AMD Ryzen AI Embedded P185 vs Intel Core Ultra 5 338H Comparison

FAQ

Q: Which processor has the higher average benchmark score?

A: The AMD Ryzen AI Embedded P185 records an average benchmark score of 62839, while the Intel Core Ultra 5 338H records 33989. The AMD part sits at the 93rd percentile of all CPUs, whereas the Intel part sits at the 84th percentile.

Q: How do the two processors compare in multithreaded performance?

A: The AMD Ryzen AI Embedded P185 scores 31817 in PassMark multithread, which is 10.8% ahead of the Intel Core Ultra 5 338H's 28717. The AMD part also has 24 threads versus 12 threads on the Intel part.

Q: Which processor wins in single-threaded performance?

A: The Intel Core Ultra 5 338H wins the PassMark single-thread test with a score of 4180, which is 4.9% ahead of the AMD Ryzen AI Embedded P185's 3977.

Q: What are the biggest scoring differences in either direction?

A: The AMD Ryzen AI Embedded P185 wins PassMark integer math by 81.5% (117832 versus 64934). The Intel Core Ultra 5 338H wins PassMark find prime numbers by 57.6% (304 versus 129).

Q: Do the two processors use the same manufacturing process?

A: No. The AMD Ryzen AI Embedded P185 uses a 4 nm process from TSMC, while the Intel Core Ultra 5 338H uses a 3 nm process from Intel. The AMD part is built on the Gorgon Point codename with Zen 5 / Zen 5c cores, and the Intel part uses Panther Lake architecture.

Q: Which processor supports ECC memory?

A: The AMD Ryzen AI Embedded P185 supports ECC memory. The Intel Core Ultra 5 338H does not list ECC memory support.

Where Each One Wins

The recorded PassMark results split five wins for the AMD Ryzen AI Embedded P185 and six wins for the Intel Core Ultra 5 338H. The AMD processor takes data compression, extended instructions, integer math, multithread, and random string sorting. The Intel processor takes data encryption, find prime numbers, floating point math, physics, single-thread, and the duplicate singlethread entry.

The AMD Ryzen AI Embedded P185 is the stronger choice for workloads that scale with thread count and integer throughput. Its 24 threads and 81.5% lead in integer math point to compilation, database operations, and content production tasks that rely on parallel integer execution. The multithread win of 10.8% reinforces this pattern, as does the 35.4% lead in data compression, a test that typically rewards wide parallel resources.

The Intel Core Ultra 5 338H is the stronger choice for workloads that favor floating point, encryption, prime number searches, and physics simulation. Its 16% lead in floating point math and 57.6% lead in find prime numbers suggest a different execution character despite having half the thread count. The 34.3% lead in physics is substantial, and the 8.2% lead in data encryption shows an advantage in cryptographic throughput.

Single-threaded work goes to Intel by 4.9%, which matters for lightly threaded applications and responsiveness. The AMD part still posts a competitive single-thread score of 3977, but the Intel part's 4180 is the higher figure. Users with mixed workloads should note that the AMD part wins the overall average by a wide margin, 62839 versus 33989, but the Intel part has dedicated strengths in several specific tests.

Architecture Differences

The AMD Ryzen AI Embedded P185 uses the Gorgon Point codename and belongs to the Ryzen AI Embedded generation built on Zen 5 / Zen 5c cores. The Intel Core Ultra 5 338H uses the Panther Lake codename and belongs to the Core Ultra Series 3 generation with Panther Lake-H architecture. These are fundamentally different designs from different foundries: AMD uses TSMC's 4 nm process, while Intel uses its own 3 nm process.

The core layouts differ sharply. The AMD part has 12 cores and 24 threads, indicating simultaneous multithreading. The Intel part has 12 cores and 12 threads, indicating no multithreading support. The cache structures also differ. AMD provides 80 KB of L1 per core, 1 MB of L2 per core, and a 16 MB L3 cache. Intel provides 192 KB of L1 per core, 2.5 MB of L2 per core, and an 18 MB shared L3 cache. The Intel part has more cache per core in L1 and L2, plus a larger L3, which may explain its wins in latency-sensitive tests like find prime numbers and physics.

The integrated graphics differ as well. AMD uses the Radeon 890M, while Intel uses the Arc B370. Memory support diverges: AMD supports both DDR5 and LPDDR5X, while Intel supports LPDDR5X only. Both use dual-channel memory buses, but Intel records a higher memory bandwidth of 136.5 GB/s versus AMD's 89.6 GB/s. PCIe capabilities also differ: AMD provides Gen 4 with 16 lanes (CPU only), while Intel provides Gen 5 with 4 lanes (CPU only). The Intel part has a faster PCIe standard but far fewer lanes.

ECC memory support is exclusive to the AMD part, which may matter for embedded and reliability-focused deployments. The AMD part also has a larger die size of 233 mm², while the Intel die size is not recorded in the database.

Specification Differences

The AMD Ryzen AI Embedded P185 and Intel Core Ultra 5 338H differ across nearly every recorded specification. Core counts match at 12, but threads differ at 24 versus 12. Base clocks are close: 2.00 GHz for AMD and 1.90 GHz for Intel. Boost clocks differ more: 5.10 GHz for AMD and 4.70 GHz for Intel. The TDP ratings are 28 W for AMD and 25 W for Intel.

Sockets are incompatible: AMD uses Socket FP8, while Intel uses BGA 2540. The process nodes differ: 4 nm TSMC versus 3 nm Intel. The AMD part has a recorded die size of 233 mm²; no die size is recorded for the Intel part.

Cache configurations differ in every level. AMD has 80 KB L1 per core, 1 MB L2 per core, and 16 MB L3. Intel has 192 KB L1 per core, 2.5 MB L2 per core, and 18 MB shared L3.

Memory support differs: AMD supports DDR5 and LPDDR5X, Intel supports LPDDR5X only. Memory bandwidth differs: AMD at 89.6 GB/s, Intel at 136.5 GB/s. ECC support is present on AMD and absent on Intel. PCIe differs: AMD Gen 4 with 16 lanes, Intel Gen 5 with 4 lanes. Integrated graphics differ: Radeon 890M versus Arc B370.

The release dates differ by less than two months. The Intel part released on 2026-01-04, and the AMD part released on 2026-02-28. Both are active production parts. Neither has a launch MSRP in the database. Both have locked multipliers. The Intel part has a recorded part number of SA4REQ9EW, while the AMD part number is unknown.

Head-to-Head Benchmarks

The largest win for the AMD Ryzen AI Embedded P185 is in PassMark integer math, where it scores 117832 against 64934, a 81.5% lead. This is the single biggest delta in the head-to-head set. The second largest AMD win is in data compression: 374429 versus 276539, a 35.4% lead. Extended instructions go to AMD by 11% (26544 versus 23906), random string sorting goes to AMD by 19% (40557 versus 34082), and multithread goes to AMD by 10.8% (31817 versus 28717).

The largest win for the Intel Core Ultra 5 338H is in PassMark find prime numbers, where it scores 304 against 129, a 57.6% lead. This is the second largest delta in the entire set, behind only AMD's integer math win. Physics goes to Intel by 34.3% (2697 versus 1772). Floating point math goes to Intel by 16% (84067 versus 70587). Data encryption goes to Intel by 8.2% (21367 versus 19612). Single-thread goes to Intel by 4.9% (4180 versus 3977).

The multithread result is notable because the AMD part wins by 10.8% despite having double the thread count. The Intel part's 12 threads produce 28717, while the AMD part's 24 threads produce 31817. The efficiency gap is large in Intel's favor, but the absolute multithread win belongs to AMD.

The single-thread result is close at 4.9%, suggesting both parts have competitive per-core performance. The Intel part's higher boost clock is 4.70 GHz versus 5.10 GHz on the AMD part, yet Intel still takes single-thread. This indicates the Intel architecture extracts more performance per clock in this test, or the boost behavior differs under the recorded conditions.

The physics result favors Intel by a wide margin, 2697 versus 1772. The find prime numbers result favors Intel even more strongly on a percentage basis. These two tests, combined with floating point math, form a cluster of Intel wins that point to strong scalar and floating point execution. The AMD wins cluster around integer, compression, sorting, and multithread, pointing to wide parallel throughput.

The overall average benchmark score is not a simple sum of these wins. The AMD part averages 62839, while the Intel part averages 33989. The Intel part's nearest rivals in the database include the Intel Core Ultra 7 165H at 34083 with a delta of -0.3%, the Intel Core i7-12800HX at 33875 with a delta of 0.3%, the Intel Xeon 6353P at 33844 with a delta of 0.4%, and the AMD EPYC 4244P at 34220 with a delta of -0.7%. The AMD part's nearest rivals include the Intel Core Ultra 7 255HX at 62738 with a delta of 0.2%, the Intel Core i7-13790F at 63080 with a delta of -0.4%, the Intel Core Ultra 7 265HX at 63173 with a delta of -0.5%, and the AMD Ryzen AI 9 PRO 465 at 62498 with a delta of 0.5%.

The Verdict

The data presents two distinct processors with different strengths. The AMD Ryzen AI Embedded P185 delivers the higher average benchmark score by a wide margin, 62839 versus 33989, and sits at the 93rd percentile of all CPUs compared to the Intel part's 84th percentile. It wins multithread, integer math, data compression, extended instructions, and random string sorting. The 24 threads and 5.10 GHz boost clock support a throughput-oriented profile. ECC memory support and 16 PCIe Gen 4 lanes also distinguish it for embedded applications.

The Intel Core Ultra 5 338H wins six of the eleven head-to-head tests, including the largest single win in the set: find prime numbers by 57.6%. It also wins physics by 34.3%, floating point math by 16%, data encryption by 8.2%, and single-thread by 4.9%. The 136.5 GB/s memory bandwidth, larger caches, and 3 nm process from Intel contribute to a profile that favors scalar, floating point, and latency-sensitive workloads. The lack of multithreading limits its parallel ceiling, but its per-thread efficiency is strong.

The choice depends on workload. For parallel integer-heavy tasks, compression, sorting, and multithreaded throughput, the AMD Ryzen AI Embedded P185 is the stronger part. For floating point math, physics simulation, prime number searches, encryption, and single-threaded responsiveness, the Intel Core Ultra 5 338H is the stronger part. The AMD part also carries the overall average score advantage and ECC support. The Intel part carries the memory bandwidth advantage and the larger L3 cache. Both are active mobile parts released within weeks of each other, but they target different execution profiles.