Intel Xeon 6780E comes in PassMark data encryption, where it scores 193004 against 90789, a 112.6% advantage. This is the single biggest delta in the entire comparison, more than doubling the 6747P's output. Random string sorting is next, with the 6780E at 326954 versus 180382, an 81.3% lead. Data compression shows 2557582 against 1833378, a 39.5% margin. Integer math delivers 641817 versus 468518, a 37% advantage. Floating-point math is closer but still favors the 6780E: 433862 versus 365904, an 18.6% win.

The Intel Xeon 6747P counters with a 40.6% lead in PassMark single-thread (3236 versus 1923). It also wins prime number finding by 38.5% (1151 versus 708). Extended instructions go to the 6747P by 20% (142557 versus 114008). Physics shows an 18.3% edge for the 6747P (13398 versus 10951). All three Cinebench multi-core tests show the 6747P ahead by exactly 14.7%: R15 at 8712 versus 7431, R20 at 36301 versus 30963, and R23 at 86432 versus 73723. PassMark multithread also lands at 14.7% in favor of the 6747P (101685 versus 86734).

The 6780E's average benchmark score of 280438 sits 17.7% above the 6747P's 238263. The 6780E has a 99th percentile ranking, matching the 6747P, but the absolute average scores reveal the core-heavy design's overall throughput advantage in the database's aggregate metrics.

Where Each One Wins

The Intel Xeon 6780E is the clear winner in security and data movement tasks. Encryption workloads show a 112.6% advantage, making it ideal for VPN gateways, secure storage, and cryptographic processing. Data compression at 39.5% ahead points to database backup, log archival, and file system compression. Integer math at 37% ahead suits financial risk modeling, integer-heavy analytics, and certain AI preprocessing. Random string sorting at 81.3% ahead indicates strength in text indexing, search engines, and data deduplication. Floating-point math also favors the 6780E, though by a smaller 18.6% margin.

The Intel Xeon 6747P wins in interactive and latency-sensitive workloads. Single-thread performance at 40.6% ahead makes it better for control plane tasks, API gateways, and any serialized code paths. Prime number finding at 38.5% ahead suggests advantages in cryptography key generation and number-theoretic computations. Extended instructions at 20% ahead point to SIMD-heavy applications like video encoding, signal processing, and scientific kernels. Physics at 18.3% ahead aligns with simulation and finite element analysis. The Cinebench multi-core wins at 14.7% across all three versions indicate strong all-around rendering and 3D workload performance.

The core count difference matters most for the 6780E's wins, while the 6747P's higher clocks and 288 MB L3 drive its victories. The 6747P also wins PassMark multithread by 14.7%, despite having only one-third the cores, which shows the efficiency of its per-core resources. For workloads that can fully saturate 144 threads, the 6780E is unmatched in the specific tasks it dominates. For workloads that need low latency and high per-thread throughput, the 6747P is the database's clear recommendation. The wins are not overlapping: the 6780E owns bulk data processing, while the 6747P owns compute-intensive and single-threaded duties.