AMD Ryzen 9 9900X vs AMD Ryzen AI Embedded P164 Comparison

The AMD Ryzen 9 9900X and the AMD Ryzen AI Embedded P164 target entirely different corners of the processor market, and the benchmark data reflects that split. The desktop Ryzen 9 9900X is a high-core-count powerhouse for demanding desktop workloads, while the Ryzen AI Embedded P164 is a mobile-class part designed for efficient, integrated systems. The recorded head-to-head results show a clean sweep: the Ryzen 9 9900X wins all 11 shared benchmark comparisons, yet the P164 remains a strong performer in its own class, holding a 91st percentile ranking among all CPUs.

Where Each One Wins

The Ryzen 9 9900X wins every single shared benchmark, but the magnitude of those wins varies dramatically by workload type. The largest gap appears in prime number finding, where the 9900X scores 436 versus the P164's 71, a 514.1% advantage. This indicates a massive lead in integer-heavy, single-thread-dependent mathematical workloads. Physics simulation also shows a large separation, with the 9900X delivering 3381 points against 1210 points, a 179.4% lead. These are workloads that scale with raw core count and clock speed, both of which favor the desktop chip.

The smallest advantage for the 9900X comes in single-threaded performance. The 9900X scores 4672 in the PassMark single-thread test, just 16% ahead of the P164's 4029. This suggests that while the desktop part is faster per core, the embedded chip's Zen 5 architecture keeps it competitive in lightly threaded tasks. The two chips also show closer relative performance in integer math and random string sorting, where the 9900X leads by 105.9% and 106.9% respectively. These are still decisive wins, but they are closer to parity than the extreme gaps seen in prime number finding.

For the P164, its wins are not in performance but in efficiency and integration. It operates at a 28 W TDP, a fraction of the 120 W TDP of the 9900X. The embedded part also integrates the Radeon 880M graphics, which is a more capable integrated GPU than the Radeon Graphics found in the desktop chip. The P164 supports both DDR5 and LPDDR5X memory, offering flexibility for compact system designs. Its 91st percentile ranking among all CPUs shows that despite the performance deficit to the 9900X, it is still a highly capable processor relative to the broader market.

Architecture Differences

Both processors are built on TSMC's 4 nm process node and use AMD's Zen 5 architecture, but they diverge significantly in configuration. The Ryzen 9 9900X, codenamed Granite Ridge, is a desktop part with 12 cores and 24 threads. It uses a dual-chiplet design with a die size of 2x 70.6 mm² and a total of 16,630 million transistors. The P164, codenamed Gorgon Point, is a mobile embedded part with 8 cores and 16 threads, built on a monolithic 233 mm² die. Its generation listing notes a combination of Zen 5 and Zen 5c cores, indicating a hybrid core design that the 9900X does not use.

Cache configurations differ substantially. The 9900X carries 64 MB of L3 cache, while the P164 has only 8 MB. Both parts share the same per-core L1 cache of 80 KB and per-core L2 cache of 1 MB. The larger L3 cache on the desktop part is a major factor in its performance advantage, particularly in workloads with large working sets. The 9900X also has a significantly higher clock envelope, with a 4.40 GHz base clock and 5.60 GHz boost clock, compared to the P164's 2.00 GHz base and 5.00 GHz boost.

Memory and I/O capabilities also separate the two. Both support dual-channel memory with 89.6 GB/s of bandwidth, and both support ECC memory. However, the 9900X uses DDR5 exclusively and provides PCIe Gen 5 with 24 CPU lanes, while the P164 supports DDR5 and LPDDR5X and provides PCIe Gen 4 with 16 lanes. The desktop part is socketed on AMD Socket AM5 and has an unlocked multiplier for overclocking, while the P164 is soldered on AMD Socket FP8 with a locked multiplier. The 9900X was released in August 2024 with a launch MSRP of $499, while the P164 is scheduled for a March 2026 release with no listed launch price.

Head-to-Head Benchmarks

The head-to-head data covers 11 PassMark tests, and the Ryzen 9 9900X wins all of them. The most extreme result is in prime number finding, where the 9900X's 436 score dwarfs the P164's 71, a 514.1% delta. This is the kind of workload that stresses integer throughput and per-core efficiency, and the desktop chip's higher clocks and larger cache deliver an overwhelming result.

Extended instructions show the next largest gap at 128.3%, with the 9900X scoring 55243 against 24193. This test exercises SIMD and specialized instruction paths, where the 9900X's higher boost clock and larger L3 cache provide a clear edge. Floating-point math follows closely at 115.2%, with scores of 120083 and 55799 respectively. The multithread test shows a 111.1% lead, with the 9900X at 54643 and the P164 at 25889, reflecting the desktop chip's 12-core, 24-thread advantage over the embedded part's 8 cores and 16 threads.

Data compression and encryption both show roughly 108% leads. The 9900X scores 683579 in compression versus 327891, and 33421 in encryption versus 16055. Random string sorting shows a 106.9% lead, with scores of 72013 and 34801. Integer math is the tightest of the large-gap tests at 105.9%, with the 9900X at 181056 and the P164 at 87940. Physics shows a 179.4% lead, with scores of 3381 and 1210.

The single-thread test is the closest overall. The 9900X scores 4672 versus 4029, a 16% lead. This confirms that while the P164's Zen 5 cores are efficient, the 9900X's much higher clock speeds and larger cache give it a meaningful per-thread advantage. The single-thread result is the only benchmark where the two chips are in the same performance neighborhood.

FAQ

Q: Which processor has more cores and threads?

A: The AMD Ryzen 9 9900X has 12 cores and 24 threads, while the AMD Ryzen AI Embedded P164 has 8 cores and 16 threads.

Q: How do the cache sizes compare between the two chips?

A: Both have 80 KB of L1 cache and 1 MB of L2 cache per core. The Ryzen 9 9900X has 64 MB of L3 cache, while the Ryzen AI Embedded P164 has 8 MB of L3 cache.

Q: What is the difference in power consumption?

A: The Ryzen 9 9900X has a TDP of 120 W, while the Ryzen AI Embedded P164 operates at a 28 W TDP.

Q: Which processor supports faster PCIe connectivity?

A: The Ryzen 9 9900X supports PCIe Gen 5 with 24 CPU lanes. The Ryzen AI Embedded P164 supports PCIe Gen 4 with 16 CPU lanes.

Q: How do the two processors rank among all CPUs?

A: The Ryzen 9 9900X ranks in the 92nd percentile among all CPUs, while the Ryzen AI Embedded P164 ranks in the 91st percentile.

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

A: The Ryzen 9 9900X scores 4672 in the PassMark single-thread test, which is 16% higher than the 4029 score of the Ryzen AI Embedded P164.

The Verdict

The benchmark data is unambiguous: the AMD Ryzen 9 9900X is the far faster processor in every measured workload. It wins all 11 head-to-head comparisons, with leads ranging from 16% in single-threaded tests to 514.1% in prime number finding. Its 12 cores, 24 threads, 64 MB of L3 cache, and 5.60 GHz boost clock make it the clear choice for desktop users who need maximum compute throughput, high-bandwidth PCIe Gen 5 connectivity, and the ability to overclock via its unlocked multiplier. Its 92nd percentile ranking and nearest-rival comparisons to server-class parts like the AMD EPYC 9015 and Intel Xeon Platinum 8260M place it in serious high-performance territory.

The AMD Ryzen AI Embedded P164 serves a different purpose. Its 28 W TDP, 8 cores, 16 threads, and integrated Radeon 880M graphics are tailored for compact, power-sensitive embedded and mobile systems. It trails the 9900X by large margins in multithreaded and math-heavy workloads, but its 16% single-thread deficit shows that Zen 5 efficiency carries over to the embedded line. Its 91st percentile ranking and nearest-rival comparisons to desktop parts like the AMD Ryzen 9 7900X indicate that it holds its own against much larger chips in average performance.

The choice comes down to platform. The 9900X is for desktop builds on Socket AM5 where power and space are not constraints. The P164 is for embedded designs on Socket FP8 where low power, integrated graphics, and memory flexibility matter more than raw compute. The data confirms that neither chip is a compromise within its intended segment; the 9900X dominates heavy desktop workloads, and the P164 delivers strong performance per watt for embedded applications.