AMD Ryzen Threadripper PRO 3955WX vs Intel Xeon Phi 7290 Comparison

FAQ

Q: Which processor has more cores and threads?

A: The Intel Xeon Phi 7290 has 72 cores and 288 threads, while the AMD Ryzen Threadripper PRO 3955WX has 16 cores and 32 threads. The Xeon Phi offers a massive thread count advantage, but the Ryzen Threadripper PRO wins in most recorded benchmarks.

Q: What is the average benchmark score difference between the two?

A: The AMD Ryzen Threadripper PRO 3955WX has an average benchmark score of 56,555, while the Intel Xeon Phi 7290 averages 53,469. The Ryzen Threadripper PRO sits at the 92nd percentile of all CPUs, and the Xeon Phi sits at the 91st percentile.

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

A: The Ryzen Threadripper PRO scores 34,131 in Cinebench R23 multi-core, while the Xeon Phi scores 15,163. The AMD part leads by 125.1% in this test.

Q: Which processor wins in single-threaded workloads?

A: The Ryzen Threadripper PRO wins decisively in single-threaded tests. In PassMark single-thread, it scores 2,679 versus 485 for the Xeon Phi, a delta of 452.4%. In Cinebench R23 single-core, the AMD chip scores 4,818 versus 2,140, a 125.1% advantage.

Q: Are there any benchmarks where the Xeon Phi wins?

A: Yes, two tests. The Xeon Phi wins PassMark extended instructions with 41,517 versus 38,952 (a 6.2% edge) and PassMark random string sorting with 68,593 versus 63,758 (a 7% edge).

Q: What are the process nodes and foundries for each chip?

A: The Ryzen Threadripper PRO is built on a 7 nm process at TSMC, while the Xeon Phi uses a 14 nm process at Intel. The Ryzen chip has 7,600 million transistors; the Xeon Phi has 8,000 million.

Where Each One Wins

The benchmark data splits cleanly into two categories. The AMD Ryzen Threadripper PRO 3955WX dominates the vast majority of workloads, winning 15 of the 17 recorded head-to-head tests. Its wins span every Cinebench generation (R15, R20, R23) in both multi-core and single-core modes, with deltas consistently around 125%. It also takes PassMark multithread (40,155 versus 17,839, a 125.1% lead), PassMark floating point math (77,979 versus 47,417, a 64.5% lead), PassMark find prime numbers (198 versus 114, a 73.7% lead), and PassMark data encryption (38,058 versus 12,505, a 204.3% lead). Even in integer math, where the gap narrows, the AMD chip wins with 132,456 versus 126,922, a 4.4% edge.

The Intel Xeon Phi 7290 wins exactly two tests. It beats the AMD part in PassMark extended instructions (41,517 versus 38,952, a 6.2% advantage) and PassMark random string sorting (68,593 versus 63,758, a 7% advantage). These are narrow wins, suggesting the Xeon Phi's many cores help in particular parallel patterns that involve string manipulation or specialized instruction throughput.

For use-case planning, the data indicates the Ryzen Threadripper PRO is the better choice for general multi-threaded workloads, single-threaded responsiveness, encryption, prime number calculations, and physics simulation. The Xeon Phi's wins are niche: workloads that heavily exercise extended instruction sets or random string sorting algorithms. If a user's software is specifically tuned for Knights Landing's instruction capabilities, the Xeon Phi holds a small edge; otherwise, the AMD part is the stronger all-rounder.

Architecture Differences

The two processors represent fundamentally different design philosophies. The AMD Ryzen Threadripper PRO 3955WX uses the Zen 2 architecture, codenamed Castle Peak, built on a 7 nm process at TSMC. It belongs to the 3000 series and the Ryzen Threadripper generation. The Intel Xeon Phi 7290 uses the Knights Landing architecture, built on a 14 nm process at Intel, and belongs to the Xeon Phi generation.

The Ryzen chip has 16 cores with 32 threads, while the Xeon Phi packs 72 cores with 288 threads. The Xeon Phi's core count is more than four times higher, but each core is far less powerful. The Ryzen chip has a base clock of 3.90 GHz and a boost clock of 4.30 GHz. The Xeon Phi runs at a base clock of 1500 MHz (1.50 GHz) and a boost clock of 1700 MHz (1.70 GHz). This clock difference explains much of the single-threaded performance gap.

Cache hierarchies also differ. The Ryzen Threadripper PRO has 64 KB of L1 cache per core, 512 KB of L2 per core, and a 64 MB L3 cache. The Xeon Phi has 32 KB of L1 per core and 512 KB of L2 per core, with no L3 cache listed. The Ryzen chip's large L3 cache is a significant advantage for workloads that benefit from shared data.

Memory support is another differentiator. Both support DDR4 and ECC memory, but the Ryzen chip has an eight-channel memory bus with 204.8 GB/s bandwidth. The Xeon Phi's memory bus and bandwidth are not recorded in the database. The Ryzen chip also supports PCIe Gen 4 with 128 lanes (CPU only), while the Xeon Phi has no PCIe data recorded.

Specification Differences

The two processors differ on nearly every specification field. The Ryzen Threadripper PRO has 16 cores and 32 threads; the Xeon Phi has 72 cores and 288 threads. The Ryzen chip runs at 3.90 GHz base and 4.30 GHz boost; the Xeon Phi runs at 1500 MHz base and 1700 MHz boost. The Ryzen chip has a TDP of 280 watts; the Xeon Phi has a TDP of 245 watts.

The socket types are different: the AMD part uses Socket WRX8, while the Intel part uses Socket 3647. The process node differs: 7 nm for AMD (TSMC) versus 14 nm for Intel. The Ryzen chip has 7,600 million transistors on a die size of 2x 74 mm²; the Xeon Phi has 8,000 million transistors with no die size recorded.

Cache configurations: the Ryzen chip has 64 KB L1 per core, 512 KB L2 per core, and 64 MB L3. The Xeon Phi has 32 KB L1 per core and 512 KB L2 per core, with no L3. Memory support: both use DDR4 and support ECC, but the Ryzen chip has an eight-channel bus with 204.8 GB/s bandwidth, while the Xeon Phi's memory bus and bandwidth are not listed. PCIe: the Ryzen chip supports Gen 4 with 128 lanes (CPU only); the Xeon Phi has no PCIe data recorded.

The Ryzen part has a launch MSRP of $1149, while the Xeon Phi has no launch MSRP recorded. The Ryzen chip was released on 2020-07-13; the Xeon Phi was released on 2016-06-19. The Ryzen chip is marked as Active in production; the Xeon Phi's production status is not recorded. The Ryzen chip is a Desktop market segment product; the Xeon Phi targets Server/Workstation. Both have locked multipliers.

Head-to-Head Benchmarks

The biggest win for the AMD Ryzen Threadripper PRO comes in single-threaded performance. In PassMark single-thread, the AMD chip scores 2,679 versus 485 for the Xeon Phi, a delta of 452.4%. This is the largest relative gap in the entire dataset. The same 452.4% delta appears in the duplicate PassMark singlethread entry. This reflects the massive clock speed and per-core efficiency advantage of the Zen 2 architecture over Knights Landing.

In multi-threaded Cinebench tests, the AMD chip is consistently about 125% ahead. Cinebench R15 multi-core: 3,440 versus 1,528 (125.1% delta). Cinebench R20 multi-core: 14,335 versus 6,368 (125.1% delta). Cinebench R23 multi-core: 34,131 versus 15,163 (125.1% delta). Notably, these deltas are identical at 125.1%, showing a consistent throughput advantage regardless of the Cinebench generation.

Single-core Cinebench results show similar patterns. Cinebench R15 single-core: 485 versus 215 (125.6% delta). Cinebench R20 single-core: 2,023 versus 898 (125.3% delta). Cinebench R23 single-core: 4,818 versus 2,140 (125.1% delta). The AMD chip wins by roughly 125% across all three.

PassMark data encryption shows the second-largest relative win: 38,058 versus 12,505, a 204.3% delta. The AMD chip also wins PassMark floating point math by 64.5% (77,979 versus 47,417) and PassMark find prime numbers by 73.7% (198 versus 114). PassMark multithread gives the AMD chip a 125.1% win (40,155 versus 17,839). PassMark physics is a closer contest: 2,460 versus 2,257, a 9% delta. PassMark data compression is also close: 601,738 versus 564,535, a 6.6% delta. PassMark integer math: 132,456 versus 126,922, a 4.4% delta.

The Xeon Phi's two wins are modest in comparison. PassMark extended instructions: 41,517 versus 38,952, a 6.2% advantage. PassMark random string sorting: 68,593 versus 63,758, a 7% advantage. Both wins are narrow, and neither approaches the magnitude of the AMD chip's victories.

The Verdict

The data points to a clear overall winner: the AMD Ryzen Threadripper PRO 3955WX. It wins 15 of 17 head-to-head benchmarks, holds a higher average benchmark score (56,555 versus 53,469), and sits at a higher percentile (92nd versus 91st). Its nearest rivals in the database, the AMD Ryzen AI 9 HX PRO 470 and AMD Ryzen AI Max 390, have average scores of 56,306 and 56,273 respectively, placing the Threadripper PRO just ahead of them.

For users who prioritize single-threaded performance, the Ryzen Threadripper PRO is the only sensible choice. The 452.4% lead in PassMark single-thread and 125% leads in Cinebench single-core tests show that the Xeon Phi cannot compete in this domain. For heavily multi-threaded workloads, the Ryzen chip still leads by 125% in Cinebench and PassMark multithread, despite having far fewer cores. This suggests that the Xeon Phi's 72 cores do not translate into practical throughput gains for these benchmarks.

The Xeon Phi's appeal is limited to two specific areas: extended instruction workloads and random string sorting. In those tests it leads by 6.2% and 7% respectively. Users running software that leverages Knights Landing's specialized instruction set, or workloads dominated by random string operations, might see a slight advantage. However, these are narrow margins compared to the AMD chip's dominant wins elsewhere.

The Ryzen Threadripper PRO is also a more recent design (2020 versus 2016), built on a smaller process node (7 nm versus 14 nm), with a higher clock speed, larger L3 cache, and support for PCIe Gen 4 with 128 lanes. Its eight-channel memory bus provides 204.8 GB/s of bandwidth, a specification the Xeon Phi does not match in the recorded data.

For a desktop or workstation user seeking a general-purpose high-end processor, the AMD Ryzen Threadripper PRO 3955WX is the clear recommendation based on the benchmark data. For a niche workload that specifically benefits from Knights Landing's extended instructions or string sorting throughput, the Intel Xeon Phi 7290 offers a narrow but real advantage. The Xeon Phi's 91st percentile ranking places it alongside the Intel Core i7-14700F and AMD Ryzen 9 7900X in average score, but its benchmark profile is far less balanced than the Ryzen Threadripper PRO's.