AMD Ryzen Threadripper PRO 9995WX vs Intel Xeon 6960P Comparison

Head-to-Head Benchmarks

The AMD Ryzen Threadripper PRO 9995WX dominates this comparison, winning 12 of the 14 shared benchmarks. The most decisive margin comes in PassMark integer math, where AMD scores 1,203,634 against Intel's 727,750 — a 65.4% advantage. This is the single largest gap in the entire dataset, and it reflects the Threadripper's 96 cores versus the Xeon's 72, a 24-core difference that scales directly into integer-heavy workloads.

Cinebench results tell a consistent story. Across all three versions — R15, R20, and R23 — the AMD part leads by exactly 33.8%. The R23 multicore score of 148,601 versus 111,060 puts roughly 37,500 points between them, which is a substantial margin for any rendering or 3D simulation workload. The consistency of that 33.8% delta across all three Cinebench versions suggests the advantage is structural rather than workload-specific.

PassMark extended instructions shows AMD ahead by 39.9%, scoring 270,647 versus 193,404. This matters for AVX-512 and similar vectorized code paths commonly found in scientific computing and AI inference. Floating-point math follows a similar pattern: 725,066 versus 527,473, a 37.5% lead. Data compression favors AMD by 33.1% (3,723,652 versus 2,797,724), while data encryption shows a 27.6% edge (206,662 versus 162,013). These are all multi-threaded, throughput-oriented tests where core count dominates.

Single-thread performance is not close either. AMD scores 4,542 in PassMark single-thread versus Intel's 3,287 — a 38.2% advantage. This is notable because the Xeon 6960P has a higher base clock (2.70 GHz versus 2.50 GHz), yet the Threadripper's 5.40 GHz boost clock and Zen 5 architecture deliver significantly better per-thread results. The Cinebench R15 single-core score of 2,114 for AMD versus Intel's absence of a comparable entry reinforces this picture.

Intel wins exactly two tests, and both are narrow in different ways. PassMark find prime numbers goes to Intel by a hair: 1,484 versus 1,476, a 0.5% margin that is essentially noise. PassMark physics is a more substantial Intel victory: 24,937 versus 16,860, a 32.4% advantage. Physics simulation often rewards different cache hierarchies and memory subsystems, and the Xeon's 432 MB shared L3 cache appears to help here. Still, one meaningful win out of fourteen benchmarks is a lopsided result.

Random string sorting is the closest AMD victory at 12.7% (418,973 versus 371,795). This test is memory-latency sensitive, and the Xeon's twelve-channel memory bus provides 614.4 GB/s of bandwidth versus AMD's eight-channel 409.6 GB/s. Despite that bandwidth disadvantage, AMD still wins. The multithread benchmark shows AMD ahead by 31% (171,200 versus 130,659), which aligns with the overall pattern: more cores, higher clocks, and superior per-core efficiency.

The Verdict

The data is unambiguous: the AMD Ryzen Threadripper PRO 9995WX is the faster processor in nearly every measurable way. It leads in 12 of 14 benchmarks, including all Cinebench tests, all PassMark math tests, and both compression and encryption workloads. The average benchmark score of 412,068 versus 365,194 puts AMD 12.8% ahead overall, and its 100th percentile ranking among all CPUs matches Intel's — both are at the very top of the database.

For workloads that prioritize raw throughput — rendering, code compilation, data processing, scientific simulation — the Threadripper PRO 9995WX is the clear choice. Its 96 cores and 192 threads provide a 33% core advantage over the Xeon's 72 cores and 144 threads, and the benchmark data shows that advantage translating into real performance gains. The 5.40 GHz boost clock also gives it a decisive edge in single-threaded tasks, which the 38.2% PassMark single-thread lead confirms.

The Intel Xeon 6960P has exactly two niches in this comparison. First, physics simulation: its 24,937 PassMark physics score beats AMD by 32.4%, suggesting that workloads with heavy physics calculations (certain engineering simulation, some game-server physics) would perform better on Intel. Second, find prime numbers: the 0.5% margin is technically a win, but it is within run-to-run variance and should not drive purchasing decisions.

The Xeon's twelve-channel memory bus and 614.4 GB/s bandwidth are objectively superior to AMD's eight-channel 409.6 GB/s configuration. However, the benchmark data shows that this bandwidth advantage does not translate into wins outside of physics. If memory bandwidth is the sole criterion, Intel wins; if performance across a broad suite matters, AMD wins decisively.

FAQ

Q: Which processor has more cores and threads?

A: The AMD Ryzen Threadripper PRO 9995WX has 96 cores and 192 threads, while the Intel Xeon 6960P has 72 cores and 144 threads — a 24-core and 48-thread advantage for AMD.

Q: How much faster is the AMD chip in Cinebench R23 multicore?

A: AMD scores 148,601 versus Intel's 111,060, a 33.8% advantage. This delta is identical across Cinebench R15, R20, and R23.

Q: Does the Intel Xeon win any benchmarks?

A: Yes, two: PassMark physics (24,937 versus 16,860, a 32.4% lead) and PassMark find prime numbers (1,484 versus 1,476, a 0.5% margin).

Q: What is the single-thread performance difference?

A: In PassMark single-thread, AMD scores 4,542 versus Intel's 3,287 — a 38.2% advantage for AMD. This is despite Intel's higher base clock of 2.70 GHz versus AMD's 2.50 GHz.

Q: Which processor has higher memory bandwidth?

A: The Intel Xeon 6960P has a twelve-channel memory bus delivering 614.4 GB/s, versus AMD's eight-channel 409.6 GB/s. Intel's bandwidth is 50% higher, yet AMD still wins most bandwidth-sensitive benchmarks.

Q: What is the average benchmark score difference?

A: AMD's average benchmark score is 412,068 versus Intel's 365,194. AMD's closest rival, the AMD EPYC 9745, scores 425,973, which is 3.3% higher than the Threadripper.

Specification Differences

The core and thread counts differ substantially: AMD offers 96 cores and 192 threads, while Intel provides 72 cores and 144 threads. Clock speeds tell a different story — Intel has a higher base clock at 2.70 GHz versus AMD's 2.50 GHz, but AMD's boost clock of 5.40 GHz far exceeds Intel's 3.90 GHz, a 1.50 GHz gap that explains AMD's single-thread dominance.

Thermal design power differs by 150 watts: AMD is rated at 350W TDP, Intel at 500W. This makes Intel the more power-hungry part despite having fewer cores, which is notable for cooling and power-delivery planning in workstation builds.

Memory architecture is another key split. AMD uses DDR5 with an eight-channel bus and 409.6 GB/s bandwidth, while Intel uses DDR5 with a twelve-channel bus and 614.4 GB/s bandwidth. Both support ECC memory. PCIe connectivity also differs: AMD provides 128 Gen 5 lanes (CPU only), Intel provides 96 Gen 5 lanes.

The socket and platform are entirely different: AMD uses Socket sTR5, Intel uses Socket 7529. AMD's multiplier is unlocked, Intel's is locked, meaning AMD allows overclocking while Intel does not. AMD's launch MSRP is $11700, Intel's is $9625. Release dates differ by roughly ten months: AMD launched on 2025-07-22, Intel on 2024-09-23.

Architecture Differences

AMD's Ryzen Threadripper PRO 9995WX is built on Zen 5 architecture with the codename Shimada Peak, manufactured on a 4 nm process at TSMC. Intel's Xeon 6960P uses Granite Rapids architecture (same codename), manufactured on a 5 nm process at Intel's own foundry. The process node difference — 4 nm versus 5 nm — partially explains AMD's clock-speed and efficiency advantages.

Cache configurations differ in structure and size. AMD provides 64 KB L1 per core and 1 MB L2 per core, with 384 MB of L3 cache. Intel provides 112 KB L1 per core and 2 MB L2 per core, with 432 MB of shared L3 cache. Intel's L3 is 48 MB larger, but it is shared across the chip rather than per-CCX, and the benchmark data shows this does not confer a broad advantage.

AMD's transistor count is listed at 99,780 million across a 12-chiplet design with each die measuring 70.6 mm². Intel does not list a transistor count, but its die size is 3x 598 mm² — three large monolithic dies. This is a fundamental architectural difference: AMD uses a chiplet approach with many small dies, Intel uses fewer, much larger dies.

Both processors target the server/workstation market segment and have no integrated graphics. AMD's part number is 100-000001361, Intel's is SRPKX. Both are currently active in production. The Xeon 6960P belongs to the Xeon 6 family (Granite Rapids-AP), while the Threadripper PRO is part of AMD's 9000 series.

Where Each One Wins

The AMD Ryzen Threadripper PRO 9995WX is the winner for rendering and 3D workloads. Cinebench R23's 33.8% lead directly translates to faster final-frame renders in applications like Blender or Maya. The same applies to video encoding and transcoding, where the 31% PassMark multithread advantage and 37.5% floating-point lead will reduce batch processing times substantially.

Data-heavy workloads favor AMD. Data compression shows a 33.1% lead, data encryption a 27.6% lead, and extended instructions (AVX-512 and similar) a 39.9% lead. For database workloads, scientific computing, or any throughput-oriented server task, the Threadripper's 96 cores provide a structural advantage that Intel's higher memory bandwidth cannot offset.

Single-threaded applications — legacy software, certain scripting engines, lightly-threaded CAD tools — clearly favor AMD. The 38.2% PassMark single-thread lead and the 5.40 GHz boost clock mean that even applications that use one or two cores will run significantly faster on AMD.

The Intel Xeon 6960P wins specifically in physics simulation. The 32.4% PassMark physics lead suggests that physics engines, rigid-body simulations, and particle systems in engineering or scientific contexts will perform better on Intel. This is likely tied to the 432 MB shared L3 cache and the twelve-channel memory bus providing 614.4 GB/s bandwidth.

Memory-bandwidth-bound workloads are Intel's theoretical domain. With 50% more bandwidth (614.4 GB/s versus 409.6 GB/s), applications that stream large datasets through memory — certain financial risk models, some HPC kernels — could benefit from Intel. However, the benchmark suite shows AMD winning random string sorting (a bandwidth-sensitive test) by 12.7%, so this theoretical advantage does not consistently materialize in practice.

For prime-number finding, Intel wins by a 0.5% margin that is statistically insignificant. This should not influence any purchasing decision. Every other benchmark, from integer math to multithread performance, points to AMD as the superior processor for virtually all workloads.