Intel Xeon 6741P vs Intel Xeon 696X Comparison

Head-to-Head Benchmarks

The recorded data presents a clear split between these two Granite Rapids processors. The Intel Xeon 696X wins 12 of the 14 head-to-head comparisons, but the Intel Xeon 6741P takes two decisive victories that reveal its unique strengths. The overall benchmark scores reinforce this: the Xeon 696X averages 286,102 points across all tests, while the Xeon 6741P averages 194,901.

The Xeon 696X dominates the Cinebench suite with consistent margins. In Cinebench R15 multicore, it scores 8,994 versus 8,624 for the 6741P, a 4.3% advantage. The R20 multicore test shows the same 4.3% delta, with scores of 37,475 and 35,935 respectively. R23 multicore follows the pattern exactly, 89,227 against 85,561, again 4.3% ahead. These consistent margins across Cinebench versions suggest a steady multicore throughput advantage rather than workload-specific optimization.

PassMark integer and floating point math tests show larger gaps. The Xeon 696X scores 572,072 in integer math, 24.9% above the 6741P's 458,058. Floating point math shows a 25.6% lead, 450,164 versus 358,423. Data compression favors the 696X by 24.7%, with scores of 2,264,907 and 1,816,408. Data encryption shows a 25.4% edge, 112,529 against 89,746. Extended instruction performance is 21.2% higher on the 696X, 172,975 versus 142,682. These PassMark workloads measure sustained computational throughput, and the 696X's extra 16 cores translate directly into roughly a quarter more work completed.

Single-thread performance also favors the Xeon 696X. In PassMark single-thread testing, it scores 3,742 against 3,195, a 17.1% improvement. This is notable because the 6741P has a higher base clock at 2.50 GHz versus 2.40 GHz, but the 696X boosts to 4.80 GHz compared to 3.80 GHz on the 6741P. The boost clock advantage appears to carry the day in lightly threaded workloads.

Random string sorting is nearly a tie. The 696X scores 180,392, just 1.7% ahead of the 6741P's 177,322. PassMark multithread shows a 4.3% edge for the 696X, 104,974 versus 100,660, a smaller margin than the raw core count difference might suggest. The 6741P's higher base clock likely narrows the gap in this aggregate metric.

The Xeon 6741P posts two significant wins. In PassMark find prime numbers, it scores 1,242 against 853 for the 696X, a 31.3% advantage. This is a striking reversal given the 696X's dominance elsewhere. The physics test shows an even larger swing: 13,890 for the 6741P versus 3,382 for the 696X, a 75.7% margin. These two workloads appear to reward the 6741P's architectural characteristics in ways that generic multicore tests do not capture.

FAQ

Q: Which processor has the higher average benchmark score?

A: The Intel Xeon 696X averages 286,102 points across all recorded benchmarks, while the Intel Xeon 6741P averages 194,901 points.

Q: How does the Xeon 696X compare to its closest rivals?

A: The 696X sits near the top of the database. It is 0.2% ahead of the AMD EPYC 9565, 0.3% behind the AMD EPYC 9555P, 2% ahead of the Intel Xeon 6780E, and 2.3% ahead of the AMD Ryzen Threadripper 9970X in average score.

Q: Where does the Xeon 6741P rank among its peers?

A: The 6741P is 0.3% ahead of the AMD EPYC 9335, 0.7% ahead of the Intel Xeon 678X, 3.8% ahead of the Intel Xeon 6740E, and 5.3% ahead of the AMD EPYC 8534P in average score.

Q: Does the Xeon 6741P win any benchmark categories?

A: Yes, it wins PassMark find prime numbers by 31.3% (1,242 versus 853) and PassMark physics by 75.7% (13,890 versus 3,382).

Q: What is the single-thread performance difference?

A: The Xeon 696X scores 3,742 in PassMark single-thread testing, 17.1% higher than the 6741P's 3,195.

Q: Are both processors in the same performance percentile?

A: Yes, both sit at the 99th percentile among all CPUs in the database.

Architecture Differences

Both processors share the Granite Rapids architecture and are built on Intel's 5 nm process node. The die size is identical at 2x 598 mm², and both use the Intel Socket 4710. The foundry is Intel for both parts. The L1 cache is 112 KB per core on each, and L2 is 2 MB per core on each.

The core counts differ substantially. The Xeon 696X provides 64 cores and 128 threads, while the Xeon 6741P offers 48 cores and 96 threads. That is a 16-core, 32-thread difference in favor of the 696X. The L3 cache also diverges: the 696X has 336 MB shared, the 6741P has 288 MB shared. The 48 MB L3 difference tracks the core count difference, as each core contributes the same L3 allocation.

The generation labels differ in the database. The 696X is listed under "Xeon 600 (Granite Rapids-WS)" while the 6741P falls under "Xeon 6 (Granite Rapids-SP)". This suggests different product tiers within the same architectural family, with the WS designation implying workstation orientation and SP implying scalable performance orientation.

Memory support is identical on paper: both use DDR5 with an eight-channel memory bus and 409.6 GB/s bandwidth. ECC memory is supported on both. Integrated graphics are absent from both parts, consistent with their server and workstation market segment. PCIe connectivity differs slightly: the 696X provides Gen 5 with 128 lanes (CPU only), while the 6741P provides Gen 5 with 136 lanes (CPU only). The 6741P offers eight additional PCIe lanes despite having fewer cores.

The multiplier is unlocked on the 696X but locked on the 6741P. This is a meaningful distinction for overclocking potential, though both parts target server and workstation environments where unlocked multipliers are less commonly exploited. The production status for both is Active, and both were released within roughly a year of each other, with the 6741P appearing first in the database's release records.

Specification Differences

The core and thread counts are the most prominent specification gap: 64 cores and 128 threads on the 696X versus 48 cores and 96 threads on the 6741P.

Clock speeds differ in both directions. The 6741P has the higher base clock at 2.50 GHz, while the 696X sits at 2.40 GHz. The boost clock reverses the order: the 696X reaches 4.80 GHz, a full 1.00 GHz above the 6741P's 3.80 GHz.

Thermal design power is higher on the 696X at 350 watts versus 300 watts on the 6741P. The 50-watt difference reflects the additional cores and higher boost capability.

L3 cache is 336 MB shared on the 696X, 288 MB shared on the 6741P. L1 and L2 cache configurations are identical per core.

PCIe lane count favors the 6741P, with 136 lanes (CPU only) against 128 lanes (CPU only) on the 696X.

The multiplier is unlocked on the 696X and locked on the 6741P. The part numbers differ: SRWQ7 for the 696X, SRVEY for the 6741P. The release dates differ, with the 6741P appearing earlier in the database.

Where Each One Wins

The Xeon 696X is the clear choice for throughput-heavy server and workstation workloads. Its 64 cores, 128 threads, and 336 MB L3 cache deliver consistent 4.3% wins in Cinebench multicore tests. The 24.7% to 25.6% advantages in data compression, encryption, integer math, and floating point math make it the stronger option for database workloads, scientific computing, financial modeling, and any task that scales across many cores. The 17.1% single-thread lead, driven by the 4.80 GHz boost clock, also gives it the edge in mixed workloads that include lightly threaded components.

The Xeon 6741P is not without its niches. Its 31.3% win in find prime numbers and 75.7% win in physics are not anomalies to dismiss. Prime number finding is a latency-sensitive workload that benefits from higher base clocks and certain instruction patterns. The physics benchmark, which often models rigid body dynamics and particle interactions, similarly rewards the 6741P's characteristics. For workloads that resemble these tests, the 6741P may outperform the larger chip despite its lower core count.

The 6741P also offers a more favorable power profile at 300 watts versus 350 watts, and it provides eight additional PCIe lanes. For systems that need more I/O connectivity per CPU and run physics or prime-heavy applications, the 6741P presents a compelling case. The 696X remains the default recommendation for maximum aggregate throughput, as its 12 benchmark wins and 46.8% higher average score demonstrate. The choice ultimately depends on whether the workload resembles the 696X's 12 winning categories or the 6741P's two specific strengths.