AMD Ryzen Threadripper PRO 9965WX vs Intel Xeon 6737P Comparison

The AMD Ryzen Threadripper PRO 9965WX and Intel Xeon 6737P are both 98th-percentile server/workstation processors, yet they deliver their performance through fundamentally different designs. The AMD part wins 14 of 16 head-to-head benchmarks, while the Intel part claims two decisive victories in specialized workloads. This analysis breaks down where each processor excels, what drives those differences, and which workloads favor which silicon.

Where Each One Wins

The AMD Ryzen Threadripper PRO 9965WX is the dominant all-rounder, taking every Cinebench test and the majority of Passmark workloads. Its wins span single-threaded performance, multi-threaded rendering, data compression, integer math, and random string sorting. The 49.3% lead in Passmark single-thread scoring is the largest margin of any comparison, indicating a substantial per-core advantage that benefits virtually every application.

The Intel Xeon 6737P wins only two benchmarks, but they are telling. It leads by 11.3% in floating-point math and by 19.6% in physics simulation. These are compute-heavy, vectorized workloads where the Xeon’s 32 cores and larger per-core L2 cache provide an edge. For users running scientific simulations, physics engines, or financial modeling that relies heavily on floating-point throughput, the Intel part is the better choice despite losing most other tests.

The AMD processor also holds a 16.3% lead in Passmark multithread score and a 19.6% lead across all Cinebench versions, from R15 to R23. This consistency suggests that the Threadripper PRO 9965WX is the safer pick for mixed workloads, while the Xeon 6737P is a specialized tool for floating-point-heavy tasks.

Architecture Differences

The two processors come from different fabs and design philosophies. The AMD Ryzen Threadripper PRO 9965WX is built on TSMC’s 4 nm process with a 33,260 million transistor count spread across four 70.6 mm² chiplets. It uses Zen 5 architecture under the "Shimada Peak" codename, featuring 24 cores and 48 threads. The Intel Xeon 6737P uses Intel’s 5 nm process with a single 598 mm² die, employing Granite Rapids architecture with 32 cores and 64 threads.

Cache hierarchies diverge significantly. AMD allocates 64 KB of L1 and 1 MB of L2 per core, with 128 MB of shared L3. Intel provides 112 KB of L1 and 2 MB of L2 per core, plus 144 MB of shared L3. The larger per-core L2 on Intel helps explain its floating-point advantage, while AMD’s larger total L3 and higher clock speeds drive its single-thread dominance.

Clock speeds tell a clear story. The AMD part runs at 4.20 GHz base and 5.40 GHz boost, compared to Intel’s 2.90 GHz base and 4.00 GHz boost. The 1.4 GHz boost advantage is enormous and directly translates to the 49.3% single-thread lead. Both processors support DDR5 memory across eight channels with identical 409.6 GB/s bandwidth and ECC support, so memory capacity and speed are not differentiating factors.

PCIe connectivity differs: AMD offers 128 Gen 5 lanes, while Intel provides 88 Gen 5 lanes. The AMD processor has an unlocked multiplier, whereas Intel’s is locked. The AMD part also draws more power at 350 W TDP versus Intel’s 270 W, which is consistent with its higher clock speeds. Both are active production parts with no integrated graphics, targeting the server and workstation market segment.

Head-to-Head Benchmarks

The most striking result is in Passmark single-thread performance, where AMD scores 4551 against Intel’s 3048, a 49.3% delta. This is the largest gap in any test and stems directly from the 5.40 GHz boost clock versus 4.00 GHz. The Cinebench single-core tests show the same pattern, with AMD leading 1152 to 963 in R15 and 4801 to 4013 in R20, each at 19.6%.

Multi-core Cinebench results are equally lopsided. In R23, AMD scores 80976 against Intel’s 67688, a 19.6% advantage. The R15 multicore test shows 8162 versus 6822, and R20 shows 34009 versus 28428, both at the same 19.6% delta. This consistency across Cinebench versions indicates a stable architectural advantage rather than a workload-specific quirk.

Passmark multithread scoring gives AMD a 16.3% win at 92604 versus 79634. Data compression follows at 16.2% (1345230 versus 1157255), and random string sorting is 15.5% higher (149617 versus 129510). Integer math shows a 5.6% AMD lead (349195 versus 330756), while prime number finding is 7.9% higher (752 versus 697). Extended instructions favor AMD by 3.1% (108753 versus 105453), and data encryption is nearly tied at 0.8% (66155 versus 65615).

The Intel wins are substantial. Floating-point math shows Intel at 258811 versus AMD’s 229685, an 11.3% lead. Physics simulation is even more decisive: Intel scores 9362 against AMD’s 7529, a 19.6% margin. These two results suggest that Intel’s architecture, despite lower clock speeds, has superior floating-point execution units or better vector instruction handling.

The Verdict

For general-purpose workstation tasks, the AMD Ryzen Threadripper PRO 9965WX is the clear choice. It wins 14 of 16 benchmarks, holds a 19.6% lead across all Cinebench versions, and offers a 49.3% single-thread advantage that makes it feel faster in nearly every interactive application. The 128 PCIe lanes and unlocked multiplier add flexibility for high-expansion systems and overclocking.

The Intel Xeon 6737P is the pick only for floating-point-heavy or physics-simulation workloads. Its 11.3% floating-point lead and 19.6% physics advantage are meaningful for computational fluid dynamics, structural analysis, or any task dominated by dense math operations. The 32 cores and 64 threads also provide more parallel capacity, though the lower clock speeds limit its general responsiveness.

The data shows that AMD’s higher clocks and Zen 5 efficiency overcome Intel’s core-count advantage in most scenarios. The Xeon’s 270 W TDP is lower, which may matter in dense server deployments, but the Threadripper PRO 9965WX’s performance per watt in benchmark scores is superior across the majority of tests. For mixed workloads, the AMD processor is the safer recommendation; for specialized floating-point compute, the Intel processor is justified.

FAQ

Q: Which processor has better single-thread performance?

A: The AMD Ryzen Threadripper PRO 9965WX leads by 49.3% in Passmark single-thread (4551 versus 3048) and by 19.6% in Cinebench R15 and R20 single-core tests.

Q: How do the core counts compare?

A: The Intel Xeon 6737P has 32 cores and 64 threads, while the AMD Ryzen Threadripper PRO 9965WX has 24 cores and 48 threads. Intel has 8 more cores and 16 more threads.

Q: What is the largest benchmark margin between them?

A: The largest margin is in Passmark single-thread, where AMD leads by 49.3%. The second-largest is a tie at 19.6% across all Cinebench tests and Passmark physics, with AMD winning the former and Intel winning the latter.

Q: Does the Intel Xeon 6737P win any benchmarks?

A: Yes, it wins Passmark floating-point math by 11.3% and Passmark physics by 19.6%.

Q: Which processor has higher clock speeds?

A: The AMD processor runs at 4.20 GHz base and 5.40 GHz boost, compared to Intel’s 2.90 GHz base and 4.00 GHz boost.

Q: How does memory bandwidth compare between the two?

A: Both support eight-channel DDR5 with identical 409.6 GB/s memory bandwidth and ECC support, so there is no difference in memory throughput.