AMD Ryzen AI Embedded P164 vs Intel Core Ultra 9 285H Comparison

FAQ

Q: Which processor has the higher average benchmark score?

A: The AMD Ryzen AI Embedded P164 records an average benchmark score of 52901, while the Intel Core Ultra 9 285H records 38312. The AMD part also sits at the 91st percentile among all CPUs, compared to the Intel part's 86th percentile.

Q: How does Intel's lead in multithreaded performance look in the recorded data?

A: The Intel Core Ultra 9 285H scores 34171 in PassMark multithread, which is 24.2% ahead of the AMD Ryzen AI Embedded P164's 25889. This is one of the larger gaps in the head-to-head results.

Q: Does the AMD processor win any of the shared benchmark tests?

A: Yes, the AMD Ryzen AI Embedded P164 wins PassMark integer math with a score of 87940, while the Intel Core Ultra 9 285H scores 85922. The AMD advantage there is 2.3%.

Q: What is the difference in single-thread performance between the two?

A: The Intel Core Ultra 9 285H scores 4415 in PassMark single thread, which is 8.7% ahead of the AMD Ryzen AI Embedded P164's 4029. Both processors share the same score across the two single-thread tests in the database.

Q: Which processor has the higher boost clock?

A: The Intel Core Ultra 9 285H has a boost clock of 5.40 GHz, while the AMD Ryzen AI Embedded P164 has a boost clock of 5.00 GHz. The Intel part also has a higher base clock at 2.90 GHz versus 2.00 GHz.

Q: How do the two processors compare in data encryption workloads?

A: The Intel Core Ultra 9 285H scores 26140 in PassMark data encryption, which is 38.6% higher than the AMD Ryzen AI Embedded P164's 16055. This is the largest percentage gap in the head-to-head results.

Architecture Differences

The AMD Ryzen AI Embedded P164 uses the Gorgon Point codename and belongs to the Ryzen AI Embedded generation built on Zen 5 and Zen 5c cores. The processor is manufactured on a 4 nm process at TSMC with a die size of 233 mm². The Intel Core Ultra 9 285H, by contrast, uses the Arrow Lake architecture with the Arrow Lake-H codename, and is built on a 3 nm process at TSMC. The die size for the Intel part is not recorded in the database.

Core configuration differs substantially. The AMD part provides 8 cores and 16 threads, while the Intel part provides 16 cores and 16 threads. This means the Intel processor has twice the core count but no hyperthreading advantage, as thread counts are equal. The AMD processor relies on simultaneous multithreading to reach 16 threads from 8 cores.

Cache layouts are distinctly different. The AMD Ryzen AI Embedded P164 has 80 KB of L1 cache per core, 1 MB of L2 cache per core, and 8 MB of L3 cache. The Intel Core Ultra 9 285H has 192 KB of L1 cache per core, 3 MB of L2 cache per core, and 24 MB of shared L3 cache. The Intel part's larger L3 cache, three times the AMD allocation, likely contributes to its performance in cache-sensitive workloads, though the database does not directly measure cache impact.

Memory support is similar on the surface: both support DDR5 and LPDDR5X memory with dual-channel buses. However, the Intel part records a memory bandwidth of 102.4 GB/s, while the AMD part records 89.6 GB/s. Both support ECC memory.

PCIe connectivity differs. The AMD Ryzen AI Embedded P164 uses Gen 4 with 16 lanes from the CPU, while the Intel Core Ultra 9 285H uses Gen 5 with 8 lanes from the CPU. The AMD part offers more lanes, but the Intel part offers a newer generation with higher per-lane bandwidth.

Integrated graphics differ as well. The AMD processor includes Radeon 880M graphics, while the Intel processor includes Arc Graphics 140T. The database does not provide graphics benchmarks for either part.

The Intel Core Ultra 9 285H carries a launch MSRP of $651, while the AMD Ryzen AI Embedded P164 has no recorded launch MSRP. Both processors are mobile-market parts, both are active in production, and neither has an unlocked multiplier.

Head-to-Head Benchmarks

The head-to-head results show a clear overall pattern: the Intel Core Ultra 9 285H wins 10 of the 11 shared tests, with only one victory for the AMD Ryzen AI Embedded P164.

The largest margin comes in PassMark find prime numbers. The Intel part scores 330 versus the AMD part's 71, a 78.5% advantage. This test often stresses integer iteration and branch prediction, and the Intel architecture's larger caches and higher clocks appear to deliver a decisive edge here.

PassMark floating point math shows the second-largest gap. Intel scores 109190, AMD scores 55799, giving Intel a 48.9% lead. The nearly double score suggests the Intel part handles floating point workloads with significantly more throughput, likely aided by its higher core count.

PassMark physics also favors Intel heavily. The Intel Core Ultra 9 285H scores 2513, while the AMD Ryzen AI Embedded P164 scores 1210, a 51.9% difference. Physics simulations often scale with core count and cache, both of which favor the Intel part.

Data encryption shows a 38.6% Intel advantage, with scores of 26140 versus 16055. This workload often benefits from dedicated instruction extensions and memory bandwidth, and the Intel part's higher bandwidth may play a role.

PassMark multithread gives Intel a 24.2% lead, with 34171 versus 25889. This is the broadest test and reflects the Intel part's core count advantage.

Random string sorting shows a 15% Intel lead, with 40931 versus 34801. Extended instructions give Intel a 9.7% edge, with 26794 versus 24193. Single-thread performance favors Intel by 8.7%, with scores of 4415 versus 4029.

Data compression is the closest Intel win. The Intel part scores 335859, the AMD part scores 327891, a margin of just 2.4%. This near parity suggests the AMD architecture is competitive in compression workloads despite its lower core count.

The lone AMD win is PassMark integer math. The AMD Ryzen AI Embedded P164 scores 87940, while the Intel Core Ultra 9 285H scores 85922, giving AMD a 2.3% edge. This is a narrow victory, but it indicates that the AMD part's integer execution pipeline is at least comparable to Intel's in this specific test.

The overall score pattern is consistent with the core and cache differences. The Intel part's 16 cores and 24 MB of L3 cache provide substantial advantages in parallel and cache-heavy workloads. The AMD part's 8 cores with simultaneous multithreading hold up better in integer math, where per-thread efficiency matters more.

Specification Differences

The two processors differ across several core specification fields. Core count: AMD has 8, Intel has 16. Thread count: both have 16, so the AMD part uses SMT while the Intel part does not. Base clock: AMD is 2.00 GHz, Intel is 2.90 GHz. Boost clock: AMD is 5.00 GHz, Intel is 5.40 GHz.

TDP differs, with AMD at 28 and Intel at 45. The Intel part consumes more power by specification, which correlates with its higher clocks and core count.

Socket types are entirely different. AMD uses Socket FP8, while Intel uses BGA 2049. These are not interchangeable platforms.

Process node differs: AMD is on 4 nm, Intel is on 3 nm, both from TSMC. Die size is recorded only for AMD at 233 mm²; Intel's die size is not listed.

Cache specifications diverge in every tier. L1 per core: AMD 80 KB, Intel 192 KB. L2 per core: AMD 1 MB, Intel 3 MB. L3: AMD 8 MB, Intel 24 MB shared.

Memory bandwidth: AMD 89.6 GB/s, Intel 102.4 GB/s. PCIe generation and lane count: AMD Gen 4 with 16 lanes, Intel Gen 5 with 8 lanes.

Integrated graphics differ: AMD Radeon 880M versus Intel Arc Graphics 140T. Release dates differ, with Intel releasing earlier and AMD later. The Intel part has a recorded launch MSRP of $651, while the AMD part does not have one.

Part numbers: AMD is listed as unknown, Intel is listed as SRQAL.

Where Each One Wins

The Intel Core Ultra 9 285H wins in almost every measured category. It leads in data compression by 2.4%, data encryption by 38.6%, extended instructions by 9.7%, find prime numbers by 78.5%, floating point math by 48.9%, multithread by 24.2%, physics by 51.9%, random string sorting by 15%, and single-thread performance by 8.7%. These wins span integer-heavy, floating-point-heavy, memory-heavy, and parallel workloads.

The Intel part's 16 cores without SMT provide raw parallel throughput that the AMD part's 8 cores with SMT cannot match in most tests. The larger L3 cache and higher memory bandwidth also support workloads that repeatedly access large data sets. The higher base and boost clocks give it an edge in latency-sensitive single-thread tasks.

The AMD Ryzen AI Embedded P164 wins only in integer math, with a 2.3% margin. This suggests the Zen 5 architecture's integer execution units are efficient, and the SMT implementation allows the 8 cores to keep up with Intel's 16 physical cores in this specific workload. The AMD part also has a lower TDP of 28 versus 45, which may make it more suitable for thermally constrained mobile designs, though the database does not directly measure power efficiency.

The AMD part's PCIe Gen 4 with 16 lanes offers more lane count than Intel's Gen 5 with 8 lanes. For systems that need many peripheral connections at Gen 4 speeds, the AMD part provides more flexibility. The Intel part's Gen 5 lanes offer higher per-lane bandwidth, which suits fewer, faster devices.

For workloads that depend on data encryption, physics simulation, floating point math, or prime number computation, the Intel Core Ultra 9 285H is the clear choice based on the recorded data. For integer math tasks and for systems where the lower TDP and higher PCIe lane count matter more, the AMD Ryzen AI Embedded P164 has specific advantages.

The Verdict

The recorded data points to a decisive overall winner in raw performance: the Intel Core Ultra 9 285H. It wins 10 of 11 head-to-head tests, with margins ranging from 2.4% in data compression to 78.5% in find prime numbers. Its average benchmark score, however, is lower than the AMD part's, at 38312 versus 52901, which reflects the different benchmark suites each part has been tested with. The Intel part appears in Cinebench and Geekbench tests that the AMD part does not have recorded, and those scores contribute to a lower average.

The Intel Core Ultra 9 285H delivers the stronger single-thread performance at 4415 versus 4029, and the stronger multithread performance at 34171 versus 25889. It also has higher clocks, more cores, more cache, and higher memory bandwidth. For any workload that scales with these resources, the Intel part is the better choice.

The AMD Ryzen AI Embedded P164 holds its own in integer math, winning that test by 2.3%, and it posts a competitive data compression score within 2.4% of Intel. Its 91st percentile ranking versus Intel's 86th indicates that, across the full database of recorded CPUs, the AMD part sits higher in the distribution. This is notable given its lower TDP of 28 and its 8-core configuration.

The choice between the two depends on the workload profile. The Intel Core Ultra 9 285H suits compute-heavy tasks that use many cores and large caches, including physics, encryption, floating point, and multithreaded applications. The AMD Ryzen AI Embedded P164 suits integer-centric workloads and designs that prioritize lower power draw and more PCIe lanes. The data does not support a single universal winner, but for the majority of the measured tests, the Intel part delivers higher scores.