AMD Ryzen Embedded 9600X vs Intel Core Ultra 5 338H Comparison

AMD Ryzen Embedded 9600X vs Intel Core Ultra 5 338H

The Verdict

The database records two distinctly positioned processors. The AMD Ryzen Embedded 9600X is a desktop-class part with 6 cores and 12 threads, while the Intel Core Ultra 5 338H is a mobile processor with 12 cores and 12 threads. The Intel part sits at the 84th percentile of all CPUs in the database, while the AMD part currently holds the 50th percentile, though the AMD entry has no recorded benchmark scores in the current dataset. The Intel Core Ultra 5 338H carries an average benchmark score of 33,989 and is closely matched with its nearest rivals: it trails the Intel Core Ultra 7 165H by 0.3 percent, leads the Intel Core i7-12800HX by 0.3 percent, leads the Intel Xeon 6353P by 0.4 percent, and trails the AMD EPYC 4244P by 0.7 percent. For users selecting strictly from the recorded data, the Intel Core Ultra 5 338H is the only option with verified performance measurements across a wide range of tests. The AMD Ryzen Embedded 9600X has no benchmark entries in the database, meaning no direct comparison can be made from measurements. However, the architecture differences are substantial and can inform a selection based on platform requirements rather than measured performance. The Intel part is a mobile processor with a 25 TDP, while the AMD part is a desktop processor with a 65 TDP. The AMD part uses a 4 nm process from TSMC, while the Intel part uses a 3 nm process from Intel. The AMD part features an unlocked multiplier, while the Intel part is locked. The AMD part supports ECC memory, while the Intel part does not. These are the primary differentiators in the absence of head-to-head benchmark data.

Architecture Differences

The AMD Ryzen Embedded 9600X belongs to the 9000 series and uses the Granite Ridge codename, based on Zen 5 architecture. It is fabricated on a 4 nm process at TSMC, with 8,315 million transistors and a die size of 70.6 mm². The processor has 6 cores and 12 threads, with a base clock of 3.90 GHz and a boost clock of 5.40 GHz. The TDP is 65, and it uses the AMD Socket AM5. The cache hierarchy includes 80 KB of L1 cache per core, 1 MB of L2 cache per core, and 32 MB of shared L3 cache. Memory support is DDR5 with a dual-channel bus, providing 89.6 GB/s of memory bandwidth. It supports ECC memory. PCIe connectivity is Gen 5 with 24 lanes from the CPU. Integrated graphics are provided by Radeon Graphics. The multiplier is unlocked, and the part number is 100-000001405E. The release date recorded is 2025-10-06.

The Intel Core Ultra 5 338H belongs to the Core Ultra Series 3 and uses the Panther Lake codename, based on Panther Lake-H architecture. It is fabricated on a 3 nm process at Intel. The processor has 12 cores and 12 threads, with a base clock of 1.90 GHz and a boost clock of 4.70 GHz. The TDP is 25, and it uses the Intel BGA 2540 socket. The cache hierarchy includes 192 KB of L1 cache per core, 2.5 MB of L2 cache per core, and 18 MB of shared L3 cache. Memory support is LPDDR5X with a dual-channel bus, providing 136.5 GB/s of memory bandwidth. It does not support ECC memory. PCIe connectivity is Gen 5 with 4 lanes from the CPU. Integrated graphics are provided by Arc B370. The multiplier is locked, and the part number is SA4REQ9EW. The release date recorded is 2026-01-04.

The core counts differ significantly: the Intel part has double the cores of the AMD part, but both have 12 threads. The Intel part has a much lower base clock of 1.90 GHz compared to 3.90 GHz, and a lower boost clock of 4.70 GHz compared to 5.40 GHz. The TDP difference is notable, with the Intel part rated at 25 and the AMD part at 65. The memory bandwidth favors the Intel part at 136.5 GB/s versus 89.6 GB/s, but the AMD part supports ECC while the Intel part does not. The PCIe lane count heavily favors the AMD part with 24 lanes versus 4 lanes. The Intel part uses a smaller 3 nm process versus 4 nm, but the AMD part has a larger L3 cache at 32 MB versus 18 MB. The AMD part has an unlocked multiplier, which is absent on the Intel part. The Intel part has more L1 and L2 cache per core, at 192 KB and 2.5 MB respectively, compared to 80 KB and 1 MB on the AMD part.

Head-to-Head Benchmarks

The database contains no head-to-head benchmark entries for these two processors. The AMD Ryzen Embedded 9600X has an empty benchmarks array, and the head-to-head comparison field is also empty. Consequently, there are no direct score comparisons to present. The Intel Core Ultra 5 338H has a full suite of recorded benchmark scores, which can be examined on its own merits. In Cinebench R15 multicore, the Intel part scores 2,504, and in single-core it scores 305. In Cinebench R20, the multicore score is 10,213 and single-core is 1,441. In Cinebench R23, the multicore score is 16,331 and single-core is 2,044. PassMark results show data compression at 276,539, data encryption at 21,367, extended instructions at 23,906, prime number finding at 304, floating point math at 84,067, integer math at 64,934, multithread at 28,717, physics at 2,697, random string sorting at 34,082, and single-thread at 4,180. The average benchmark score is 33,989.

The nearest rivals for the Intel part are closely clustered. The Intel Core Ultra 7 165H averages 34,083, which is 0.3 percent higher than the 338H. The Intel Core i7-12800HX averages 33,875, which is 0.3 percent lower. The Intel Xeon 6353P averages 33,844, which is 0.4 percent lower. The AMD EPYC 4244P averages 34,220, which is 0.7 percent higher. These deltas indicate that the Core Ultra 5 338H sits in a competitive band where performance differences against these specific rivals are within a single percentage point. The 338H is effectively at parity with the Core Ultra 7 165H and the EPYC 4244P, while holding a small edge over the i7-12800HX and the Xeon 6353P. The recorded percentile of 84 places it above the majority of CPUs in the database. The AMD Ryzen Embedded 9600X has no such data, so its relative standing cannot be quantified from the database.

FAQ

Q: Which processor has more cores?

A: The Intel Core Ultra 5 338H has 12 cores, while the AMD Ryzen Embedded 9600X has 6 cores. Both processors have 12 threads.

Q: What is the memory bandwidth difference?

A: The Intel Core Ultra 5 338H supports LPDDR5X memory with a bandwidth of 136.5 GB/s, while the AMD Ryzen Embedded 9600X supports DDR5 memory with a bandwidth of 89.6 GB/s. The Intel part has a higher recorded memory bandwidth.

Q: Does either processor support ECC memory?

A: The AMD Ryzen Embedded 9600X supports ECC memory, while the Intel Core Ultra 5 338H does not.

Q: What are the TDP ratings?

A: The AMD Ryzen Embedded 9600X has a TDP of 65, and the Intel Core Ultra 5 338H has a TDP of 25.

Q: What is the process node for each?

A: The AMD Ryzen Embedded 9600X is fabricated on a 4 nm process at TSMC, and the Intel Core Ultra 5 338H is fabricated on a 3 nm process at Intel.

Q: How does the Intel Core Ultra 5 338H compare to its nearest rivals?

A: The Intel Core Ultra 5 338H has an average benchmark score of 33,989. It is 0.3 percent behind the Intel Core Ultra 7 165H, 0.3 percent ahead of the Intel Core i7-12800HX, 0.4 percent ahead of the Intel Xeon 6353P, and 0.7 percent behind the AMD EPYC 4244P.

Q: Which processor has an unlocked multiplier?

A: The AMD Ryzen Embedded 9600X has an unlocked multiplier, while the Intel Core Ultra 5 338H does not.

Q: What are the PCIe lane counts?

A: The AMD Ryzen Embedded 9600X has Gen 5 with 24 lanes from the CPU, and the Intel Core Ultra 5 338H has Gen 5 with 4 lanes from the CPU.

Where Each One Wins

The Intel Core Ultra 5 338H wins in the domain of measured performance, as it is the only part with recorded benchmarks in the database. Its average benchmark score of 33,989 and 84th percentile ranking indicate solid performance across the tested workloads. The single-thread PassMark score of 4,180 and the Cinebench R23 single-core score of 2,044 show strong per-thread capability. The multicore Cinebench R23 score of 16,331 and PassMark multithread score of 28,717 reflect the benefit of 12 cores. The memory bandwidth of 136.5 GB/s is higher than the AMD part’s 89.6 GB/s, which can favor memory-intensive workloads. The lower TDP of 25 makes it suitable for thermally constrained environments, such as mobile or compact systems. The 3 nm process from Intel indicates a smaller fabrication node, which can influence power efficiency. The larger L1 and L2 caches per core, at 192 KB and 2.5 MB respectively, may benefit latency-sensitive tasks.

The AMD Ryzen Embedded 9600X wins in several architectural categories despite lacking benchmark scores. The higher boost clock of 5.40 GHz versus 4.70 GHz suggests a potential advantage in lightly threaded workloads, though this is not measured in the database. The larger shared L3 cache of 32 MB versus 18 MB can improve hit rates for data sets that fit within that capacity. The PCIe connectivity is substantially stronger, with 24 lanes versus 4 lanes, which supports more expansion devices or high-bandwidth peripherals. ECC memory support is present on the AMD part, which is a critical feature for data integrity in embedded or server-type applications. The unlocked multiplier allows for overclocking, which is not available on the locked Intel part. The desktop form factor with AMD Socket AM5 and a 65 TDP indicates a different platform positioning than the mobile BGA 2540 socket. The 4 nm process from TSMC and the Granite Ridge Zen 5 architecture are distinct from the Intel Panther Lake architecture. The AMD part also has a higher base clock of 3.90 GHz versus 1.90 GHz, which can reduce latency in idle-to-active transitions.

The use-case split is clear from the data. The Intel Core Ultra 5 338H is the choice when measured performance is required, as it has verified scores across Cinebench and PassMark suites. It suits workloads that benefit from 12 cores and 12 threads, higher memory bandwidth, and lower power consumption. The AMD Ryzen Embedded 9600X is the choice when platform features matter more than recorded benchmarks, such as ECC memory support, 24 PCIe lanes, an unlocked multiplier, and a larger L3 cache. The absence of benchmark data for the AMD part means that no performance claim can be made from the database, but its architectural specifications position it for embedded or desktop builds that prioritize connectivity, memory integrity, and overclocking flexibility. The Intel part’s 84th percentile and average score of 33,989 place it in a competitive range against the Core Ultra 7 165H, Core i7-12800HX, Xeon 6353P, and EPYC 4244P, with all deltas within 0.7 percent. The AMD part’s 50th percentile is recorded but with a zero average score, indicating an incomplete dataset. For a direct comparison, the database lacks the necessary head-to-head measurements, so any selection between the two must rely on the architectural traits listed above rather than benchmark outcomes.