Intel Xeon 638 vs Intel Xeon 6731P Comparison

Intel
INTEL

Intel Xeon 638

CORE STATE Granite Rapids
CORE SPECS 16 Cores / 32 Threads
CLOCK SPEED 3.2 Base / 4.8 GHz Turbo
CACHE 72 MB (shared)
MAX TDP 180W
ARCHITECTURE Granite Rapids
nm
PROCESS 5 nm
LAUNCH DATE 2026
VS
Intel
INTEL

Xeon 6731P

CORE STATE Granite Rapids
CORE SPECS 32 Cores / 64 Threads
CLOCK SPEED 2.5 Base / 4.1 GHz Turbo
CACHE 144 MB (shared)
MAX TDP 245W
ARCHITECTURE Granite Rapids
nm
PROCESS 5 nm
LAUNCH DATE 2025

PERFORMANCE BENCHMARKS

cinebench_cinebench_r15_multicore
4,757
4,522
cinebench_cinebench_r15_singlecore
671
638
cinebench_cinebench_r20_multicore
19,824
18,845
cinebench_cinebench_r20_singlecore
2,798
2,660
cinebench_cinebench_r23_multicore
47,202
44,871
cinebench_cinebench_r23_singlecore
6,663
6,334
passmark_data_compression
725,818
799,474
passmark_data_encryption
36,030
40,087
passmark_extended_instructions
56,498
65,656
passmark_find_prime_numbers
381
541
passmark_floating_point_math
144,757
157,330
passmark_integer_math
184,884
198,761
passmark_multithread
55,651
52,790
passmark_physics
4,704
7,105
passmark_random_string_sorting
74,318
88,019
passmark_single_thread
3,670
2,107
passmark_singlethread
3,670
2,107

Analysis: Intel Xeon 638 vs Intel Xeon 6731P

Head-to-Head Benchmarks

The benchmark data reveals a fascinating split between these two Granite Rapids Xeons. The Intel Xeon 6731P and the Intel Xeon 638 trade blows across the test suite, with the 638 edging out a narrow victory in total wins (9 wins versus 8 wins for the 6731P). However, the magnitude of the victories tells a more nuanced story than the raw win count suggests.

Looking at the Cinebench results, the Intel Xeon 638 holds a consistent advantage across all six tests. In Cinebench R15, R20, and R23, both for single-core and multi-core workloads, the 638 leads by exactly 4.9% in every instance. The multi-core scores show the 638 achieving 4757 in R15, 19824 in R20, and 47202 in R23, while the 6731P trails at 4522, 18845, and 44871 respectively. The single-core results follow the same pattern, with the 638 posting 671 in R15, 2798 in R20, and 6663 in R23, versus the 6731P's 638, 2660, and 6334.

The Passmark single-thread test delivers the most dramatic difference between the two processors. Here, the Intel Xeon 638 achieves a score of 3670, which is a massive 42.6% ahead of the 6731P's 2107. This is by far the largest percentage gap in either direction across the entire benchmark suite. The Passmark multithread test also goes to the 638, though by a much narrower margin of 5.1%, with scores of 55651 versus 52790.

The Intel Xeon 6731P, however, dominates in several Passmark sub-tests, often with substantial margins. The most striking victory comes in the Passmark physics test, where the 6731P scores 7105, a full 51% ahead of the 638's 4704. The find prime numbers test also shows a major gap, with the 6731P posting 541 against 381, a 42% advantage. In random string sorting, the 6731P leads by 18.4% (88019 versus 74318). The extended instructions test shows a 16.2% lead for the 6731P (65656 versus 56498), while data encryption favors the 6731P by 11.3% (40087 versus 36030). Data compression sees the 6731P ahead by 10.1% (799474 versus 725818), and floating point math shows an 8.7% edge (157330 versus 144757). Integer math rounds out the 6731P's wins with a 7.5% advantage (198761 versus 184884).

Where Each One Wins

The data paints a clear picture of two processors optimized for different workloads. The Intel Xeon 638 excels in scenarios where raw single-thread performance and lightly threaded multi-core tasks matter most. Its consistent 4.9% lead across all Cinebench versions, both single and multi-core, suggests that its higher clock speeds translate directly into better performance for rendering tasks that scale with per-core speed. The enormous 42.6% advantage in Passmark single-thread performance reinforces this interpretation, making the 638 the clear choice for applications that rely heavily on single-threaded execution or that cannot efficiently distribute work across many cores.

The Intel Xeon 6731P, on the other hand, demonstrates its strength in specialized compute tasks. The 51% lead in physics calculations and the 42% advantage in prime number finding indicate that this processor handles algorithmic and scientific workloads with notably greater efficiency. The 16.2% lead in extended instructions suggests that workloads leveraging advanced CPU instruction sets will see meaningful benefits from the 6731P. Data compression and encryption tasks, critical for database and storage workloads, also favor the 6731P by margins between 10% and 12%. The 6731P's 8.7% and 7.5% leads in floating point and integer math respectively round out a profile that suits computational research, financial modeling, and data processing applications.

Architecture Differences

Both processors share the same Granite Rapids architecture and are built on Intel's 5 nm process node. They occupy the same socket, Intel Socket 4710, and share an identical die size of 598 mm². The fundamental difference lies in how Intel has configured these chips within the same architectural family.

The Intel Xeon 6731P doubles the core count of the 638, offering 32 cores and 64 threads versus 16 cores and 32 threads. This core advantage is complemented by a larger L3 cache: the 6731P carries 144 MB shared, while the 638 has 72 MB shared. Both processors maintain the same per-core L1 cache of 112 KB and per-core L2 cache of 2 MB.

Clock speeds tell the opposite story. The Intel Xeon 638 operates at a base clock of 3.20 GHz and boosts to 4.80 GHz, while the 6731P runs at 2.50 GHz base and 4.10 GHz boost. This clock advantage explains the 638's dominance in single-threaded and lightly threaded benchmarks. The 6731P does command a higher TDP at 245 watts compared to the 638's 180 watts, a necessary trade-off for doubling the core count.

Memory architecture also differs substantially. The 6731P supports eight-channel DDR5 memory with a bandwidth of 409.6 GB/s, while the 638 uses a quad-channel configuration with 204.8 GB/s bandwidth. Both support ECC memory. PCIe connectivity follows a similar pattern, with the 6731P offering 136 Gen 5 lanes versus 80 lanes on the 638. The 638 has an unlocked multiplier, while the 6731P does not, and the two carry different part numbers: SRVNR for the 6731P and SA2DN for the 638.

FAQ

Q: Why does the Intel Xeon 638 win in Cinebench multi-core despite having half the cores?

A: The 638's higher clock speeds, 3.20 GHz base and 4.80 GHz boost, combined with the benchmark's scaling characteristics, give it a 4.9% edge across all Cinebench tests. The 6731P's 32 cores cannot overcome the clock disadvantage in these specific workloads.

Q: Which processor is better for scientific computing workloads?

A: The Intel Xeon 6731P shows clear advantages in compute-heavy tasks. It leads by 51% in Passmark physics, 42% in prime number finding, and 16.2% in extended instructions, making it the stronger choice for algorithmic and simulation workloads.

Q: How significant is the single-thread performance gap?

A: The Passmark single-thread test shows the 638 at 3670 versus 2107 for the 6731P, a 42.6% difference. This is the largest performance gap between the two processors in any benchmark and indicates a substantial advantage for the 638 in single-threaded applications.

Q: What memory bandwidth do these processors support?

A: The 6731P supports eight-channel DDR5 memory with 409.6 GB/s bandwidth, while the 638 uses quad-channel DDR5 with 204.8 GB/s. Both support ECC memory.

Q: Are both processors on the same manufacturing process?

A: Yes, both are built on Intel's 5 nm process node and share the same Granite Rapids architecture with an identical die size of 598 mm².

Q: Which processor has more PCIe lanes?

A: The Intel Xeon 6731P offers 136 Gen 5 lanes, while the Intel Xeon 638 provides 80 Gen 5 lanes, both counted for CPU-only connectivity.

The Verdict

The benchmark data indicates that the choice between these processors depends entirely on workload profile. The Intel Xeon 638 is the superior option for general-purpose computing, single-threaded applications, and tasks that rely on high clock speeds. Its 4.9% lead across all Cinebench tests and 42.6% advantage in single-thread performance make it the clear pick for workstations running typical productivity software, development tools, and rendering applications that favor per-core speed.

The Intel Xeon 6731P, despite its lower win count, shows commanding leads in several specialized areas. The 51% advantage in physics calculations and 42% in prime number finding suggest that this processor is better suited for computational research, simulation, and data-intensive workloads. The 16.2% lead in extended instructions and the 10-12% advantages in data compression and encryption make it the stronger candidate for database servers, scientific computing clusters, and security applications.

The 6731P's doubled core count, larger L3 cache, and double memory bandwidth (409.6 GB/s versus 204.8 GB/s) provide the foundational resources for these compute-heavy tasks. The 638's higher clocks and lower TDP (180 watts versus 245 watts) make it more efficient for workloads that cannot utilize the 6731P's additional cores. Users with mixed workloads should weigh whether their applications favor the 638's single-thread dominance or the 6731P's specialized compute advantages.

Specification Differences

| Specification | Intel Xeon 6731P | Intel Xeon 638 |

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

| Cores | 32 | 16 |

| Threads | 64 | 32 |

| Base Clock | 2.50 GHz | 3.20 GHz |

| Boost Clock | 4.10 GHz | 4.80 GHz |

| TDP | 245 W | 180 W |

| L3 Cache | 144 MB (shared) | 72 MB (shared) |

| Memory Bus | Eight-channel | Quad-channel |

| Memory Bandwidth | 409.6 GB/s | 204.8 GB/s |

| PCIe | Gen 5, 136 Lanes (CPU only) | Gen 5, 80 Lanes (CPU only) |

| Generation | Xeon 6 (Granite Rapids-SP) | Xeon 600 (Granite Rapids-WS) |

| Multiplier Unlocked | No | Yes |

| Part Number | SRVNR | SA2DN |

| Launch MSRP | $2700 | $899 |

DETAILED SPECIFICATIONS

SPECIFICATION
638
6731P
Core Specs
Cores
16
32 +100.0%
Threads
32
64 +100.0%
Base Clock (GHz)
3.2
2.5 -21.9%
Boost Clock (GHz)
4.8
4.1 -14.6%
Frequency (GHz)
3.2
2.5 -21.9%
Turbo Clock (GHz)
4.8
4.1 -14.6%
Multiplier
32
25 -21.9%
SMP CPUs
1
1 0.0%
Cache
L1 Cache
112 KB (per core)
112 KB (per core)
L2 Cache
2 MB (per core)
2 MB (per core)
L3 Cache
72 MB (shared)
144 MB (shared)
Power
TDP (W)
180
245 +36.1%
Architecture
Architecture
Granite Rapids
Granite Rapids
Codename
Granite Rapids
Granite Rapids
Generation
Xeon 600 (Granite Rapids-WS)
Xeon 6 (Granite Rapids-SP)
Process Size
5 nm
5 nm
Die Size
598 mm²
598 mm²
Foundry
Intel
Intel
Memory
Memory Support
DDR5
DDR5
Memory Bus
Quad-channel
Eight-channel
Memory Bandwidth
204.8 GB/s
409.6 GB/s
ECC Memory
Yes
Yes
Platform
Socket
Intel Socket 4710
Intel Socket 4710
Chipsets
W890
—
PCIe
Gen 5, 80 Lanes(CPU only)
Gen 5, 136 Lanes(CPU only)
AMD Multi-Die
IO Process Size
10 nm
10 nm
Interconnect
CXL
Gen 2.0 (Shared with PCI-E)
Gen 2.0, 64 Lanes (Shared with PCI-E)
Other
Market
Server/Workstation
Server/Workstation
Production Status
Active
Active
Launch Price
$899
$2700
Part Number
SA2DN
SRVNR
Package
FC-LGA18N
FC-LGA18N
Tj Max
100°C
102°C
Bundled Cooler
None
None
View Xeon 638 Details View Xeon 6731P Details