AMD Ryzen 9 8945HX vs Intel Core Ultra 5 336H Comparison

Head-to-Head Benchmarks

The head-to-head data shows a decisive overall victory for the AMD Ryzen 9 8945HX, which wins 13 of the 17 recorded benchmark comparisons. The Intel Core Ultra 5 336H takes 4 wins, but the magnitude of the AMD wins is far larger in most tests.

Starting with the Cinebench suite, the AMD Ryzen 9 8945HX leads by a consistent 78% margin across all six tests. In Cinebench R23 multicore, the AMD scores 42713 against 23991 for the Intel part. The single-core R23 result shows 6030 versus 3387, a 78% advantage. The R20 multicore result is 17939 against 10076, and R20 single-core is 2532 versus 1422, a 78.1% delta. The R15 tests follow the same pattern: 4305 versus 2418 in multicore and 607 versus 341 in single-core.

The Passmark suite reveals even larger gaps in specific workloads. The biggest single delta is in integer math, where the AMD scores 195180 against 62070 for the Intel, a 214.5% advantage. Data compression shows 677755 versus 280340, a 141.8% delta. Random string sorting favors AMD by 129%, with scores of 78781 and 34402. Extended instructions show a 103% delta, with 49823 against 24542. Data encryption is 40836 versus 21291, a 91.8% difference. Multithread performance sits at 51405 versus 28545, an 80.1% delta. Floating point math is 117453 versus 82415, a 42.5% advantage for AMD.

The Intel Core Ultra 5 336H wins are narrower but real. In Passmark single-thread, Intel scores 4013 against 3907, a 2.6% delta. The same result appears under the Passmark singlethread entry, confirming the same 4013 versus 3907 outcome. In find prime numbers, Intel scores 299 against 262, a 12.4% delta. In physics, Intel scores 2682 against 2168, a 19.2% delta. These wins show that the Intel part holds its own in specific algorithmic workloads, but they do not offset the scale of AMD's multicore and throughput advantages.

The Verdict

The recorded data points to a clear performance hierarchy. The AMD Ryzen 9 8945HX sits in the 95th percentile of all CPUs in the database, with an average benchmark score of 76212. The Intel Core Ultra 5 336H sits in the 84th percentile with an average score of 34485. That gap of roughly 2.2x in average score is consistent with the head-to-head deltas.

The nearest rivals for the AMD part are all high-end desktop and workstation processors. It sits within 0.6% of the AMD Ryzen 9 9950X3D, within 0.1% of the Intel Core Ultra 9 285HX, and within 0.4% of both the AMD EPYC Embedded 8224P and AMD Ryzen Threadripper PRO 9945WX. This places the 8945HX firmly in flagship territory despite being a mobile part.

The Intel Core Ultra 5 336H, by contrast, sits within 0.8% of the AMD Ryzen AI 7 350, within 0.7% of the AMD Ryzen 7 3700X, within 0.4% of the Intel Core i5-13450HX, and within 0.2% of the Intel Core i7-13700HX. Those rivals are upper-midrange parts, which is the performance class the data assigns to the 336H.

For users whose workloads match the Cinebench rendering tests, the Passmark throughput tests, or the compression and encryption workloads, the AMD part delivers roughly 78% to 214% more performance. For users whose workloads match single-thread Passmark, prime number finding, or physics simulation, the Intel part delivers a modest 2.6% to 19.2% advantage. The data does not support recommending the Intel part for general throughput, but it does show specific niches where it leads.

Architecture Differences

The AMD Ryzen 9 8945HX uses the Zen 4 architecture under the Dragon Range codename, built on a 5 nm process at TSMC. It belongs to the 8000 series and the Ryzen 9 generation. The chip is fabricated from 13,140 million transistors across a die size of 2x 71 mm². It uses an AMD Socket FL1 and has an unlocked multiplier.

The Intel Core Ultra 5 336H uses the Panther Lake architecture under the Panther Lake codename, built on a 3 nm process at Intel. It belongs to the Core Ultra Series 3 and the Ultra 5 generation, specifically Panther Lake-H. It uses an Intel BGA 2540 socket and does not have an unlocked multiplier. The database does not list transistor count or die size for the Intel part.

Cache structures differ substantially. The AMD part uses 64 KB of L1 per core, 1 MB of L2 per core, and 64 MB of L3 cache. The Intel part uses 192 KB of L1 per core, 2.5 MB of L2 per core, and 18 MB of shared L3 cache. The AMD L3 advantage of 64 MB versus 18 MB is significant for workloads that fit within the larger cache. The Intel design has larger per-core L1 and L2 allocations.

Both parts support DDR5 memory in dual-channel configuration, but the Intel part also supports LPDDR5X. The memory bandwidth figures differ: the AMD part records 83.2 GB/s, while the Intel part records 115.2 GB/s. Neither part supports ECC memory.

PCI Express support differs in lane count. The AMD part provides Gen 5 with 28 lanes from the CPU, while the Intel part provides Gen 5 with 12 lanes from the CPU. Both use integrated graphics, with the AMD part using Radeon 610M and the Intel part using Intel Xe3 Graphics.

Specification Differences

The core and thread counts differ despite both parts having 16 cores. The AMD Ryzen 9 8945HX supports 32 threads, indicating simultaneous multithreading. The Intel Core Ultra 5 336H supports 16 threads, matching its core count with no multithreading. This explains part of the large multicore deltas.

Clock speeds differ. The AMD part has a base clock of 2.50 GHz and a boost clock of 5.40 GHz. The Intel part has a base clock of 1.90 GHz and a boost clock of 4.60 GHz. The AMD part runs both higher at base and higher at boost.

Thermal design power differs by more than 2x. The AMD part is rated at 55 W TDP, while the Intel part is rated at 25 W TDP. This power envelope difference is consistent with the performance gap, but the database does not record actual power consumption.

Release dates differ. The AMD part entered production on 2025-04-22, while the Intel part entered production on 2026-01-04. Both are listed as Active in production status. Neither part has a launch MSRP recorded in the database.

The part numbers differ: the AMD unit is 100-000001848, and the Intel unit is SA4RD. Market segment for both is Mobile.

FAQ

Q: Which processor has more threads?

A: The AMD Ryzen 9 8945HX has 32 threads, while the Intel Core Ultra 5 336H has 16 threads. Both have 16 cores.

Q: What is the largest performance difference between the two?

A: The largest delta is in Passmark integer math, where the AMD part scores 195180 against 62070, a 214.5% advantage. The next largest is data compression at 141.8% in favor of AMD.

Q: Does the Intel processor win any benchmarks?

A: Yes, the Intel Core Ultra 5 336H wins four benchmarks: Passmark single-thread (4013 versus 3907, a 2.6% delta), Passmark singlethread (same scores), Passmark find prime numbers (299 versus 262, a 12.4% delta), and Passmark physics (2682 versus 2168, a 19.2% delta).

Q: How do the cache sizes compare?

A: The AMD part has 64 KB L1 per core, 1 MB L2 per core, and 64 MB L3. The Intel part has 192 KB L1 per core, 2.5 MB L2 per core, and 18 MB shared L3.

Q: What process nodes are used?

A: The AMD Ryzen 9 8945HX is built on a 5 nm process at TSMC. The Intel Core Ultra 5 336H is built on a 3 nm process at Intel.

Q: How do the average benchmark scores compare?

A: The AMD part has an average benchmark score of 76212, placing it in the 95th percentile. The Intel part has an average score of 34485, placing it in the 84th percentile.

Where Each One Wins

The AMD Ryzen 9 8945HX dominates rendering and throughput workloads. The Cinebench R23 multicore score of 42713 versus 23991 indicates that the AMD part is the appropriate choice for 3D rendering, video encoding, and other tasks that scale with thread count. The 32 threads against 16 threads give it a structural advantage. The Passmark multithread score of 51405 versus 28545 confirms this. The data compression score of 677755 versus 280340 makes the AMD part the better pick for archiving, file compression, and database workloads. The integer math score of 195180 versus 62070 points to strong general compute performance. The extended instructions score of 49823 versus 24542 and the encryption score of 40836 versus 21291 reinforce the AMD advantage in cryptographic and SIMD-heavy workloads.

The AMD part also wins the memory-latency-sensitive workloads through its 64 MB L3 cache, which is more than 3.5x the Intel part's 18 MB. The 5.40 GHz boost clock gives it a single-core Cinebench advantage of 78%, even though the Intel part wins the Passmark single-thread test by a narrow 2.6%. The AMD part also offers more PCI Express lanes, with 28 Gen 5 lanes against 12 Gen 5 lanes, which matters for users connecting multiple high-bandwidth devices.

The Intel Core Ultra 5 336H wins in specific algorithmic workloads. The physics score of 2682 versus 2168, a 19.2% delta, suggests that the Intel architecture handles physics simulation more efficiently. The find prime numbers score of 299 versus 262, a 12.4% delta, points to an advantage in integer-heavy algorithmic loops. The single-thread Passmark score of 4013 versus 3907, a 2.6% delta, indicates a slight edge in low-thread-count general tasks. The Intel part also has a higher memory bandwidth figure at 115.2 GB/s versus 83.2 GB/s, and it supports LPDDR5X memory, which may benefit bandwidth-sensitive applications. The 25 W TDP makes it the lower-power option by a 30 W margin, which could matter in thermally constrained chassis.

The database also shows the Intel part supports more L1 and L2 cache per core, with 192 KB L1 and 2.5 MB L2 versus 64 KB L1 and 1 MB L2 on the AMD part. This per-core cache advantage may explain the physics and prime number wins. The Intel part's 3 nm process gives it a node advantage over the AMD's 5 nm process, though the recorded performance data does not translate that into overall wins.

The 3.5x L3 cache advantage, the 2x thread advantage, and the higher clocks all favor the AMD part for sustained multicore work. The Intel part's wins are real but isolated to workloads that favor its per-core cache design and specific instruction patterns. For any mixed workload that includes rendering, compression, or encryption, the data shows the AMD part is more than three quarters faster in many cases.