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

AMD Ryzen AI Embedded P164 and AMD Ryzen AI Max PRO 385 are both 8-core, 16-thread mobile processors built on TSMC's 4 nm node, but the benchmark data shows a decisive performance gap between them. The P164 wins only two recorded tests, both single-thread related, while the Max PRO 385 dominates in all multi-thread and workload-specific metrics. The database places the P164 at the 91st percentile of all CPUs, while the Max PRO 385 sits at the 88th percentile, an interesting inversion given the Max PRO 385's clear lead in most raw scores.

Head-to-Head Benchmarks

The largest single win for the AMD Ryzen AI Max PRO 385 comes in the passmark_find_prime_numbers test, where it scores 157 against the P164's 71. That is a 54.8% advantage, the most pronounced delta in the entire head-to-head set. Prime number finding is a heavily integer-dependent workload, and the Max PRO 385's larger shared L3 cache and higher base clock appear to translate directly into a massive lead here.

In passmark_data_compression, the Max PRO 385 scores 379448 versus 327891 for the P164, a 13.6% edge. Data compression workloads benefit from memory bandwidth and cache capacity, and the Max PRO 385's quad-channel LPDDR5X interface with 256.0 GB/s of bandwidth versus the P164's dual-channel 89.6 GB/s explains much of this gap. The passmark_data_encryption test shows a 15.4% win for the Max PRO 385, with scores of 18978 and 16055 respectively. Encryption is another workload that scales with memory throughput and instruction efficiency.

The passmark_extended_instructions test delivers a 23.1% win for the Max PRO 385, scoring 31442 against 24193. This test exercises SIMD and extended instruction set paths, and the Max PRO 385's Zen 5 architecture, which is listed explicitly as "Zen 5" in the database, gives it an edge over the P164's mixed Zen 5 / Zen 5c configuration. Floating point math shows a 19.8% advantage for the Max PRO 385, with 69580 versus 55799. Integer math follows the same pattern, with the Max PRO 385 scoring 105056 against 87940, a 16.3% win.

The passmark_multithread score, a broad measure of parallel performance, favors the Max PRO 385 by 19.3%, with 32075 versus 25889. The physics test, which is another multi-threaded workload, shows a 29.3% gap, with the Max PRO 385 scoring 1711 against 1210. Random string sorting, a memory-latency-sensitive test, gives the Max PRO 385 a 14.7% win, 40784 versus 34801.

The only two wins for the AMD Ryzen AI Embedded P164 are in the passmark_single_thread and passmark_singlethread tests, which are duplicate entries of the same measurement. The P164 scores 4029, while the Max PRO 385 scores 3995, a slim 0.9% margin. Both processors boost to 5.00 GHz, so the single-thread difference comes down to the P164's lower base clock of 2.00 GHz versus the Max PRO 385's 3.60 GHz, which is not a direct factor in boost-limited single-thread performance. The P164's 0.9% win is real but marginal, and it does not offset the Max PRO 385's broader dominance.

The Verdict

The data indicates that the AMD Ryzen AI Max PRO 385 is the stronger processor for every multi-threaded and memory-intensive workload recorded in the database. It wins 9 of the 11 head-to-head tests, with leads ranging from 13.6% to 54.8%. The Max PRO 385 delivers 256.0 GB/s of memory bandwidth over a quad-channel bus, while the P164 is limited to 89.6 GB/s over dual-channel memory. That bandwidth difference is the single most explanatory factor for the Max PRO 385's wins in data compression, encryption, and random string sorting.

The AMD Ryzen AI Embedded P164 is the better choice only for scenarios where single-thread performance is the sole criterion, and even then the margin is just 0.9%. The P164 also has a lower TDP of 28 watts versus 55 watts for the Max PRO 385, which makes it a more power-conscious option for embedded designs. The P164's 8 MB of L3 cache is a fraction of the Max PRO 385's 32 MB shared L3, and that difference shows up across the benchmark suite.

The percentile rankings complicate the picture slightly. The P164 sits at the 91st percentile of all CPUs, while the Max PRO 385 sits at the 88th percentile. This is driven by the P164's higher average benchmark score of 52901 against the Max PRO 385's 43326, but that average includes the Cinebench results that are only recorded for the Max PRO 385. The head-to-head PassMark tests, which are the only directly comparable measurements, overwhelmingly favor the Max PRO 385.

Architecture Differences

The AMD Ryzen AI Max PRO 385 uses the Strix Halo codename and is built on the Zen 5 architecture across all cores. The AMD Ryzen AI Embedded P164 uses the Gorgon Point codename and is classified as a Zen 5 / Zen 5c hybrid generation, meaning it mixes full Zen 5 cores with denser Zen 5c cores. Both parts are manufactured by TSMC on a 4 nm process, and both use the same L1 cache configuration of 80 KB per core and L2 cache of 1 MB per core.

The L3 cache configuration diverges sharply. The P164 has 8 MB of L3, while the Max PRO 385 has 32 MB of shared L3. This fourfold difference in cache capacity is a primary driver of the performance gap in workloads that repeatedly access the same data. The Max PRO 385's memory subsystem is also fundamentally different, using quad-channel LPDDR5X with 256.0 GB/s of bandwidth, while the P164 uses dual-channel DDR5 and LPDDR5X with 89.6 GB/s.

Both processors support ECC memory, both use PCIe Gen 4 with 16 lanes from the CPU, and both are locked multipliers. The integrated graphics differ, with the P164 using a Radeon 880M and the Max PRO 385 using a Radeon 8050S. The sockets are different as well, with the P164 on AMD Socket FP8 and the Max PRO 385 on AMD Socket FP11. The Max PRO 385 has a part number of 100-000001422, while the P164's part number is listed as unknown.

FAQ

Q: Which processor has the higher single-thread score?

A: The AMD Ryzen AI Embedded P164 scores 4029 in the passmark_single_thread test, while the AMD Ryzen AI Max PRO 385 scores 3995. The P164 wins by 0.9%.

Q: How much memory bandwidth does each processor provide?

A: The AMD Ryzen AI Max PRO 385 provides 256.0 GB/s over a quad-channel LPDDR5X bus. The AMD Ryzen AI Embedded P164 provides 89.6 GB/s over a dual-channel DDR5 and LPDDR5X bus.

Q: What is the L3 cache size difference?

A: The AMD Ryzen AI Max PRO 385 has 32 MB of shared L3 cache. The AMD Ryzen AI Embedded P164 has 8 MB of L3 cache.

Q: Which processor wins the passmark_find_prime_numbers test?

A: The AMD Ryzen AI Max PRO 385 scores 157, while the AMD Ryzen AI Embedded P164 scores 71. The Max PRO 385 wins by 54.8%, the largest delta in the head-to-head set.

Q: Do both processors support ECC memory?

A: Yes, both the AMD Ryzen AI Embedded P164 and the AMD Ryzen AI Max PRO 385 list ECC memory support as true.

Q: What are the TDP values for these two processors?

A: The AMD Ryzen AI Embedded P164 has a TDP of 28 watts. The AMD Ryzen AI Max PRO 385 has a TDP of 55 watts.

Where Each One Wins

The AMD Ryzen AI Max PRO 385 wins in every multi-threaded workload category recorded in the head-to-head data. It is the clear choice for data compression, where its 379448 score beats the P164's 327891 by 13.6%. It wins encryption with 18978 versus 16055, a 15.4% margin. Extended instructions go to the Max PRO 385 by 23.1%, and floating point math by 19.8%. Integer math, multithread, physics, and random string sorting all favor the Max PRO 385, with deltas of 16.3%, 19.3%, 29.3%, and 14.7% respectively. The largest win is in prime number finding, where the Max PRO 385's 157 score is 54.8% higher than the P164's 71.

The AMD Ryzen AI Embedded P164 wins only the single-thread test, and by a 0.9% margin. Its 4029 score edges out the Max PRO 385's 3995. This makes the P164 marginally better for lightly threaded applications where clock speed is the limiting factor, though both processors boost to 5.00 GHz, so the difference is minimal. The P164 also has a lower TDP of 28 watts, which makes it the more power-efficient option for embedded systems where thermal and power budgets are tight. The Max PRO 385's 55 watt TDP is a significant increase, but the benchmark data shows that the extra power buys substantial performance gains across the board.

Specification Differences

The AMD Ryzen AI Embedded P164 and AMD Ryzen AI Max PRO 385 share the same core and thread counts, both offering 8 cores and 16 threads. Both have a boost clock of 5.00 GHz, but the base clocks differ, with the P164 at 2.00 GHz and the Max PRO 385 at 3.60 GHz. The TDP also differs, with the P164 at 28 watts and the Max PRO 385 at 55 watts. The sockets are different, FP8 for the P164 and FP11 for the Max PRO 385.

Cache configurations diverge in L3 capacity, with the P164 holding 8 MB and the Max PRO 385 holding 32 MB shared. Memory support differs, with the P164 supporting DDR5 and LPDDR5X over a dual-channel bus with 89.6 GB/s bandwidth, while the Max PRO 385 supports only LPDDR5X over a quad-channel bus with 256.0 GB/s bandwidth. The integrated graphics are different, Radeon 880M for the P164 and Radeon 8050S for the Max PRO 385. The codenames and generations differ, with the P164 using Gorgon Point and a Zen 5 / Zen 5c generation label, while the Max PRO 385 uses Strix Halo and a pure Zen 5 generation label. The Max PRO 385 has a listed part number of 100-000001422, while the P164's part number is unknown.