AMD Ryzen Threadripper 9970X vs Intel Core Ultra X7 368H Comparison

Head-to-Head Benchmarks

The benchmark data shows a complete sweep: the AMD Ryzen Threadripper 9970X wins all 11 head-to-head tests against the Intel Core Ultra X7 368H. The largest margins appear in heavily parallel workloads. In data compression, AMD scores 1,757,998 versus Intel's 312,927, a 461.8% advantage. Extended instructions show a 454.5% gap (142,342 versus 25,669), and random string sorting delivers 191,445 against 38,093, a 402.6% difference. Integer math follows at 428.3% (465,378 versus 88,083). These four tests all exceed a 4x performance ratio, indicating the AMD processor's raw throughput advantage in algorithmic and data-processing tasks.

The multithreaded PassMark score reinforces the pattern. AMD records 107,399, while Intel manages 32,956, a 225.9% lead. Floating-point math shows a 193.1% margin (309,719 versus 105,681), and physics simulation scores 6,835 versus 2,857, a 139.2% gap. Data encryption delivers 86,765 against 25,228, a 243.9% difference. The smallest victories are in single-threaded tests: AMD scores 4,530 versus Intel's 4,005, a 13.1% advantage. Find prime numbers shows a 91.6% gap (615 versus 321), which is notable because this workload typically scales less with core count.

The average benchmark score tells the same story. AMD's average is 279,778, placing it in the 99th percentile of all CPUs. Intel's average is 40,518, in the 87th percentile. The nearest rivals for AMD include the Intel Xeon 6780E (280,438 average, 0.2% lower), AMD EPYC 9565 (285,471 average, 2% lower), Intel Xeon 696X (286,102 average, 2.2% lower), and AMD EPYC 9555P (287,066 average, 2.5% lower). For Intel, the closest competitors are the Intel Core 7 253PE (40,557 average, 0.1% lower), Intel Core 5 223PE (40,585 average, 0.2% lower), AMD Ryzen 9 7940H (40,431 average, 0.2% higher), and Intel Xeon 6507P (40,426 average, 0.2% higher). In every head-to-head comparison, the AMD processor outperforms the Intel chip, and the deltas are substantial.

Where Each One Wins

The AMD Ryzen Threadripper 9970X dominates every measured category. Its largest advantages are in data compression (461.8%), extended instructions (454.5%), random string sorting (402.6%), and integer math (428.3%). These workloads benefit from the 32 cores and 64 threads, which allow massive parallel execution. The AMD processor also leads in multithreaded performance by 225.9%, making it the clear choice for rendering, compilation, and simulation tasks. Its single-thread lead of 13.1% means it is also faster in lightly threaded applications, though the margin is far smaller.

The Intel Core Ultra X7 368H shows no benchmark win in any category. Its best relative performance comes in single-threaded tests, where it trails by only 13.1%. The data indicates that Intel's chip is competitive in per-core efficiency but cannot match the AMD processor's aggregate throughput. For workloads that rely on a single thread, such as some legacy applications or lightly threaded games, the Intel part is closer to parity but still behind. The recorded data shows no scenario where Intel wins outright.

The average benchmark scores place the AMD processor in the 99th percentile, while Intel sits in the 87th percentile. The nearest rival data confirms that AMD's competition is in the server and workstation segment, with the Xeon 6780E trailing by only 0.2%, while Intel's mobile chip competes with mid-range laptop processors like the Ryzen 9 7940H, which is 0.2% ahead. This positioning matters for use-case analysis: the AMD processor belongs in desktop workstation environments, while the Intel chip targets mobile platforms with lower power envelopes.

Architecture Differences

The AMD Ryzen Threadripper 9970X uses the Zen 5 architecture on a 4 nm process from TSMC, with the codename Shimada Peak. It belongs to the Ryzen Threadripper generation (Zen 5, Shimada Peak). The processor has 32 cores and 64 threads, with a base clock of 4.00 GHz and a boost clock of 5.40 GHz. It features 128 MB of L3 cache, with 64 KB of L1 per core and 1 MB of L2 per core. The transistor count is 33,260 million, spread across four dies, each measuring 70.6 mm². Memory support includes DDR5 with a quad-channel bus, delivering 204.8 GB/s of bandwidth. ECC memory is supported. The PCIe interface offers Gen 5 with 80 lanes from the CPU. It uses AMD Socket sTR5, has a 350 W TDP, and includes no integrated graphics. The multiplier is unlocked, and the launch MSRP is $2499.

The Intel Core Ultra X7 368H uses the Panther Lake architecture on a 3 nm process from Intel. It belongs to the Core Ultra Series 3 generation (Ultra X7, Panther Lake-H). The processor has 16 cores and 16 threads, with a base clock of 2.00 GHz and a boost clock of 5.00 GHz. It features 18 MB of shared L3 cache, with 192 KB of L1 per core and 2.5 MB of L2 per core. Memory support includes LPDDR5X with a dual-channel bus, delivering 153.6 GB/s of bandwidth. ECC memory is not supported. The PCIe interface offers Gen 5 with only 4 lanes from the CPU. It uses Intel BGA 2540, has a 25 W TDP, and includes integrated Arc B390 graphics. The multiplier is locked. No transistor count or die size data is recorded for this processor.

The architectural differences are stark. AMD offers 32 cores versus Intel's 16, with simultaneous multithreading doubling AMD's thread count to 64, while Intel provides no additional threads. AMD's L3 cache is 128 MB versus Intel's 18 MB, a 7.1x difference. The memory bus is quad-channel versus dual-channel, and bandwidth is 204.8 GB/s versus 153.6 GB/s. The process nodes differ (4 nm versus 3 nm), but AMD uses more transistors (33,260 million versus no recorded figure for Intel). The TDP gap is enormous: 350 W versus 25 W, reflecting the desktop versus mobile market segments. AMD supports ECC memory, Intel does not. AMD provides 80 PCIe Gen 5 lanes, Intel provides 4. These differences explain the benchmark outcomes: more cores, more cache, wider memory, and more PCIe bandwidth translate directly into the measured performance advantages.

The Verdict

The data supports a clear conclusion: the AMD Ryzen Threadripper 9970X is the significantly faster processor in every benchmark recorded. For workloads that demand maximum multithreaded throughput, data compression, encryption, extended instruction processing, integer math, floating-point math, physics simulation, random string sorting, and even single-threaded performance, AMD wins. The margins range from 13.1% in single-threaded tests to 461.8% in data compression. The average benchmark score of 279,778 versus 40,518 reflects a 6.9x difference, and the percentile ranking (99th versus 87th) confirms that the AMD processor operates in a different performance class.

Users who need a desktop workstation processor for heavy parallel tasks, such as 3D rendering, scientific computing, or large-scale data processing, should choose the AMD Ryzen Threadripper 9970X. Its 32 cores, 64 threads, 128 MB of L3 cache, quad-channel DDR5 memory, and 80 PCIe Gen 5 lanes provide the infrastructure for sustained high-throughput workloads. The 350 W TDP and unlocked multiplier indicate that this is a power-hungry, overclockable part designed for maximum performance.

The Intel Core Ultra X7 368H suits a different scenario. With 16 cores, 16 threads, 18 MB of L3 cache, dual-channel LPDDR5X memory, and 4 PCIe Gen 5 lanes, it is a mobile processor with a 25 W TDP. It includes integrated Arc B390 graphics, which AMD lacks. Its performance is competitive with other mobile chips, as the nearest rival data shows (Core 7 253PE is 0.1% lower, Core 5 223PE is 0.2% lower, Ryzen 9 7940H is 0.2% higher). For users seeking a laptop processor with modest power consumption and integrated graphics, the Intel part is appropriate. The single-thread score of 4,005 is only 13.1% behind AMD, so lightly threaded tasks are less disadvantaged.

The benchmark data does not support any scenario where Intel wins a head-to-head test. The wins column shows 11 for AMD and 0 for Intel. However, the market segments differ so fundamentally that direct comparison may be less relevant than the architecture differences suggest. The AMD processor is a desktop workstation part with a 350 W TDP, while the Intel processor is a mobile part with a 25 W TDP. The performance gap is expected given the power and thermal budgets. The verdict is straightforward: for raw performance, choose AMD; for mobile efficiency with integrated graphics, choose Intel.

FAQ

Q: Which processor has more cores?

A: The AMD Ryzen Threadripper 9970X has 32 cores, while the Intel Core Ultra X7 368H has 16 cores.

Q: What is the single-threaded performance difference?

A: In the PassMark single-thread test, AMD scores 4,530 and Intel scores 4,005, giving AMD a 13.1% lead.

Q: Does the Intel processor support ECC memory?

A: No. The Intel Core Ultra X7 368H does not support ECC memory, while the AMD Ryzen Threadripper 9970X does.

Q: What is the multithreaded performance gap?

A: In the PassMark multithread test, AMD scores 107,399 and Intel scores 32,956, so AMD is 225.9% ahead.

Q: Which processor has integrated graphics?

A: The Intel Core Ultra X7 368H includes Arc B390 integrated graphics. The AMD Ryzen Threadripper 9970X has no integrated graphics.

Q: What are the memory bandwidth figures?

A: AMD provides 204.8 GB/s with quad-channel DDR5, while Intel provides 153.6 GB/s with dual-channel LPDDR5X.

Specification Differences

| Specification | AMD Ryzen Threadripper 9970X | Intel Core Ultra X7 368H |

|----------------|------------------------------|--------------------------|

| Cores | 32 | 16 |

| Threads | 64 | 16 |

| Base clock | 4.00 GHz | 2.00 GHz |

| Boost clock | 5.40 GHz | 5.00 GHz |

| TDP | 350 W | 25 W |

| Socket | AMD Socket sTR5 | Intel BGA 2540 |

| Process node | 4 nm | 3 nm |

| L1 cache | 64 KB per core | 192 KB per core |

| L2 cache | 1 MB per core | 2.5 MB per core |

| L3 cache | 128 MB | 18 MB shared |

| Memory support | DDR5 | LPDDR5X |

| Memory bus | Quad-channel | Dual-channel |

| Memory bandwidth | 204.8 GB/s | 153.6 GB/s |

| ECC memory | Yes | No |

| PCIe lanes | Gen 5, 80 lanes | Gen 5, 4 lanes |

| Integrated graphics | N/A | Arc B390 |

| Market segment | Desktop | Mobile |

| Multiplier unlocked | Yes | No |

| Launch MSRP | $2499 | Not recorded |

| Transistors | 33,260 million | Not recorded |

| Die size | 4x 70.6 mm² | Not recorded |

| Release date | 2025-07-29 | 2026-01-04 |