CPU Comparison

The Verdict

The benchmark data splits these two processors into clearly different roles. The AMD Ryzen Threadripper 9980X wins 10 of 14 head-to-head tests, including every Cinebench multi-core run, and holds a decisive single-thread advantage. The Intel Xeon 6960P wins 4 tests, but those wins are massive in magnitude, a 211.7% lead in Passmark physics and a 93% lead in prime number finding. The Xeon 6960P posts an average benchmark score of 365,194 against the Threadripper’s 321,753, a gap driven by its lopsided victories in specialized workloads. The Threadripper 9980X is the general-purpose compute champion; the Xeon 6960P is the specialist for integer-heavy, physics-simulation, and sorting tasks. For users whose workloads mirror the Cinebench suite, the AMD part is the clear pick. For those running physics solvers, prime-number searches, or string-sorting pipelines, the Intel part’s extreme deltas justify its selection despite losing the overall benchmark count.

Architecture Differences

The two processors diverge fundamentally at every architectural layer. The Intel Xeon 6960P uses the Granite Rapids architecture on a 5 nm Intel process, built as three 598 mm² dies. It packs 72 cores and 144 threads, with a base clock of 2.70 GHz and a boost clock of 3.90 GHz. Its cache hierarchy is generous: 112 KB of L1 per core, 2 MB of L2 per core, and a shared 432 MB of L3. Memory connectivity spans twelve channels of DDR5, yielding 614.4 GB/s of bandwidth. PCIe support reaches Gen 5 with 96 CPU lanes. The chip draws a 500 W TDP and fits the Intel Socket 7529, targeting the server and workstation segment. Its multiplier is locked.

The AMD Ryzen Threadripper 9980X belongs to the Zen 5 architecture, codenamed Shimada Peak, fabricated on a 4 nm TSMC process using eight 70.6 mm² dies. It contains 64 cores and 128 threads, with a higher base clock of 3.20 GHz and a much higher boost clock of 5.40 GHz. Cache amounts are smaller per core, 64 KB L1 and 1 MB L2, with a 256 MB shared L3. Memory runs quad-channel DDR5 at 204.8 GB/s, a third of the Xeon’s bandwidth. PCIe provides Gen 5 across 80 lanes. The TDP is 350 W, the socket is AMD Socket sTR5, and the market segment is desktop. The multiplier is unlocked, and the die carries 66,520 million transistors. The Xeon counters with no integrated graphics on either part, but the Intel chip’s die size and cache capacity are substantially larger, while the AMD part’s process node is smaller and its clock speeds are far higher.

Head-to-Head Benchmarks

The Cinebench suite shows a consistent pattern. In Cinebench R15 multi-core, the Threadripper scores 13,157 against the Xeon’s 11,194, a 14.9% advantage. The same 14.9% delta appears in Cinebench R20 (54,822 vs 46,645) and Cinebench R23 (130,529 vs 111,060). The AMD part wins all three multi-core Cinebench tests by an identical margin, suggesting a stable architectural efficiency gain rather than a workload-specific quirk.

Single-thread performance is the AMD part’s largest win. Passmark single-thread scores show 4,537 for the Threadripper versus 3,287 for the Xeon, a 27.6% lead. Cinebench single-core results reinforce this: 18,427 in R23 and 7,739 in R20, numbers the Xeon does not have benchmark entries for in the data. The Threadripper’s 5.40 GHz boost clock clearly carries the day.

The Intel Xeon’s wins are narrower in number but extreme in scale. Passmark physics shows the Xeon at 24,937 versus the Threadripper’s 8,001, a 211.7% advantage. Prime number finding goes to the Xeon at 1,484 against 769, a 93% lead. Random string sorting favors the Xeon by 27.3% (371,795 vs 292,083). Data encryption is a closer Intel win at 3.1% (162,013 vs 157,137).

The remaining Passmark tests split toward AMD. Integer math shows 872,071 vs 727,750, a 16.5% AMD lead. Extended instructions go 228,959 vs 193,404, a 15.5% margin. Multithread scores 141,641 vs 130,659, a 7.8% edge. Floating-point math is close at 559,003 vs 527,473, a 5.6% AMD win. Data compression also favors AMD at 2,974,534 vs 2,797,724, a 5.9% margin.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Xeon 6960P has 72 cores and 144 threads, while the AMD Ryzen Threadripper 9980X has 64 cores and 128 threads.

Q: What is the single-thread performance difference?

A: The Threadripper 9980X scores 4,537 in Passmark single-thread, which is 27.6% higher than the Xeon 6960P’s 3,287.

Q: How do they compare in Cinebench R23 multi-core?

A: The Threadripper 9980X scores 130,529, which is 14.9% higher than the Xeon 6960P’s 111,060.

Q: Where does the Xeon 6960P have its biggest advantage?

A: In Passmark physics, the Xeon scores 24,937 versus 8,001 for the Threadripper, a 211.7% lead. It also leads prime number finding by 93% and random string sorting by 27.3%.

Q: What are the memory bandwidth specifications?

A: The Xeon 6960P supports twelve-channel DDR5 with 614.4 GB/s bandwidth, while the Threadripper 9980X supports quad-channel DDR5 with 204.8 GB/s.

Q: Which chip has a higher boost clock?

A: The Threadripper 9980X boosts to 5.40 GHz, compared to the Xeon 6960P’s 3.90 GHz.

Where Each One Wins

The AMD Ryzen Threadripper 9980X wins every rendering and content-creation workload in the dataset. Its Cinebench R15, R20, and R23 multi-core scores all lead by 14.9%, making it the stronger choice for 3D rendering, video encoding, and any task that scales across threads with balanced workloads. The single-thread advantage of 27.6% extends to general desktop responsiveness, legacy single-threaded applications, and lightly threaded code. In Passmark integer math, the AMD part leads by 16.5%, which benefits compilers, database operations, and financial modeling. Extended instructions show a 15.5% AMD edge, relevant for SIMD-heavy scientific code. Multithread performance favors AMD by 7.8%, and floating-point math by 5.6%, covering most numerical simulation workloads. Data compression also goes to AMD by 5.9%, aiding archiving and storage pipelines.

The Intel Xeon 6960P wins where raw throughput per instruction matters more than clock speed. Its 211.7% lead in Passmark physics indicates a massive advantage for rigid-body dynamics, particle simulations, and game-physics-style computations. The 93% lead in prime number finding points to integer-heavy, branch-predictable loops where the Xeon’s larger cache and core count dominate. Random string sorting, a memory-latency-sensitive workload, goes to Intel by 27.3%, suggesting better cache behavior for pointer-chasing patterns. Data encryption favors Intel by 3.1%, a modest edge for security workloads.

The overall benchmark averages reflect this split: the Xeon’s 365,194 average score edges out the Threadripper’s 321,753, but that is driven by the Xeon’s extreme outliers in physics and prime numbers. The Threadripper wins the majority of tests and all the mainstream rendering benchmarks, making it the default recommendation for mixed-use workstations. The Xeon is the specialist’s choice for physics-heavy HPC, cryptographic workloads, and sorting-intensive data processing. Its 432 MB L3 cache and 614.4 GB/s memory bandwidth provide the infrastructure for those wins, while the Threadripper’s 5.40 GHz boost clock and 4 nm process deliver higher per-thread efficiency. The Xeon’s 500 W TDP and locked multiplier position it as a fixed-configuration server part; the Threadripper’s 350 W TDP and unlocked multiplier offer overclocking flexibility for desktop enthusiasts.