Intel Xeon 6740P vs Intel Xeon 6747P Comparison

Head-to-Head Benchmarks

The Intel Xeon 6747P wins every recorded benchmark in the head-to-head comparison against the Intel Xeon 6740P, a clean sweep of 14 wins to 0. The margin is not uniform, however, and the distribution of those deltas reveals where the higher-clocked part pulls away most dramatically.

The largest single gap appears in PassMark's find prime numbers test, where the 6747P scores 1151 against 822 for the 6740P, a 40% advantage. This is a classic indicator of raw integer throughput and clock-driven latency sensitivity, and it is the standout outlier in the entire comparison. Close behind is PassMark physics, where the 6747P's 13398 result beats the 6740P's 9705 by 38.1%. Physics workloads often scale with sustained all-core frequency, and the data here strongly supports that interpretation.

In the floating-point and extended instruction workloads, the 6747P leads by roughly a fifth. PassMark floating point math shows a 23.6% delta (365904 versus 296102), while extended instructions come in at 21.9% (142557 versus 116992). Integer math follows at 20.6% (468518 versus 388500). These are substantial, consistent margins across the math-heavy suite, indicating that the 6747P is not just faster in one narrow test but carries a broad throughput advantage in compute-intensive operations.

Data compression favors the 6747P by 19.3% (1833378 versus 1537129), a notable result for storage and database workloads. Encryption shows a smaller but still healthy 10.7% lead (90789 versus 82028). The multithread aggregate score from PassMark gives the 6747P a 15.5% edge (101685 versus 88061), which mirrors the Cinebench results exactly.

All three Cinebench multicore runs, R15, R20, and R23, produce the same 15.5% delta in favor of the 6747P: 8712 versus 7544, 36301 versus 31437, and 86432 versus 74851 respectively. This consistency across Cinebench generations suggests a stable, frequency-driven performance gap rather than workload-specific noise. The single-thread results are closer, with the 6747P ahead by 8.8% in both PassMark single-thread and singlethread tests (3236 versus 2975). Random string sorting is the tightest contest at just 3.1% (180382 versus 175015), indicating that memory-access patterns and latency dominate that workload more than raw core speed.

Architecture Differences

Both processors share the same fundamental silicon. They are built on Granite Rapids architecture, part of the Xeon 6 (Granite Rapids-SP) generation, fabricated on Intel's 5 nm process node. Both use the same dual-die design with a combined die size of 2x 598 mm². Core counts are identical at 48 cores and 96 threads, and the cache hierarchy is exactly the same: 112 KB of L1 per core, 2 MB of L2 per core, and 288 MB of shared L3. Memory support is also identical, with DDR5 across an eight-channel bus delivering 409.6 GB/s of bandwidth, and ECC memory is supported on both.

The distinguishing factor is clock speed and power envelope. The 6747P has a base clock of 2.70 GHz and a boost clock of 3.90 GHz, while the 6740P runs at 2.10 GHz base and 3.80 GHz boost. That 0.60 GHz base-clock gap is the primary driver behind the benchmark deltas observed above. The boost clocks are closer, differing by only 0.10 GHz, which explains why single-thread performance is much closer than all-core performance. Sustained all-core workloads on the 6747P benefit from the higher base frequency.

The thermal design power scales accordingly. The 6747P is rated at 330 W TDP, while the 6740P is rated at 270 W. This 60 W difference is the cost of the higher base clock and the corresponding performance gains. Both parts use the same Intel Socket 4710, and both support PCIe Gen 5 with 88 lanes (CPU only). Neither has integrated graphics, and neither has an unlocked multiplier. Both were released on the same date, 2025-02-23, and both are currently active production parts. The part numbers are SRVEZ for the 6747P and SRV5R for the 6740P.

The Verdict

The benchmark data is unambiguous: the Intel Xeon 6747P is the faster processor in every measured workload. If raw performance is the sole criterion, the 6747P is the correct choice. The question is whether the performance delta justifies the difference in launch MSRP, which is $6497 for the 6747P and $4650 for the 6740P.

The average benchmark score tells the broader story. The 6747P has an average score of 238263, placing it in the 99th percentile of all CPUs. Its nearest rivals include the AMD EPYC 9634 (average score 244274, 2.5% higher), the Intel Xeon 6980P (251516, 5.3% higher), the AMD EPYC 9455P (217854, 9.4% lower), and the Intel Xeon w9-3595X (209881, 13.5% lower). The 6740P, by contrast, has an average score of 176227, placing it in the 98th percentile. Its nearest rivals are the AMD EPYC 7763 (179916, 2.1% higher), the AMD Ryzen Threadripper PRO 9975WX (182700, 3.5% higher), the AMD Ryzen Threadripper PRO 3995WX (171748, 2.6% lower), and the AMD EPYC 7C13 (167788, 5% lower).

The 6747P sits in a higher performance tier altogether, competing with top-end EPYC and Xeon parts, while the 6740P trades blows with previous-generation and mid-range Threadripper and EPYC offerings. For workloads where the 15.5% to 40% gains in the head-to-head tests translate directly to throughput, the 6747P's higher launch MSRP may be justified. For deployments where the 6740P's lower TDP and lower launch MSRP matter more, and where the performance gap can be absorbed by scaling out additional nodes, the 6740P remains a strong option.

FAQ

Q: Which processor has the higher boost clock?

A: The Intel Xeon 6747P has a boost clock of 3.90 GHz, while the Intel Xeon 6740P has a boost clock of 3.80 GHz.

Q: How large is the L3 cache on both processors?

A: Both the Intel Xeon 6747P and the Intel Xeon 6740P have 288 MB of shared L3 cache.

Q: What is the TDP difference between the two?

A: The Intel Xeon 6747P is rated at 330 W TDP, and the Intel Xeon 6740P is rated at 270 W TDP.

Q: Do both processors support the same memory configuration?

A: Yes, both support DDR5 memory across an eight-channel bus with 409.6 GB/s bandwidth, and both support ECC memory.

Q: In which test does the 6747P show the largest advantage?

A: The largest advantage is in PassMark find prime numbers, where the 6747P scores 1151 versus 822 for the 6740P, a 40% delta.

Q: Are there any benchmarks where the 6740P wins?

A: No. The recorded data shows the 6747P wins all 14 head-to-head benchmarks, with the 6740P winning none.

Where Each One Wins

The Intel Xeon 6747P wins across every benchmark category in this comparison, but the magnitude of its wins varies by workload type. The biggest gaps are in prime number finding (40%), physics (38.1%), floating-point math (23.6%), and extended instructions (21.9%). These are compute-heavy, frequency-sensitive workloads where the 6747P's higher base clock of 2.70 GHz versus 2.10 GHz provides a decisive edge. Integer math (20.6%) and data compression (19.3%) also show strong gains, making the 6747P the better fit for database compression, scientific computing, and financial modeling that relies on integer arithmetic.

The 6747P also leads in the Cinebench suite by a consistent 15.5% across R15, R20, and R23, which suggests that 3D rendering and content-creation workloads will see a predictable improvement. The multithread PassMark score follows the same 15.5% pattern, reinforcing the all-core advantage.

The Intel Xeon 6740P does not win any recorded benchmark, but its profile is not without merit. The smallest delta is in random string sorting, where the 6747P leads by only 3.1%. This indicates that memory-bound or latency-bound workloads will see the least benefit from upgrading to the 6747P. The 6740P's lower TDP of 270 W versus 330 W makes it more suitable for power-constrained deployments, and its lower launch MSRP of $4650 versus $6497 positions it as the more economical option in dense server configurations where per-socket cost and power density are primary constraints.

For single-threaded tasks, the gap narrows to 8.8% in PassMark single-thread tests. Applications that are lightly threaded and latency-sensitive will still favor the 6747P, but the difference is modest compared to the all-core workloads.

Specification Differences

The following fields differ between the Intel Xeon 6747P and the Intel Xeon 6740P:

• Base Clock: 2.70 GHz on the 6747P, 2.10 GHz on the 6740P
• Boost Clock: 3.90 GHz on the 6747P, 3.80 GHz on the 6740P
• TDP: 330 W on the 6747P, 270 W on the 6740P
• Launch MSRP: $6497 for the 6747P, $4650 for the 6740P
• Part Number: SRVEZ for the 6747P, SRV5R for the 6740P

All other recorded specifications are identical, including core count (48), threads (96), architecture (Granite Rapids), process node (5 nm), die size (2x 598 mm²), cache hierarchy (112 KB L1 per core, 2 MB L2 per core, 288 MB L3 shared), memory support (DDR5, eight-channel, 409.6 GB/s, ECC), PCIe (Gen 5, 88 lanes CPU only), socket (Intel Socket 4710), integrated graphics (N/A), market segment (Server/Workstation), production status (Active), release date (2025-02-23), and unlocked multiplier (false on both).