AMD Ryzen Threadripper PRO 9945WX vs Intel Xeon 6724P Comparison

Where Each One Wins

The benchmark data splits these two workstation processors into distinct personality profiles. The AMD Ryzen Threadripper PRO 9945WX claims 15 of the 17 recorded head-to-head comparisons, while the Intel Xeon 6724P takes only 2. That asymmetry alone tells a story, but the nature of those wins matters more than the raw count.

The AMD part dominates almost every category that stresses general compute throughput. In Cinebench, it wins all six recorded tests, both single-core and multi-core, across R15, R20, and R23. The margins are remarkably consistent at roughly 10.7% in every Cinebench test. That uniformity suggests a fundamental per-thread efficiency advantage rather than a workload-specific quirk. The Threadripper also wins PassMark's multithread test by 10.7%, data compression by 7.1%, data encryption by 6.4%, integer math by 7.7%, physics by 22.3%, random string sorting by 23.4%, and the single-thread test by a striking 39.5%.

The Intel Xeon 6724P, by contrast, wins only two recorded tests: PassMark's find prime numbers and floating point math. The prime number win is modest at 9.9%, but the floating point math victory is substantial at 17.6%. That floating point result is the largest margin in favor of either processor in the entire comparison, even larger than AMD's 39.5% single-thread win in percentage terms? No, the 39.5% is larger. The floating point win is the biggest Intel advantage, and it signals that Granite Rapids has genuine muscle in numerically intensive workloads.

The use-case split is therefore clear. The Threadripper PRO 9945WX is the better all-rounder, with wins in rendering, encryption, compression, integer arithmetic, physics simulation, and every single-threaded test. The Xeon 6724P is the specialist for floating point heavy code, particularly workloads that resemble PassMark's floating point math suite, and it has a smaller edge in prime number generation.

The Verdict

The data points to a straightforward recommendation. The AMD Ryzen Threadripper PRO 9945WX should be the default choice for users whose workloads span rendering, general productivity, encryption, compression, and simulation. Its average benchmark score of 76513 places it in the 95th percentile of all CPUs in the database, and it sits within 1% of rivals like the AMD Ryzen 9 9950X3D and within 0.5% of the Intel Core Ultra 9 285HX. The Intel Xeon 6724P, with an average score of 72396, lands in the 94th percentile, just 0.1% ahead of the Intel Xeon 6517P and 0.7% ahead of the Intel Core Ultra 7 265KF.

The Threadripper wins 15 of 17 benchmarks, including every Cinebench test and the PassMark multithread test. Its 10.7% advantage in Cinebench R23 multi-core, 48325 against 43643, is the kind of margin that matters for rendering farms and batch compilation. Its 39.5% single-thread lead, 4573 against 3279 in PassMark, is enormous and will be felt in any application that still relies on single-core responsiveness.

The Xeon 6724P is the pick only for workloads that are dominated by floating point math, where it leads by 17.6%, and prime number finding, where it leads by 9.9%. Those are narrow niches. For a general workstation, the recorded data says the AMD part is the stronger processor. The Xeon does have a launch MSRP of $3622, and it draws 210 W TDP against the Threadripper's 350 W, but the benchmark results do not show a performance advantage that would justify choosing it for mixed workloads.

Head-to-Head Benchmarks

The Cinebench results are the cleanest demonstration of AMD's advantage. In Cinebench R23 multi-core, the Threadripper scores 48325 against the Xeon's 43643, a 10.7% lead. The single-core R23 result is 6822 against 6161, also a 10.7% margin. The pattern repeats in R20, where multi-core is 20296 against 18330 and single-core is 2865 against 2587. In R15, multi-core is 4871 against 4399 and single-core is 687 against 620. Every Cinebench test lands at 10.7% or 10.8% in AMD's favor. This consistency indicates that the Threadripper's Zen 5 architecture delivers a per-clock performance advantage that scales uniformly across both single-threaded and multi-threaded rendering workloads.

PassMark's single-thread test produces the largest margin of the entire comparison. The Threadripper scores 4573, the Xeon scores 3279, and the delta is 39.5%. That is not a small edge; it is a generational gap in single-core throughput. The same two numbers appear for both "passmark_single_thread" and "passmark_singlethread" in the database, confirming the result.

The physics test is another strong AMD win. The Threadripper scores 6118 against 5004, a 22.3% advantage. Random string sorting also favors AMD heavily, 84498 against 68477, a 23.4% lead. These are substantial margins in workloads that stress memory access patterns and branch prediction.

The Intel wins are concentrated in two PassMark tests. Floating point math shows the Xeon ahead at 135404 against 111566, a 17.6% lead. Find prime numbers shows the Xeon at 372 against 335, a 9.9% lead. The floating point result is particularly interesting because it contradicts the general trend. The Xeon's Granite Rapids architecture, with its larger per-core L1 and L2 caches, appears to handle floating point heavy instruction streams better than the Zen 5 design.

The closest contest is PassMark's extended instructions, where the Threadripper wins by only 0.6%, 54406 against 54101. That is effectively a tie. Data encryption is a 6.4% AMD win, 36540 against 34332. Data compression is a 7.1% AMD win, 671963 against 627185. Integer math is a 7.7% AMD win, 185421 against 172216.

FAQ

Q: Which processor is faster in Cinebench R23 multi-core?

A: The AMD Ryzen Threadripper PRO 9945WX scores 48325, which is 10.7% higher than the Intel Xeon 6724P's 43643.

Q: Does the Intel Xeon 6724P win any benchmark?

A: Yes. It wins PassMark's floating point math test by 17.6%, 135404 against 111566, and PassMark's find prime numbers test by 9.9%, 372 against 335.

Q: How large is the single-thread performance gap?

A: The Threadripper PRO 9945WX leads by 39.5% in PassMark's single-thread test, scoring 4573 against the Xeon's 3279.

Q: Which processor has the higher average benchmark score?

A: The Threadripper PRO 9945WX has an average score of 76513, placing it in the 95th percentile, while the Xeon 6724P averages 72396 and sits in the 94th percentile.

Q: Are these processors in the same market segment?

A: Yes, both are classified as Server/Workstation processors. Both support DDR5 memory with an eight-channel bus and 409.6 GB/s bandwidth, and both support ECC memory.

Q: What is the closest benchmark between the two?

A: PassMark's extended instructions test, where the Threadripper wins by only 0.6%, 54406 against 54101.

Architecture Differences

The two processors come from fundamentally different design philosophies. The AMD Ryzen Threadripper PRO 9945WX uses the Zen 5 architecture under the codename Shimada Peak, built on a 4 nm process at TSMC. The Intel Xeon 6724P uses Granite Rapids architecture, also its codename, built on a 5 nm process at Intel. The process node difference, 4 nm versus 5 nm, likely contributes to the Threadripper's per-thread efficiency advantage.

The transistor counts and die sizes are only recorded for the AMD part. The Threadripper packs 16,630 million transistors across two 70.6 mm² dies. No such figures are available for the Xeon in the database.

Cache organization differs significantly. The Threadripper has 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of shared L3. The Xeon has larger per-core caches: 112 KB of L1 per core, 2 MB of L2 per core, and 72 MB of shared L3. Despite having more cache per core and more L3 overall, the Xeon still loses most benchmarks, which suggests that cache size alone does not determine performance in these workloads.

Both are server/workstation parts with no integrated graphics. Both support DDR5 memory with an eight-channel bus and 409.6 GB/s bandwidth, and both support ECC memory. The PCIe lane counts differ: the Threadripper provides Gen 5 with 128 lanes (CPU only), while the Xeon provides Gen 5 with 88 lanes (CPU only). That is a meaningful difference for systems with many GPUs or NVMe drives, and it is not measured in the benchmark scores.

The Threadripper has an unlocked multiplier, while the Xeon is locked. The Threadripper belongs to AMD's 9000 series and the Ryzen Threadripper generation described as Zen 5 (Shimada Peak). The Xeon belongs to the Xeon 6 generation described as Granite Rapids-SP.

Specification Differences

The core and thread counts differ. The Threadripper PRO 9945WX has 12 cores and 24 threads, while the Xeon 6724P has 16 cores and 32 threads. Despite having fewer cores, the AMD part wins the multi-threaded benchmarks, which points to a substantial per-thread performance advantage.

Clock speeds favor AMD. The Threadripper has a base clock of 4.70 GHz and a boost clock of 5.40 GHz. The Xeon has a base clock of 3.60 GHz and a boost clock of 4.30 GHz. The 1.10 GHz gap in boost clock is the most obvious explanation for the single-thread performance difference.

TDP (thermal design power) favors Intel. The Xeon is rated at 210 W, while the Threadripper is rated at 350 W. The AMD part draws more power but delivers more performance in most tests. The Xeon's launch MSRP is $3622; the Threadripper's launch MSRP is not recorded in the database.

Sockets differ completely. The Threadripper uses AMD Socket sTR5, while the Xeon uses Intel Socket 4710. These are not interchangeable platforms, so the motherboard choice will largely determine which processor is even an option.

The release dates differ. The Xeon 6724P was released on 2025-02-23, while the Threadripper PRO 9945WX was released on 2025-06-30. The Xeon is therefore the earlier release by several months. Both processors are marked as Active in production status.

The part numbers are recorded: the Threadripper is 100-000000726, and the Xeon is SRVUA. The Xeon is the only one of the two with a recorded launch MSRP, listed at $3622.

The PCIe lane count is a notable differentiator. The Threadripper provides 128 Gen 5 lanes, the Xeon provides 88 Gen 5 lanes. For multi-GPU workstations or systems with extensive NVMe storage, that 40-lane difference could be decisive even though it does not appear in CPU benchmark scores.