Intel Xeon 6740E vs Intel Xeon 6747P Comparison

Where Each One Wins

The benchmark results split these two Xeon 6 processors into clearly different roles. The Intel Xeon 6747P wins 12 of the 14 recorded head-to-head tests, establishing itself as the stronger all-around performer for compute-heavy workloads. Its victories span Cinebench multi-core tests, PassMark integer and floating point math, prime number finding, physics simulation, extended instruction sets, and single-thread performance. The Xeon 6747P also edges ahead in data compression by 2.6%, a narrow but consistent margin.

The Intel Xeon 6740E wins only two tests, but those wins are telling. It leads in data encryption by 33.8% and random string sorting by 18.3%. Both workloads benefit from high memory bandwidth and parallel throughput across many cores. The 6740E packs 96 cores versus the 6747P's 48 cores, which explains its advantage in tasks that scale with core count rather than per-core speed. However, the 6740E trails in every Cinebench multi-core test by 33.5%, which is surprising given its core advantage, indicating that the 6747P's higher clock speeds and architecture choices outweigh raw core counts in rendering workloads.

For single-threaded tasks, the 6747P is decisively ahead. Its PassMark single-thread score of 3236 versus 1997 for the 6740E represents a 62% advantage. This carries over to physics simulation, where the 6747P scores 76.1% higher, and extended instructions, where it leads by 80.5%. The 6740E is best suited for workloads that are memory-bound or encryption-heavy, while the 6747P dominates CPU-bound computation.

Architecture Differences

The two processors come from different Intel architectures. The Xeon 6747P uses Granite Rapids, part of the Xeon 6 Granite Rapids-SP generation, while the Xeon 6740E uses Sierra Forest, from the Sierra Forest-SP generation. Both are built on Intel's 5 nm process node and use the same Intel Socket 4710, but the silicon designs diverge significantly.

Core counts differ dramatically. The 6747P has 48 cores and 96 threads, while the 6740E has 96 cores and 96 threads. The 6740E has no hyperthreading, meaning each core handles one thread, whereas the 6747P doubles its thread count. Clock speeds also favor the 6747P: 2.70 GHz base and 3.90 GHz boost versus 2.40 GHz base and 3.20 GHz boost for the 6740E. Thermal design power reflects this, with the 6747P rated at 330 watts and the 6740E at 250 watts, though the 6740E delivers more cores within a lower power envelope.

Cache hierarchies are structured differently. The 6747P uses 112 KB L1 per core, 2 MB L2 per core, and a shared 288 MB L3 cache. The 6740E uses 96 KB L1 per core, 4 MB L2 per module, and a shared 96 MB L3 cache. The die sizes also differ: the 6747P uses a dual-die design totaling 2x 598 mm², while the 6740E uses a single 578 mm² die. The 6747P's larger L3 cache is a key advantage for workloads with large working sets.

Both processors support DDR5 memory with an eight-channel bus and 409.6 GB/s bandwidth, plus Gen 5 PCIe with 88 CPU lanes. Neither has integrated graphics. The 6747P was released later, in February 2025, while the 6740E arrived in June 2024. The 6747P has a launch MSRP of $6497, and the 6740E has a launch MSRP of $5265.

Head-to-Head Benchmarks

The most consistent gap appears in Cinebench multi-core tests. Across Cinebench R15, R20, and R23, the 6747P scores 8712, 36301, and 86432 respectively, while the 6740E scores 6525, 27191, and 64741. The delta is identical at 33.5% in each case. This consistency suggests a fundamental per-core performance advantage for the 6747P that scales linearly across rendering workloads.

PassMark integer math is close, with the 6747P scoring 468518 against 457614, a 2.4% margin. Floating point math shows a larger gap: 365904 versus 309333, an 18.3% lead for the 6747P. Prime number finding is where the biggest delta appears: the 6747P scores 1151 versus 526, an 118.8% advantage. Physics simulation follows at 76.1% (13398 versus 7608). Extended instructions show an 80.5% gap (142557 versus 78968).

The 6740E's wins are substantial. Data encryption scores 137106 versus 90789, a 33.8% margin. Random string sorting scores 220684 versus 180382, an 18.3% margin. Data compression is nearly even, with the 6747P at 1833378 and the 6740E at 1786845, a 2.6% edge for the 6747P. PassMark multithread follows the Cinebench pattern with a 33.5% lead for the 6747P (101685 versus 76167).

Single-thread tests appear twice in the data, once as "single_thread" and once as "singlethread," with identical scores. The 6747P scores 3236 in both, the 6740E scores 1997 in both, and the 62% delta repeats. This confirms the 6747P's per-core superiority is consistent across repeated measurements.

FAQ

Q: Which processor has more cores?

A: The Intel Xeon 6740E has 96 cores, while the Intel Xeon 6747P has 48 cores. However, the 6747P has 96 threads due to hyperthreading, matching the 6740E's thread count.

Q: Why does the 6747P win multi-core benchmarks despite having fewer cores?

A: The 6747P has higher clock speeds (2.70 GHz base, 3.90 GHz boost versus 2.40 GHz and 3.20 GHz) and a much larger L3 cache (288 MB versus 96 MB). These factors lead to a 33.5% advantage in Cinebench R15, R20, and R23 multi-core tests.

Q: In which workloads does the 6740E outperform the 6747P?

A: The 6740E wins in data encryption by 33.8% (137106 versus 90789) and random string sorting by 18.3% (220684 versus 180382). These tasks benefit from its higher core count and module-based L2 cache design.

Q: What is the single-thread performance difference?

A: The 6747P scores 3236 in PassMark single-thread tests, which is 62% higher than the 6740E's score of 1997. This makes the 6747P clearly better for latency-sensitive, single-threaded applications.

Q: How do their memory systems compare?

A: Both processors support DDR5 with an eight-channel memory bus and 409.6 GB/s bandwidth. They also both have ECC memory support and Gen 5 PCIe with 88 lanes.

Q: Which processor has a larger L3 cache?

A: The 6747P has a shared 288 MB L3 cache, three times larger than the 6740E's shared 96 MB L3 cache. This benefits workloads with large data sets that need frequent cache hits.

The Verdict

The data points to a clear split. For general-purpose server and workstation workloads that mix compute, rendering, physics, and single-threaded tasks, the Intel Xeon 6747P is the stronger choice. It wins 12 of 14 benchmarks and leads by significant margins in most categories. Its 33.5% advantage across all Cinebench multi-core tests and 62% single-thread lead make it the default pick for applications where per-core performance matters.

The Intel Xeon 6740E is the better option for specific workloads. Its 96 cores and 33.8% encryption advantage make it suitable for security-focused tasks, while its 18.3% lead in random string sorting suggests strength in text processing or data manipulation pipelines. Its lower 250 watt TDP and larger core count also make it a candidate for power-conscious deployments that prioritize throughput over per-core speed.

Against its nearest rivals, the 6747P sits at the 99th percentile of all CPUs, with an average benchmark score of 238263. It trails the AMD EPYC 9634 by 2.5% and the Intel Xeon 6980P by 5.3%, while leading the AMD EPYC 9455P by 9.4% and the Intel Xeon w9-3595X by 13.5%. The 6740E also sits at the 99th percentile with an average score of 187718, leading the AMD EPYC 8534P by 1.4% and the AMD Ryzen Threadripper PRO 9975WX by 2.7%, while trailing the Intel Xeon 678X by 3% and the AMD EPYC 9335 by 3.4%.

For buyers choosing between these two, the decision rests on workload character. If encryption and string-heavy tasks dominate, the 6740E's 96 cores deliver measurable wins. For everything else, the 6747P's higher clocks, larger cache, and superior single-thread performance make it the more versatile processor.

Specification Differences

| Specification | Intel Xeon 6747P | Intel Xeon 6740E |

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

| Cores | 48 | 96 |

| Threads | 96 | 96 |

| Base Clock | 2.70 GHz | 2.40 GHz |

| Boost Clock | 3.90 GHz | 3.20 GHz |

| TDP | 330 W | 250 W |

| Architecture | Granite Rapids | Sierra Forest |

| Codename | Granite Rapids | Sierra Forest |

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

| Die Size | 2x 598 mm² | 578 mm² |

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

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

| L3 Cache | 288 MB (shared) | 96 MB (shared) |

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

| Launch MSRP | $6497 | $5265 |

| Part Number | SRVEZ | SRPFV |