AMD Ryzen Threadripper PRO 9945WX vs Intel Xeon 6517P Comparison

The Verdict

The benchmark data splits these two workstation processors into distinct roles. The AMD Ryzen Threadripper PRO 9945WX wins 16 of the 17 recorded head-to-head comparisons, including every Cinebench test and most Passmark workloads. The Intel Xeon 6517P claims only one victory, in floating point math, where it leads by 12.5%. The AMD part also holds a higher average benchmark score, 76513 versus 72350, and sits at the 95th percentile among all CPUs compared to the Intel's 94th percentile.

For users prioritizing single-thread responsiveness, physics simulation, encryption, or general multithreaded throughput, the AMD Ryzen Threadripper PRO 9945WX is the clear choice from the recorded data. Its single-thread Passmark score of 4573 beats the Xeon's 3311 by 38.1%, and its physics score of 6118 beats 4452 by 37.4%. The Intel Xeon 6517P, with its launch MSRP of $1195, becomes relevant only for workloads that heavily depend on floating point math, where its 127497 score exceeds the AMD's 111566.

The data suggests buyers should choose based on workload character. Floating point heavy scientific computing leans toward Intel. Everything else in the database leans toward AMD, often by substantial margins. The AMD part is also unlocked, allowing multiplier adjustments, while the Intel part is locked.

Architecture Differences

The two processors come from different manufacturing and design philosophies. The AMD Ryzen Threadripper PRO 9945WX uses Zen 5 architecture on a 4 nm TSMC process node, with the codename Shimada Peak. It belongs to the Ryzen Threadripper (Zen 5 (Shimada Peak)) generation and the 9000 series. The Intel Xeon 6517P uses Granite Rapids architecture on a 5 nm Intel process, belonging to the Xeon 6 (Granite Rapids-SP) generation.

Core counts differ notably. The Intel Xeon 6517P has 16 cores and 32 threads, while the AMD has 12 cores and 24 threads. Despite having fewer cores, the AMD part wins most multithreaded tests, which points to higher per-core efficiency and clock speeds. The AMD base clock is 4.70 GHz with a boost of 5.40 GHz. The Intel runs at 3.20 GHz base and 4.20 GHz boost. That clock advantage is substantial and appears throughout the single-thread results.

Cache layouts also differ. The AMD part has 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of L3 cache. The Intel part has 112 KB of L1 per core, 2 MB of L2 per core, and 72 MB of shared L3 cache. The Intel has more cache per core in L1 and L2, plus a larger L3 pool, but that does not translate into benchmark wins outside floating point math.

Both support DDR5 memory with an eight-channel bus and 409.6 GB/s memory bandwidth. Both support ECC memory. The AMD offers Gen 5 PCIe with 128 lanes from the CPU, while the Intel offers Gen 5 with 88 lanes. The AMD part uses Socket sTR5, the Intel uses Socket 4710. The AMD is built on 16,630 million transistors across a die size of 2x 70.6 mm², while the Intel's transistor count and die size are not recorded in the database. The AMD has an unlocked multiplier; the Intel does not.

Where Each One Wins

The AMD Ryzen Threadripper PRO 9945WX dominates across rendering, compression, encryption, integer math, sorting, physics, and all single-thread tests. Cinebench R15, R20, and R23 multicore and singlecore results all favor AMD by exactly 14.1%. Passmark data compression shows a narrower AMD win at 2.9%, while data encryption shows a 12.8% advantage. Extended instructions go to AMD by 4.8%. Integer math favors AMD by 14%, multithread by 14.2%, physics by 37.4%, random string sorting by 25.2%, and single-thread by 38.1%. Prime numbers testing is a tie at 335 for both.

The Intel Xeon 6517P wins only floating point math, scoring 127497 against 111566, a 12.5% margin. This is a meaningful signal for users running numerical simulations, scientific computing, or other FPU-heavy tasks. The Intel part also has more cores, 16 versus 12, which may matter in workloads not captured by these specific benchmarks, though the recorded tests do not show a multithreaded advantage from the extra cores.

The data implies a clear split: AMD for general workstation throughput and single-thread latency, Intel for floating point density. The AMD's physics and single-thread wins are the largest margins in the entire comparison, suggesting its architecture extracts far more performance per clock in latency-sensitive tasks.

FAQ

Q: Which processor has the higher average benchmark score?

A: The AMD Ryzen Threadripper PRO 9945WX has an average benchmark score of 76513, while the Intel Xeon 6517P scores 72350.

Q: Does the Intel Xeon 6517P win any benchmark?

A: Yes, it wins Passmark floating point math with a score of 127497, beating the AMD's 111566 by 12.5%.

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

A: The AMD leads by 38.1% in Passmark single-thread (4573 versus 3311) and by 14.1% in every Cinebench singlecore test.

Q: Which processor has more cores?

A: The Intel Xeon 6517P has 16 cores and 32 threads, while the AMD Ryzen Threadripper PRO 9945WX has 12 cores and 24 threads.

Q: Do both processors support ECC memory?

A: Yes, both support ECC memory, and both use DDR5 with an eight-channel bus and 409.6 GB/s bandwidth.

Q: What is the difference in PCIe lanes?

A: The AMD offers Gen 5 with 128 lanes from the CPU, while the Intel offers Gen 5 with 88 lanes.

Head-to-Head Benchmarks

The largest AMD victory appears in Passmark single-thread, where it scores 4573 against the Intel's 3311, a 38.1% margin. This same result appears twice in the database, once as "single_thread" and once as "singlethread," both with identical scores. The physics test shows the second largest gap: AMD at 6118 versus Intel at 4452, a 37.4% difference. Random string sorting goes to AMD by 25.2%, with scores of 84498 and 67480.

The Cinebench suite is uniform. Every generation, R15, R20, and R23, and every mode, singlecore and multicore, shows a 14.1% AMD advantage. R15 multicore is 4871 versus 4268, R15 singlecore is 687 versus 602, R20 multicore is 20296 versus 17787, R20 singlecore is 2865 versus 2511, R23 multicore is 48325 versus 42352, and R23 singlecore is 6822 versus 5979. The consistency of that exact percentage across all six tests suggests the clock speed and IPC advantage scales evenly across rendering workloads.

Passmark multithread gives AMD a 14.2% win, 56854 versus 49786, nearly matching the Cinebench multicore margin. Integer math goes to AMD by 14%, 185421 versus 162671. Data encryption shows a 12.8% AMD lead, 36540 versus 32385. Extended instructions favor AMD by 4.8%, 54406 versus 51891. Data compression is the closest AMD win at 2.9%, 671963 versus 653338. Prime numbers is a dead tie at 335.

The Intel's single win in floating point math is substantial at 12.5%, 127497 versus 111566. This is the only test where the Intel's extra cores and larger cache appear to translate into a performance advantage. The fact that the Intel loses every other multithreaded test despite having 16 cores against 12 suggests the AMD's higher clocks and architecture efficiency overcome the core deficit.

Specification Differences

The two processors differ across nearly every major specification category. The AMD has 12 cores and 24 threads, while the Intel has 16 cores and 32 threads. AMD base clock is 4.70 GHz versus Intel's 3.20 GHz. AMD boost clock is 5.40 GHz versus Intel's 4.20 GHz. The AMD TDP is 350 watts, while the Intel TDP is 190 watts. The AMD uses Socket sTR5, the Intel uses Socket 4710.

The architecture and manufacturing differ completely: AMD uses Zen 5 on TSMC's 4 nm process, Intel uses Granite Rapids on Intel's 5 nm process. The AMD codename is Shimada Peak, the Intel codename is Granite Rapids. The AMD generation is listed as Ryzen Threadripper (Zen 5 (Shimada Peak)), the Intel generation is Xeon 6 (Granite Rapids-SP). The AMD is part of the 9000 series, while no series is recorded for the Intel.

The AMD has 16,630 million transistors and a die size of 2x 70.6 mm², while the Intel's transistor count and die size are not recorded. Cache differs per core: AMD has 64 KB L1 and 1 MB L2 per core, Intel has 112 KB L1 and 2 MB L2 per core. L3 cache is 64 MB on AMD versus 72 MB shared on Intel.

PCIe capability differs: AMD provides Gen 5 with 128 lanes from the CPU, Intel provides Gen 5 with 88 lanes. The AMD multiplier is unlocked, the Intel is locked. The AMD part number is 100-000000726, the Intel part number is SRVU4. The AMD release date is recorded as 2025-06-30, the Intel release date as 2025-02-23. The Intel has a launch MSRP of $1195, while no MSRP is recorded for the AMD. Both support DDR5, eight-channel memory, 409.6 GB/s bandwidth, and ECC memory. Neither has integrated graphics. Both are active production parts in the server and workstation segment.