AMD Ryzen AI Embedded P164 vs Intel Xeon Gold 5320H Comparison

The AMD Ryzen AI Embedded P164 and Intel Xeon Gold 5320H represent two fundamentally different approaches to computing, and the benchmark data reflects that starkly. The AMD part wins decisively in single-threaded performance and encryption, while the Intel Xeon dominates in nearly every multi-threaded and throughput-oriented workload. With the AMD claiming 3 head-to-head wins and the Intel taking 8, the choice between them is less about overall capability and more about matching the silicon to the specific task at hand.

Head-to-Head Benchmarks

The most dramatic separation occurs in the PassMark single-thread test, where the AMD Ryzen AI Embedded P164 scores 4029 against the Intel Xeon Gold 5320H’s 2428. That is a 65.9% advantage for the AMD part, the largest delta of any benchmark in the comparison. This result aligns with the AMD’s 5.00 GHz boost clock versus the Intel’s 4.20 GHz, though the architectural efficiency of the Zen 5 design is clearly a factor as well. For any workload that relies on per-core responsiveness, the data is unambiguous: the AMD is the faster processor.

The AMD also wins the data encryption test, scoring 16055 versus the Intel’s 11462, a 40.1% margin. This is a notable result because encryption workloads often scale with core count, yet the AMD’s 8 cores and 16 threads outperform the Intel’s 20 cores and 40 threads in this specific task. The implication is that the AMD’s modern instruction set and memory subsystem are more effective at cryptographic operations, making it the better choice for security-sensitive embedded applications.

However, the Intel Xeon Gold 5320H is the clear victor in raw throughput. The largest margin is in the find prime numbers test, where the Intel scores 156 against the AMD’s 71, a 54.5% advantage. This is a purely integer-heavy, parallel workload, and the Intel’s 20 physical cores provide a massive parallel advantage. The random string sorting test tells a similar story, with the Intel at 62730 versus the AMD’s 34801, a 44.5% gap. These results underscore the Intel’s strength in data manipulation and sorting tasks that can utilize its many threads.

The multi-thread test, which is the closest proxy for general-purpose parallel performance, shows the Intel ahead at 31718 versus 25889, an 18.4% margin. While this is a clear win, it is smaller than the core-count disparity might suggest, indicating that the AMD’s per-core efficiency helps it close some of the gap. The physics test shows a larger 49.5% advantage for the Intel (2395 vs 1210), suggesting that the Xeon’s higher base clock of 2.40 GHz is beneficial in sustained, compute-heavy physics simulations. The Intel also leads in integer math (119955 vs 87940, 26.7% ahead), floating-point math (73873 vs 55799, 24.5% ahead), data compression (500828 vs 327891, 34.5% ahead), and extended instructions (36564 vs 24193, 33.8% ahead). In every one of these parallel workloads, the Intel’s additional cores translate into higher scores, though the AMD’s relative competitiveness in multi-thread and integer math shows its architecture is not being left far behind.

Architecture Differences

The architectural gap between these two processors is generational. The AMD Ryzen AI Embedded P164 is built on a 4 nm process at TSMC and uses the Gorgon Point codename, part of the Ryzen AI Embedded line based on Zen 5 / Zen 5c cores. The Intel Xeon Gold 5320H, in contrast, is a Cooper Lake-SP part built on Intel’s 14 nm process. This process difference is a primary driver of the efficiency and clock speed disparities seen in the benchmarks. The AMD’s smaller transistor geometry allows for higher boost clocks at a much lower thermal envelope, while the Intel’s older, larger node requires more power to reach lower clock speeds.

Core configurations also diverge sharply. The AMD has 8 cores and 16 threads, while the Intel has 20 cores and 40 threads. The AMD compensates with a much higher 5.00 GHz boost clock versus 4.20 GHz on the Intel, but the base clocks tell a different story: the Intel runs at 2.40 GHz base, while the AMD sits at 2.00 GHz. The TDP figures are extreme, with the AMD rated at 28 watts and the Intel at 150 watts, making the AMD a low-power embedded part and the Intel a high-power server chip. Cache hierarchies also differ: the AMD has 80 KB of L1 and 1 MB of L2 per core, with 8 MB of shared L3, whereas the Intel has 64 KB of L1 and 1 MB of L2 per core, but a much larger 27.5 MB of shared L3. This larger L3 on the Intel likely helps with its data compression and sorting scores, as more data can reside closer to the cores.

Memory support is another major differentiator. The AMD supports DDR5 and LPDDR5X over a dual-channel bus, offering 89.6 GB/s of bandwidth. The Intel supports only DDR4 over a six-channel bus, providing 128.0 GB/s of bandwidth. The Intel’s higher memory bandwidth is a likely contributor to its wins in data compression and random string sorting, which are bandwidth-sensitive. Both parts support ECC memory, but the AMD’s dual-channel configuration is a limiting factor for its throughput potential. PCIe connectivity also differs: the AMD offers Gen 4 with 16 CPU lanes, while the Intel provides Gen 3 with 48 CPU lanes, giving the Intel more expansion headroom for server-style workloads. The AMD includes an integrated Radeon 880M GPU, while the Intel has no integrated graphics, making the AMD a more self-contained embedded solution. The Intel’s die size is not listed, but its process node and core count suggest a much larger physical package.

The Verdict

The data paints a clear picture for distinct use cases. The AMD Ryzen AI Embedded P164 is the superior processor for single-threaded performance and encryption tasks, and its low 28-watt TDP makes it a compelling option for power-constrained embedded environments. Its 65.9% lead in single-thread performance over the Intel is a decisive factor for any application that prioritizes per-core speed, such as real-time control or interactive workloads. The 40.1% win in data encryption also indicates that the AMD is better suited for secure communications and data protection tasks, despite having fewer cores.

The Intel Xeon Gold 5320H is the clear choice for multi-threaded, throughput-heavy server workloads. Its wins in data compression (34.5% ahead), extended instructions (33.8% ahead), and prime number finding (54.5% ahead) show that it excels in parallel compute tasks that can utilize its 20 cores and 40 threads. The 18.4% lead in the multi-thread test and the 44.5% lead in random string sorting further reinforce its position as a server/workstation processor designed for heavy data processing. The higher memory bandwidth of 128.0 GB/s and the larger 27.5 MB L3 cache directly support these workloads.

Neither processor is a universal winner. The AMD’s 8 cores limit its parallel ceiling, while the Intel’s older architecture and lower boost clock cripple its single-thread performance. The AMD’s 91st percentile ranking among all CPUs is matched by the Intel, but their average benchmark scores are nearly identical: 52901 for the AMD and 52431 for the Intel. The choice comes down to the workload. If the task is single-thread-bound or encryption-heavy, the AMD Ryzen AI Embedded P164 is the better part. If the task is multi-threaded and data-intensive, the Intel Xeon Gold 5320H is the stronger performer.

FAQ

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

A: The AMD Ryzen AI Embedded P164 scores 4029 in the PassMark single-thread test, which is 65.9% higher than the Intel Xeon Gold 5320H’s score of 2428.

Q: How much faster is the Intel Xeon in multi-threaded workloads?

A: The Intel Xeon Gold 5320H scores 31718 in the PassMark multi-thread test, compared to the AMD’s 25889, representing an 18.4% advantage.

Q: What is the core and thread count difference?

A: The AMD Ryzen AI Embedded P164 has 8 cores and 16 threads, while the Intel Xeon Gold 5320H has 20 cores and 40 threads.

Q: Which processor has a higher boost clock?

A: The AMD Ryzen AI Embedded P164 has a 5.00 GHz boost clock, which is higher than the Intel Xeon Gold 5320H’s 4.20 GHz boost clock.

Q: Which processor is more efficient in data encryption?

A: The AMD Ryzen AI Embedded P164 is faster in data encryption, scoring 16055 compared to the Intel’s 11462, a 40.1% margin.

Q: Which processor supports more memory bandwidth?

A: The Intel Xeon Gold 5320H supports 128.0 GB/s of memory bandwidth via six-channel DDR4, while the AMD supports 89.6 GB/s via dual-channel DDR5/LPDDR5X.

Where Each One Wins

The AMD Ryzen AI Embedded P164 wins decisively in single-threaded applications and cryptography. Specifically, it wins in the PassMark single-thread test (4029 vs 2428) and data encryption (16055 vs 11462). The data suggests this processor is ideal for embedded systems that require fast, responsive interactions, real-time control loops, or secure data handling where per-core performance is critical. Its 28-watt TDP also makes it suitable for fanless or battery-powered designs where thermal management is a priority.

The Intel Xeon Gold 5320H wins in all other measured categories, making it the go-to choice for multi-threaded, data-parallel workloads. It wins in data compression (500828 vs 327891), extended instructions (36564 vs 24193), find prime numbers (156 vs 71), floating-point math (73873 vs 55799), integer math (119955 vs 87940), multi-thread (31718 vs 25889), physics (2395 vs 1210), and random string sorting (62730 vs 34801). This makes it the better option for server virtualization, database management, scientific computing, and any workload that can leverage its 40 threads and higher 128.0 GB/s memory bandwidth.