AMD Ryzen AI Embedded P164 vs AMD Ryzen AI Max 385 Comparison

Head-to-Head Benchmarks

The benchmark data presents a clear hierarchical relationship between these two AMD processors. The AMD Ryzen AI Max 385 wins all 11 recorded head-to-head comparisons, though the margin varies dramatically by workload type. The most decisive victory comes in the prime number search test, where the Max 385 scores 165 against the P164's 71, a 57% advantage. This is the largest percentage gap in the entire dataset and indicates a substantial difference in pure integer recursion and branch-heavy workload performance.

Physics simulation also shows a major separation. The Max 385 records 1889 points versus 1210 for the P164, a 35.9% lead. This test typically stresses thread scheduling and floating-point throughput under sustained load, and the gap here suggests the Max 385's higher power envelope translates directly into sustained compute capability. The extended instruction test, which exercises AVX-style vector workloads, shows the Max 385 at 33873 against 24193, a 28.6% advantage. That result reinforces the pattern: the Max 385 is substantially stronger in both scalar and vector intensive tasks.

Multi-threaded performance, as measured by PassMark's multithread test, favors the Max 385 by 23.2%, with scores of 33705 versus 25889. This is a meaningful gap for any workload that scales across all 16 threads. The data compression test shows a 19.3% lead for the Max 385 (406505 vs 327891), while data encryption favors it by 19.4% (19926 vs 16055). Both tests are sensitive to memory bandwidth and cache hierarchy, which aligns with the architectural differences between the two chips.

Floating-point math delivers a 21.5% advantage for the Max 385 (71105 vs 55799), while integer math shows a 17.8% lead (107046 vs 87940). Random string sorting, a test that stresses memory access patterns and pointer chasing, gives the Max 385 a 20.4% edge (43725 vs 34801). These results collectively indicate that the Max 385's advantage is not confined to one execution domain; it is consistently ahead across compression, encryption, math, and sorting workloads.

The single-thread tests are the outlier. Here the two processors are nearly identical, with the Max 385 scoring 4060 and the P164 scoring 4029, a mere 0.8% difference. Both chips share a 5.00 GHz boost clock, and the data confirms that single-core performance is effectively a tie. This is a critical finding because it isolates the advantage of the Max 385 to multi-threaded and memory-bandwidth-sensitive scenarios rather than raw per-core speed.

Looking at the broader database context, the P164 sits at the 91st percentile among all CPUs with an average benchmark score of 52901. Its nearest rivals include the AMD Ryzen 5 9500F (52873, 0.1% slower), the Intel Xeon 634 (52974, 0.1% faster), and the AMD EPYC 7313P (53206, 0.6% faster). The Max 385, meanwhile, holds the 88th percentile with an average score of 44309. Its nearest rivals are the Intel Core i9-13950HX (44342, 0.1% faster), the Intel Core i5-13600 (44240, 0.2% slower), and the Intel Core Ultra X9 388H (44466, 0.4% faster). Despite the Max 385 winning every head-to-head test against the P164, its overall average score is lower because the PassMark aggregate includes a different weighting across all recorded workloads, and the P164's scores in other unlisted tests likely pull its average higher.

The Verdict

The data dictates a straightforward selection criterion. For any workload that uses multiple threads, the AMD Ryzen AI Max 385 is the faster processor in every measured category. Its advantages range from 17.8% in integer math to 57% in prime number search, with multi-threaded performance ahead by 23.2%. If the task is parallel compute, the Max 385 is the clear choice.

For single-threaded performance, the two are functionally equivalent. The 0.8% gap in single-thread score is within noise territory, and both chips share the same 5.00 GHz boost clock. A workload that is purely single-threaded would see no meaningful difference between the two.

The AMD Ryzen AI Embedded P164 does have one distinguishing characteristic in the data: its 91st percentile ranking versus the Max 385's 88th, and its higher average benchmark score of 52901 versus 44309. This suggests that in the broader PassMark suite, which includes tests not broken out in the head-to-head section, the P164 performs comparatively better relative to the entire CPU population. However, for the specific tests where both chips were measured, the Max 385 dominates.

The power envelope is a decisive differentiator. The P164 has a 28 W TDP, while the Max 385 operates at 55 W. Systems constrained by thermal or power budgets will prefer the P164. Systems that can accommodate the higher power draw will get substantially better multi-threaded throughput from the Max 385. The choice hinges on whether the workload is power-sensitive or performance-sensitive.

Architecture Differences

The two processors come from different AMD design lineages. The P164 uses the Gorgon Point codename and belongs to the Ryzen AI Embedded generation built on Zen 5 and Zen 5c cores. The Max 385 uses the Strix Halo codename and belongs to the Ryzen AI Max generation built on Zen 5 cores exclusively.

Both are manufactured on a 4 nm process at TSMC, but the die configurations differ. The P164 uses a single die measuring 233 mm², while the Max 385 uses two dies, each measuring 70.6 mm², for a combined area of roughly 141.2 mm². This is a notable structural difference: the P164 has a larger monolithic die, while the Max 385 uses a chiplet-style approach with two smaller dies.

The cache hierarchy diverges significantly. Both chips have 80 KB of L1 cache per core and 1 MB of L2 cache per core. The L3 cache, however, differs substantially. The P164 has 8 MB of L3 cache, while the Max 385 has 32 MB of shared L3 cache. This fourfold difference in L3 capacity is a major factor in the multi-threaded performance gap, as larger shared cache reduces latency for data shared across cores.

The integrated graphics also differ. The P164 features a Radeon 880M, while the Max 385 features a Radeon 8050S. The database does not include graphics benchmark scores, so the performance implications are not quantified in the recorded data.

Memory architecture is another point of divergence. The P164 supports both DDR5 and LPDDR5X memory, while the Max 385 supports LPDDR5X only. The memory bus width differs as well: the P164 uses dual-channel memory, while the Max 385 uses quad-channel. This directly relates to the memory bandwidth figures: the P164 delivers 89.6 GB/s, while the Max 385 delivers 256.0 GB/s. The Max 385's bandwidth advantage is roughly 2.86 times that of the P164, which explains its superior performance in bandwidth-sensitive tests like data compression and random string sorting.

Both processors use AMD Socket FP8 and FP11 respectively, with the P164 on FP8 and the Max 385 on FP11. Both support PCIe Gen 4 with 16 lanes (CPU only), and both support ECC memory. Neither has an unlocked multiplier, so overclocking is not a differentiator.

Specification Differences

The core and thread counts are identical: both have 8 cores and 16 threads. The base clock differs significantly. The P164 runs at 2.00 GHz, while the Max 385 runs at 3.60 GHz. The boost clock is identical at 5.00 GHz for both.

The TDP is a substantial differentiator. The P164 is rated at 28 W, while the Max 385 is rated at 55 W. This nearly doubles the thermal design power and explains the performance gap in sustained workloads.

The socket types differ: the P164 uses AMD Socket FP8, while the Max 385 uses AMD Socket FP11. These are not interchangeable, so motherboard compatibility is a hard constraint.

Memory support differs in both type and width. The P164 supports DDR5 and LPDDR5X with a dual-channel bus and 89.6 GB/s bandwidth. The Max 385 supports LPDDR5X only, but with a quad-channel bus and 256.0 GB/s bandwidth. Both support ECC memory.

The integrated graphics differ: Radeon 880M on the P164 versus Radeon 8050S on the Max 385. The release dates differ as well: the Max 385 was released on 2025-01-05, while the P164 was released on 2026-03-08. The part number is recorded as unknown for the P164, while the Max 385 has part number 100-000001424.

FAQ

Q: Which processor has the higher multi-threaded benchmark score?

A: The AMD Ryzen AI Max 385 scores 33705 in the PassMark multithread test, compared to 25889 for the AMD Ryzen AI Embedded P164, a 23.2% advantage.

Q: How do the two compare in single-threaded performance?

A: They are nearly identical. The Max 385 scores 4060 and the P164 scores 4029 in the PassMark single-thread test, a 0.8% difference.

Q: What is the memory bandwidth difference?

A: The Max 385 delivers 256.0 GB/s over a quad-channel LPDDR5X bus, while the P164 delivers 89.6 GB/s over a dual-channel DDR5/LPDDR5X bus.

Q: Which processor has more L3 cache?

A: The Max 385 has 32 MB of shared L3 cache, while the P164 has 8 MB of L3 cache.

Q: What is the TDP difference between the two?

A: The P164 is rated at 28 W, while the Max 385 is rated at 55 W.

Q: Do both processors use the same socket?

A: No. The P164 uses AMD Socket FP8, and the Max 385 uses AMD Socket FP11. They are not compatible.