AMD Ryzen Threadripper PRO 9945WX vs Intel Core Ultra 9 275HX Comparison

The AMD Ryzen Threadripper PRO 9945WX and Intel Core Ultra 9 275HX occupy opposite ends of the computing spectrum—one is a workstation-class server processor, the other a high-end mobile chip—yet their average benchmark scores sit within 0.6% of each other, making this a fascinating comparison of architectural priorities.

FAQ

Q: Which processor has the higher average benchmark score?

A: The AMD Ryzen Threadripper PRO 9945WX scores 76,513 on average, while the Intel Core Ultra 9 275HX scores 76,024. The AMD part leads by 0.6%, a margin that places both in the 97th percentile of all CPUs.

Q: How do the two chips compare in multi-threaded Cinebench tests?

A: The Threadripper wins every Cinebench multi-core test by 0.7%. In Cinebench R23 multi-core, AMD scores 48,325 versus Intel’s 48,008, a difference of 317 points.

Q: What is the largest single benchmark victory for either processor?

A: Intel’s biggest win is in PassMark floating point math, where it scores 193,324 versus AMD’s 111,566—a 42.3% advantage. AMD’s largest win is in PassMark physics, scoring 6,118 versus Intel’s 3,377, a 81.2% lead.

Q: Do the processors have the same core and thread counts?

A: No. The Threadripper PRO 9945WX has 12 cores and 24 threads, while the Core Ultra 9 275HX has 24 cores and 24 threads. They share identical thread counts but differ in core count.

Q: Which processor supports ECC memory?

A: Only the AMD Ryzen Threadripper PRO 9945WX supports ECC memory. The Intel Core Ultra 9 275HX does not list ECC support in its specifications.

Q: What is the memory bandwidth difference between the two?

A: AMD’s eight-channel memory bus delivers 409.6 GB/s of bandwidth, while Intel’s dual-channel setup provides 102.4 GB/s. This is a 4x difference in theoretical peak bandwidth.

Architecture Differences

The architectural divide is stark. The AMD Ryzen Threadripper PRO 9945WX is built on Zen 5, using the Shimada Peak codename, and is fabricated on a 4 nm process at TSMC. It comes in a server/workstation market segment, designed for socket sTR5. The Intel Core Ultra 9 275HX uses Arrow Lake architecture, specifically Arrow Lake-HX, on a more advanced 3 nm process, also from TSMC, but targets the mobile segment with a BGA 2114 socket.

Core counts diverge significantly. AMD packs 12 cores and 24 threads, relying on simultaneous multithreading to reach 24 threads. Intel offers 24 physical cores but no hyperthreading, also yielding 24 threads. This means Intel’s thread count comes entirely from core count, while AMD achieves the same thread count from half as many cores. Clock speeds tell a similar story: AMD’s base clock is 4.70 GHz versus Intel’s 2.70 GHz, but both boost to 5.40 GHz—identical maximum frequencies.

Cache hierarchies differ in scale. AMD provides 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3 cache. Intel counters with 192 KB of L1 per core, 3 MB of L2 per core, but only 36 MB of shared L3. The Threadripper’s larger L3 cache is double that of the Core Ultra.

Memory architecture is another major differentiator. AMD supports DDR5 across an eight-channel memory bus, yielding 409.6 GB/s of bandwidth and ECC support. Intel also supports DDR5 but only on a dual-channel bus, capping bandwidth at 102.4 GB/s, with no ECC capability. PCI Express connectivity mirrors this: AMD offers Gen 5 with 128 lanes (CPU only), while Intel provides Gen 5 with 20 lanes.

The Intel part includes integrated graphics—Arc Xe-LPG Graphics with 64 execution units—while AMD lists no integrated graphics. Both processors have unlocked multipliers, though the Threadripper’s 350 W TDP versus Intel’s 55 W TDP reflects their different thermal envelopes and market positions.

Head-to-Head Benchmarks

The benchmark data reveals a processor that wins through sheer consistency and another that wins through specialized strengths. Across 17 head-to-head tests, AMD takes 12 wins to Intel’s 5, but the margins tell a more nuanced story.

In Cinebench, AMD is uniformly ahead, though narrowly. Every Cinebench test—R15, R20, and R23, both single-core and multi-core—shows the Threadripper winning by 0.6% to 0.7%. In R23 multi-core, AMD scores 48,325 versus Intel’s 48,008; in R23 single-core, AMD posts 6,822 versus 6,777. These are small but consistent advantages, suggesting that Zen 5’s architecture holds a slight edge in both single-threaded and multi-threaded rendering workloads.

PassMark tests reveal a more polarized picture. AMD dominates in integer-heavy and physics-based workloads. In PassMark integer math, AMD scores 185,421 versus Intel’s 156,314—an 18.6% lead. The physics test is AMD’s biggest win: 6,118 versus 3,377, an 81.2% gap. AMD also leads in extended instructions (54,406 versus 47,687, +14.1%), random string sorting (84,498 versus 75,710, +11.6%), and data compression (671,963 versus 618,338, +8.7%). Multi-threaded performance is close but favors AMD: 56,854 versus 56,415, a 0.8% margin.

Intel’s wins are concentrated in specific computational areas. The most dramatic is floating point math: Intel scores 193,324 versus AMD’s 111,566, a 42.3% advantage that suggests fundamentally different FPU designs. Intel also wins data encryption decisively: 48,016 versus 36,540, a 23.9% lead. Prime number finding favors Intel at 457 versus 335 (+26.7%), and single-threaded PassMark tests show Intel ahead at 4,727 versus 4,573 (+3.3%). These five wins indicate Intel’s architecture excels in cryptography, floating-point throughput, and certain integer operations.

Specification Differences

The specifications where these two processors diverge are numerous and fundamental:

• Cores: 12 (AMD) versus 24 (Intel)
• Threads: 24 for both
• Base clock: 4.70 GHz (AMD) versus 2.70 GHz (Intel)
• Boost clock: 5.40 GHz for both
• TDP: 350 W (AMD) versus 55 W (Intel)
• Socket: AMD Socket sTR5 versus Intel BGA 2114
• Architecture: Zen 5 (Shimada Peak) versus Arrow Lake (Arrow Lake-HX)
• Process node: 4 nm versus 3 nm
• L1 cache per core: 64 KB versus 192 KB
• L2 cache per core: 1 MB versus 3 MB
• L3 cache (shared): 64 MB versus 36 MB
• Memory bus: Eight-channel versus dual-channel
• Memory bandwidth: 409.6 GB/s versus 102.4 GB/s
• ECC memory: Supported versus not supported
• PCIe lanes: 128 (Gen 5) versus 20 (Gen 5)
• Integrated graphics: None versus Arc Xe-LPG Graphics 64EU
• Market segment: Server/Workstation versus Mobile
• Release date: Not listed versus 2025-01-12

The Verdict

The data presents a clear picture: the AMD Ryzen Threadripper PRO 9945WX is the more consistent all-round performer, winning 12 of 17 benchmarks and taking every Cinebench test. Its 81.2% lead in physics, 18.6% lead in integer math, and 14.1% lead in extended instructions make it the stronger choice for general workstation tasks, rendering, and integer-heavy computations. The eight-channel memory bus with 409.6 GB/s bandwidth and ECC support further cement its workstation credentials.

The Intel Core Ultra 9 275HX is not simply a loser, however. Its 42.3% advantage in floating point math and 23.9% lead in data encryption point to specific workloads where Intel’s architecture is superior. The 24-core design without hyperthreading delivers these wins while maintaining a 55 W TDP—a fraction of AMD’s 350 W envelope. For mobile users needing high floating-point throughput or encryption performance in a power-constrained package, the Core Ultra is the data-backed selection.

The choice hinges on workload and platform. For a workstation user who needs maximum memory bandwidth, ECC reliability, and broad multi-threaded performance, the Threadripper PRO 9945WX is the clear winner—its average score of 76,513 edges out Intel’s 76,024, and it wins the majority of tests. For a mobile user who prioritizes floating-point math, encryption, and power efficiency, the Core Ultra 9 275HX offers specialized strengths that the Threadripper cannot match. Both sit in the 97th percentile of all CPUs, but they achieve that ranking through fundamentally different means.