AMD Ryzen Threadripper PRO 5965WX vs Intel Xeon 654 Comparison

FAQ

Q: Which processor delivers the higher multi-threaded performance in Cinebench?

A: The AMD Ryzen Threadripper PRO 5965WX wins every Cinebench multi-core test. It scores 5685 in Cinebench R15 multi-core versus 5256 for the Intel Xeon 654, an 8.2% lead. In Cinebench R23 multi-core, the AMD scores 56400 against 52150, an 8.1% advantage.

Q: Does the Intel Xeon 654 win any benchmark categories?

A: Yes, the Intel Xeon 654 wins four of the 17 recorded head-to-head tests. It leads in PassMark floating point math (163093 vs 157708, a 3.3% edge), PassMark physics (5596 vs 4251, a 24% advantage), and PassMark single-thread (3778 vs 3337, an 11.7% lead).

Q: How do the two processors compare in data encryption workloads?

A: The AMD Ryzen Threadripper PRO 5965WX is substantially faster. It records a PassMark data encryption score of 63113 versus 40675 for the Intel Xeon 654, a 55.2% difference. This is the largest performance gap in the entire benchmark set.

Q: What is the difference in average benchmark scores between the two?

A: The AMD Ryzen Threadripper PRO 5965WX has an average benchmark score of 98504, placing it in the 97th percentile of all CPUs. The Intel Xeon 654 averages 90717, placing it in the 96th percentile. The AMD holds a roughly 8.6% lead in average score.

Q: Which processor has a higher boost clock?

A: The Intel Xeon 654 has a higher boost clock at 4.80 GHz compared to 4.50 GHz for the AMD Ryzen Threadripper PRO 5965WX. However, the AMD has a higher base clock at 3.80 GHz versus 3.10 GHz for the Intel.

Q: What are the memory bandwidth specifications for each processor?

A: The AMD Ryzen Threadripper PRO 5965WX supports DDR4 memory with an eight-channel bus and a bandwidth of 204.8 GB/s. The Intel Xeon 654 supports DDR5 memory with an eight-channel bus and a bandwidth of 409.6 GB/s, exactly double the AMD’s figure.

Architecture Differences

The AMD Ryzen Threadripper PRO 5965WX is built on the Zen 3 architecture, specifically the Chagall PRO codename, using a 7 nm process from TSMC. It packs 24 cores and 48 threads. The die size is reported as 4x 81 mm², with a transistor count of 16,600 million. The cache hierarchy consists of 64 KB of L1 per core, 512 KB of L2 per core, and a large 128 MB L3 cache. It uses the AMD Socket WRX8 and supports PCIe Gen 4 with 128 lanes from the CPU.

The Intel Xeon 654 uses the Granite Rapids architecture, also referred to as Granite Rapids-WS, on a 5 nm process from Intel. It has 18 cores and 36 threads. The die size is 2x 598 mm², which is substantially larger physically than the AMD chip. Its cache layout differs notably: 112 KB of L1 per core, 2 MB of L2 per core, and a shared 72 MB L3 cache. The Intel part uses Socket 4710 and supports PCIe Gen 5 with 128 lanes from the CPU.

Several architectural contrasts stand out. The AMD offers 6 more cores and 12 more threads, which aligns with its multi-core benchmark dominance. The Intel part has a higher boost clock by 0.30 GHz, but the AMD has a higher base clock by 0.70 GHz. The AMD’s L3 cache is 56 MB larger, while the Intel’s per-core L2 cache is 4x larger. The Intel Xeon 654 moves to DDR5 memory support, doubling the memory bandwidth to 409.6 GB/s compared to the AMD’s 204.8 GB/s with DDR4. The AMD’s TDP is 280 W, while the Intel is rated at 200 W, a difference of 80 W. The Intel part has an unlocked multiplier, whereas the AMD does not. Neither processor includes integrated graphics.

Head-to-Head Benchmarks

The recorded data shows the AMD Ryzen Threadripper PRO 5965WX winning 13 of the 17 head-to-head benchmark comparisons. The Intel Xeon 654 wins the remaining four.

The AMD’s dominance is most pronounced in compute-heavy integer and encryption tasks. In PassMark integer math, the AMD scores 281247 against 207745 for the Intel, a 35.4% advantage. PassMark data encryption shows an even larger gap: 63113 versus 40675, a 55.2% lead. The AMD also wins PassMark find prime numbers by 34.1% (523 vs 390) and PassMark data compression by 23.3% (1010067 vs 818902). PassMark random string sorting goes to the AMD by 19.9% (99287 vs 82828).

In the Cinebench suite, the AMD’s wins are consistent but narrower. Across all six Cinebench tests (R15, R20, R23, both single-core and multi-core), the AMD leads by between 8.1% and 8.2%. For instance, Cinebench R20 multi-core shows 23688 for the AMD versus 21903 for the Intel, an 8.1% edge. The single-core results follow the same pattern, with the AMD ahead by 8.1% or 8.2% in each case. PassMark multithread also favors the AMD by 8.1% (66353 vs 61353). PassMark extended instructions goes to the AMD by 5.9% (67300 vs 63539).

The Intel Xeon 654’s wins are concentrated in specific workloads. PassMark physics shows the Intel ahead by 24% (5596 vs 4251), which is its largest margin. PassMark single-thread (and the duplicate singlethread entry) gives the Intel an 11.7% lead (3778 vs 3337). The Intel also wins PassMark floating point math, but by a smaller margin of 3.3% (163093 vs 157708).

The overall pattern is clear: the AMD wins nearly all throughput-oriented tests, often by large margins, while the Intel takes physics, single-thread, and floating point math. The average benchmark score difference reinforces this, with the AMD at 98504 versus 90717 for the Intel.

The Verdict

The data points to a clear overall winner for general workstation throughput. The AMD Ryzen Threadripper PRO 5965WX wins 13 of 17 benchmark comparisons and holds an average benchmark score of 98504, which is 8.6% higher than the Intel Xeon 654’s 90717. The AMD’s 24 cores and 48 threads, combined with its 128 MB L3 cache, deliver decisive advantages in integer math, encryption, compression, and multi-core rendering. For workloads that scale with core count and cache size, the AMD is the stronger choice.

The Intel Xeon 654 has specific strengths that should not be ignored. Its 24% lead in PassMark physics and 11.7% lead in PassMark single-thread indicate that it handles certain latency-sensitive or lightly threaded tasks better. The Intel also wins floating point math by 3.3%. Its DDR5 memory support doubles the recorded memory bandwidth to 409.6 GB/s, which could matter in memory-bound scenarios that are not captured in these particular benchmarks. Its 200 W TDP is lower than the AMD’s 280 W, which may be relevant in power-constrained environments.

Who should pick which? Users whose primary workloads involve multi-threaded rendering, data encryption, compression, or integer-heavy computation should select the AMD Ryzen Threadripper PRO 5965WX based on its consistent and often large margins. Users who prioritize physics simulation, single-thread responsiveness, or floating point math may find the Intel Xeon 654 more suitable. The Intel’s higher boost clock of 4.80 GHz and unlocked multiplier also offer tuning headroom that the AMD lacks. The choice depends on workload mix, not overall capability.

Specification Differences

The two processors differ across nearly every core specification. The AMD has 24 cores and 48 threads, while the Intel has 18 cores and 36 threads. The AMD’s base clock is 3.80 GHz, higher than the Intel’s 3.10 GHz. The Intel’s boost clock is 4.80 GHz, higher than the AMD’s 4.50 GHz. The AMD’s TDP is 280 W, while the Intel’s is 200 W.

The socket and platform differ completely. The AMD uses Socket WRX8, while the Intel uses Socket 4710. The AMD is based on a 7 nm process from TSMC, while the Intel uses a 5 nm process from Intel. The AMD’s die size is 4x 81 mm², while the Intel’s is 2x 598 mm². The AMD has 16,600 million transistors; the Intel’s transistor count is not recorded.

Cache configurations diverge significantly. The AMD has 64 KB L1 per core, 512 KB L2 per core, and 128 MB L3. The Intel has 112 KB L1 per core, 2 MB L2 per core, and 72 MB shared L3. The memory support differs: the AMD uses DDR4 with 204.8 GB/s bandwidth, while the Intel uses DDR5 with 409.6 GB/s bandwidth. Both support eight-channel memory and ECC. PCIe support differs by generation: the AMD has Gen 4 with 128 lanes, the Intel has Gen 5 with 128 lanes.

The Intel Xeon 654 has an unlocked multiplier, while the AMD does not. The Intel lists integrated graphics as N/A, and the AMD has no integrated graphics. The release dates differ by nearly four years, with the AMD launching earlier and the Intel later. The Intel’s launch MSRP is $1199.

Where Each One Wins

The AMD Ryzen Threadripper PRO 5965WX wins in the majority of recorded workloads, making it the default pick for multi-threaded throughput. It excels in data encryption with a 55.2% lead, integer math with a 35.4% lead, and prime number finding with a 34.1% lead. These results suggest strong performance in cryptography, financial modeling, and integer-heavy scientific computing. Its 23.3% advantage in data compression points to wins in archiving and database workloads. The AMD also leads all six Cinebench tests by roughly 8%, indicating solid performance in 3D rendering and video encoding. PassMark multithread and random string sorting also favor the AMD, with leads of 8.1% and 19.9% respectively.

The Intel Xeon 654 wins in four specific areas. Its 24% lead in PassMark physics suggests an edge in simulation and physics-based workloads, likely due to its higher boost clock and different cache topology. Its 11.7% lead in PassMark single-thread indicates better raw single-core performance, which benefits older or less parallel software. Its 3.3% lead in floating point math points to an advantage in certain scientific and engineering applications that rely heavily on FPU throughput. Additionally, the Intel’s DDR5 memory support and doubled memory bandwidth of 409.6 GB/s may provide an advantage in memory-bandwidth-bound tasks, even though this is not directly reflected in the recorded benchmark wins.

For a user deciding between the two, the AMD offers a broader base of wins across most categories, with particularly large margins in security and integer workloads. The Intel offers targeted wins in physics and single-thread performance, plus a lower TDP and higher boost clock. The data supports choosing the AMD for general workstation use, while the Intel is the better fit for workloads that match its specific winning categories.