AMD Ryzen AI Max PRO 390 vs Intel Core Ultra 5 338H Comparison

FAQ

Q: Which processor has the higher overall average benchmark score?

A: The AMD Ryzen AI Max PRO 390 records an average benchmark score of 63765, while the Intel Core Ultra 5 338H records 33989. The AMD part sits at the 93rd percentile among all CPUs, compared to the 84th percentile for the Intel part.

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

A: The Intel Core Ultra 5 338H leads in single-threaded workloads. It wins Cinebench R23 single-core with a score of 2044 versus 1976 for the AMD Ryzen AI Max PRO 390, a 3.3% advantage. In PassMark single-thread, Intel scores 4180 versus 3967, a 5.1% margin.

Q: Which processor wins in multi-threaded workloads?

A: The AMD Ryzen AI Max PRO 390 dominates multi-threaded tests. In Cinebench R23 multi-core, AMD scores 24828 versus 16331 for Intel, a 52% advantage. In Cinebench R15 multi-core, AMD scores 3918 versus 2504, a 56.5% lead. PassMark multithread shows AMD at 42912 versus 28717, a 49.4% margin.

Q: What are the core and thread counts for each processor?

A: Both processors have 12 cores. The AMD Ryzen AI Max PRO 390 supports 24 threads, while the Intel Core Ultra 5 338H supports 12 threads. The AMD part uses simultaneous multithreading, effectively doubling thread count; the Intel part does not.

Q: Which processor has higher memory bandwidth?

A: The AMD Ryzen AI Max PRO 390 uses a quad-channel memory bus with 256.0 GB/s bandwidth. The Intel Core Ultra 5 338H uses a dual-channel bus with 136.5 GB/s bandwidth. AMD's memory bandwidth is nearly double that of Intel.

Q: What are the manufacturing process nodes for each chip?

A: The AMD Ryzen AI Max PRO 390 is built on a 4 nm process at TSMC, while the Intel Core Ultra 5 338H uses a 3 nm process at Intel. Despite the smaller process node for Intel, the AMD part delivers substantially higher multi-threaded performance in the recorded benchmarks.

Architecture Differences

The AMD Ryzen AI Max PRO 390 and Intel Core Ultra 5 338H represent different architectural approaches from their respective manufacturers. AMD uses the Zen 5 architecture under the Strix Halo codename, part of the Ryzen AI Max PRO generation. Intel uses the Panther Lake architecture, part of the Core Ultra Series 3, under the Panther Lake-H generation.

The process nodes differ. AMD fabricates the Ryzen AI Max PRO 390 on a 4 nm process at TSMC, while Intel builds the Core Ultra 5 338H on a 3 nm process at Intel's own foundry. This gives Intel a process geometry advantage, but the benchmark data shows that architectural efficiency and other factors play a larger role in overall performance.

Core organization differs significantly. The AMD part has 12 cores and 24 threads, indicating simultaneous multithreading support. The Intel part also has 12 cores but only 12 threads, meaning each core handles a single thread. This directly affects multi-threaded workload performance, as the AMD processor can process twice as many threads concurrently.

Cache hierarchies show notable differences. The AMD Ryzen AI Max PRO 390 has 80 KB of L1 cache per core, 1 MB of L2 cache per core, and 64 MB of shared L3 cache. The Intel Core Ultra 5 338H has 192 KB of L1 cache per core, 2.5 MB of L2 cache per core, and 18 MB of shared L3 cache. Intel has larger per-core caches at the L1 and L2 levels, but AMD's shared L3 cache is substantially larger at 64 MB versus 18 MB.

Memory architecture diverges sharply. AMD uses a quad-channel LPDDR5X memory bus delivering 256.0 GB/s of bandwidth. Intel uses a dual-channel LPDDR5X bus delivering 136.5 GB/s. The AMD part also supports ECC memory, while the Intel part does not. For memory-intensive workloads, the AMD processor has a clear structural advantage.

PCI Express support differs. The AMD Ryzen AI Max PRO 390 provides Gen 4 with 16 lanes for the CPU. The Intel Core Ultra 5 338H provides Gen 5 with 4 lanes for the CPU. Intel's newer PCIe generation offers higher per-lane bandwidth, but AMD offers more lanes.

Integrated graphics also differ. The AMD part pairs with the Radeon 8050S, while the Intel part uses the Arc B370. Both processors target the mobile segment and are currently in active production. The AMD part uses Socket FP11, while Intel uses BGA 2540.

Head-to-Head Benchmarks

The head-to-head benchmark data shows a clear overall winner, with the AMD Ryzen AI Max PRO 390 winning 11 of the 15 recorded tests and the Intel Core Ultra 5 338H winning 4. The magnitude of the AMD wins, however, varies considerably across workload types.

Multi-core rendering heavily favors AMD. In Cinebench R23 multi-core, AMD scores 24828 against Intel's 16331, a 52% advantage. Cinebench R15 multi-core shows a similar pattern: AMD scores 3918, Intel scores 2504, a 56.5% lead. These results reflect the thread count difference, as AMD processes 24 threads versus Intel's 12.

Integer math demonstrates the largest performance gap. PassMark integer math shows AMD scoring 148508 against Intel's 64934, a 128.7% advantage. This is the single biggest delta in the entire comparison. AMD more than doubles Intel's integer throughput, which suggests a substantial per-core efficiency advantage in this workload in addition to the thread count benefit.

Data compression heavily favors AMD. PassMark data compression shows AMD at 506170 versus 276539 for Intel, an 83% advantage. Extended instructions testing shows AMD scoring 40888 against 23906, a 71% margin. Random string sorting shows AMD at 55248 versus 34082, a 62.1% lead. These results indicate that AMD's architecture handles data manipulation and sorting tasks with significantly higher throughput.

Floating point math favors AMD by a smaller margin. PassMark floating point math shows AMD at 94666 versus 84067 for Intel, a 12.6% advantage. This is a more modest lead than the integer results, suggesting that the floating point units are more closely matched.

Data encryption shows AMD ahead by 21.8%. The AMD part scores 26027 against Intel's 21367. Prime number finding shows a narrow AMD lead of 6.3%, with scores of 323 and 304 respectively. Physics simulation shows AMD ahead by 2.8%, with scores of 2773 and 2697.

PassMark multithread shows AMD at 42912 versus 28717 for Intel, a 49.4% advantage. This aligns with the Cinebench multi-core results, confirming that the AMD part maintains a strong lead across different multi-threaded benchmark suites.

The Intel Core Ultra 5 338H wins all four single-threaded tests. Cinebench R23 single-core shows Intel at 2044 versus 1976 for AMD, a 3.3% lead. Cinebench R15 single-core shows Intel at 305 versus 303, a narrow 0.7% margin. PassMark single-thread shows Intel at 4180 versus 3967, a 5.1% lead. The single-thread data appears twice in the head-to-head results as both passmark_single_thread and passmark_singlethread, both showing identical scores.

The overall picture is one of AMD dominance in multi-threaded and throughput-oriented workloads, with Intel holding a modest but consistent edge in single-threaded tasks. The AMD part's advantages range from 2.8% to 128.7% depending on the workload, while Intel's advantages range from 0.7% to 5.1%.

Specification Differences

The two processors differ in several key specification fields. Clock speeds show AMD with a 3.20 GHz base clock and 5.00 GHz boost clock, while Intel runs at 1.90 GHz base and 4.70 GHz boost. AMD has a higher base clock by 1.30 GHz and a higher boost clock by 0.30 GHz.

Thermal design power differs substantially. The AMD Ryzen AI Max PRO 390 has a TDP of 55 watts, while the Intel Core Ultra 5 338H has a TDP of 25 watts. This means the Intel part is designed for a lower power envelope, which may influence its placement in thinner or more power-constrained systems.

Thread count differs as noted: AMD supports 24 threads, Intel supports 12 threads, both with 12 cores. Memory bus width differs: AMD uses quad-channel, Intel uses dual-channel. Memory bandwidth differs: AMD delivers 256.0 GB/s, Intel delivers 136.5 GB/s.

ECC memory support differs: AMD supports it, Intel does not. PCIe configuration differs: AMD provides Gen 4 with 16 lanes, Intel provides Gen 5 with 4 lanes. Cache sizes differ at every level. AMD has 80 KB L1 per core, 1 MB L2 per core, and 64 MB shared L3. Intel has 192 KB L1 per core, 2.5 MB L2 per core, and 18 MB shared L3.

Process node and foundry differ: AMD uses 4 nm at TSMC, Intel uses 3 nm at Intel. Sockets differ: AMD uses Socket FP11, Intel uses BGA 2540. Integrated graphics differ: AMD pairs with Radeon 8050S, Intel with Arc B370. Release dates differ: AMD launched on 2025-01-05, Intel on 2026-01-04.

Part numbers also differ: the AMD unit is 100-000001421, while the Intel unit is SA4REQ9EW. Both processors are locked, with the multiplier unlocked field set to false. Both are mobile market segments and in active production. Neither has a recorded launch MSRP.

The Verdict

The benchmark data indicates that the AMD Ryzen AI Max PRO 390 is the stronger processor for throughput-oriented workloads. Its 24 threads, 64 MB of shared L3 cache, and 256.0 GB/s memory bandwidth give it decisive advantages in multi-threaded rendering, integer math, data compression, and encryption. The 52% lead in Cinebench R23 multi-core and the 128.7% lead in PassMark integer math are substantial margins that would translate into meaningful real-world differences in CPU-intensive tasks.

The Intel Core Ultra 5 338H holds the advantage in single-threaded workloads, but the margins are small. Its largest single-threaded win is 5.1% in PassMark single-thread, and its Cinebench R23 single-core lead is only 3.3%. These margins are unlikely to compensate for the large multi-threaded deficits in most usage scenarios.

The Intel part's lower TDP of 25 watts versus AMD's 55 watts suggests it is designed for power-constrained environments. The data does not include power efficiency metrics, so no direct efficiency comparison is possible from the recorded measurements. The Intel part also uses a 3 nm process, which is a smaller geometry than AMD's 4 nm, but the benchmark results do not show a performance benefit from this advantage.

The average benchmark score difference is substantial: 63765 for AMD versus 33989 for Intel. The AMD part ranks at the 93rd percentile of all CPUs, compared to the 84th percentile for Intel. Its nearest rivals include the Intel Core i9-13900KS with a delta of -0.4% and the AMD Ryzen AI 7 450G with a delta of 0.7%. The Intel part's nearest rivals include the Intel Core Ultra 7 165H with a delta of -0.3% and the Intel Core i7-12800HX with a delta of 0.3%. These rival comparisons show that the AMD part competes with high-end desktop processors, while the Intel part competes with mid-range mobile and workstation parts.

Where Each One Wins

The AMD Ryzen AI Max PRO 390 wins in every multi-threaded and throughput-oriented category. This includes Cinebench R15 and R23 multi-core, PassMark multithread, data compression, data encryption, extended instructions, find prime numbers, floating point math, integer math, physics, and random string sorting. Users running CPU rendering, video encoding, data analytics, scientific computation, or database workloads would benefit from the AMD part's higher throughput.

The largest AMD advantages come in integer math (128.7% ahead), data compression (83% ahead), and extended instructions (71% ahead). These workloads benefit from the combination of 24 threads, the large 64 MB L3 cache, and the 256.0 GB/s memory bandwidth. The data suggests that the AMD part is particularly well suited to workloads that process large datasets or require heavy parallel computation.

The Intel Core Ultra 5 338H wins in single-threaded performance categories. Its advantages are 5.1% in PassMark single-thread, 3.3% in Cinebench R23 single-core, and 0.7% in Cinebench R15 single-core. These wins are consistent across all single-threaded tests, indicating a genuine per-core performance advantage for the Intel architecture. Applications that depend heavily on single-threaded responsiveness, such as certain legacy software or lightly threaded interactive applications, would see a small benefit from the Intel part.

The Intel part also has a lower TDP at 25 watts, which may make it more suitable for systems with limited cooling or battery capacity. The data does not include battery life or thermal measurements, so this remains a specification-level observation rather than a tested result.

Overall, the AMD Ryzen AI Max PRO 390 delivers the higher performance in the majority of recorded benchmarks, with its strongest results in multi-threaded and memory-intensive workloads. The Intel Core Ultra 5 338H offers a modest single-threaded advantage within a lower power envelope. The choice between them depends on whether the workload prioritizes parallel throughput or single-threaded responsiveness.