AMD Ryzen Threadripper 9970X vs Intel Core Ultra 9 285 Comparison

Where Each One Wins

The benchmark split is decisive. The AMD Ryzen Threadripper 9970X wins 9 of the 11 recorded head-to-head tests, while the Intel Core Ultra 9 285 wins only 2. That nearly clean sweep tells a clear story: this is a processor built to dominate heavy, parallel workloads.

The Threadripper 9970X takes every multithreaded and compute-heavy discipline. Data compression shows a 192% lead, integer math is 182.3% ahead, and extended instructions (SIMD-heavy code) runs 213.8% faster. Random string sorting, a classic many-thread test, lands at 159.9% ahead. Physics simulation, which scales well with core count, shows a 90% advantage. Floating-point math is 58.8% faster, data encryption is 84.8% faster, and the PassMark multithread score is 89.7% higher. Even prime number generation, a workload that often favors frequency, goes to AMD by 34%.

The Intel Core Ultra 9 285 wins exactly two tests, both single-thread measurements: PassMark single-thread at 4881 versus 4530, a 7.2% advantage. That is the only category where the Intel part leads, and it does so by a modest margin compared to the scale of AMD's multithreaded wins.

What this means in practice: any workload that can use more than a handful of threads will be dramatically faster on the Threadripper. The Intel part is the better choice for lightly threaded tasks where raw single-core speed matters more than core count, but the gap is narrow. The Threadripper's single-thread score is not weak; it trails by a small percentage, and its multithreaded dominance makes it the stronger overall package for almost everything else.

Architecture Differences

The two processors come from fundamentally different design philosophies. The AMD Ryzen Threadripper 9970X uses the Zen 5 architecture on the Shimada Peak codename, built on TSMC's 4 nm process. It packs 32 cores and 64 threads, with a base clock of 4.00 GHz and a boost clock of 5.40 GHz. The transistor count is listed at 33,260 million, spread across four chiplets at 70.6 mm² each. Cache structure is generous: 64 KB of L1 per core, 1 MB of L2 per core, and 128 MB of shared L3. Memory support is DDR5 over a quad-channel bus, delivering 204.8 GB/s of bandwidth. ECC memory is supported. PCIe connectivity is Gen 5 with 80 lanes from the CPU. The socket is AMD Socket sTR5, and the TDP is 350 watts. The multiplier is unlocked.

The Intel Core Ultra 9 285 uses the Arrow Lake architecture on the Arrow Lake-S codename, built on TSMC's 3 nm process. It has 24 cores and 24 threads, meaning no hyperthreading; each core is a single thread. Base clock is 2.50 GHz, boost is 5.60 GHz. Transistor count is 17,800 million on a 243 mm² die. Cache is smaller: 192 KB of L1 per core, 3 MB of L2 per core, and 36 MB of shared L3. Memory support is DDR5 over a dual-channel bus, giving 102.4 GB/s of bandwidth. ECC memory is supported. PCIe is Gen 5 with 20 lanes from the CPU. The socket is Intel Socket 1851, and the TDP is 65 watts. The multiplier is locked. It also includes integrated graphics: Arc Xe-LPG Graphics 64EU.

The core count difference is the headline: 32 cores and 64 threads versus 24 cores and 24 threads. The Threadripper has more than double the thread count, which directly explains its multithreaded benchmark dominance. The Intel part compensates with a slightly higher boost clock (5.60 GHz versus 5.40 GHz) and a smaller, denser process node, which contributes to its single-thread win. Memory bandwidth is also lopsided: 204.8 GB/s versus 102.4 GB/s, a direct result of quad-channel versus dual-channel memory buses. PCIe lane count is 80 versus 20, which matters for expansion-heavy workstation builds.

The TDP figures are not directly comparable in the sense of efficiency, but the numbers are what they are: 350 watts for AMD versus 65 watts for Intel. The Intel part draws far less power on paper, while the AMD part spends its much higher budget on feeding 64 threads.

FAQ

Q: Which processor is faster in single-threaded workloads?

A: The Intel Core Ultra 9 285. It scores 4881 in PassMark single-thread versus 4530 for the AMD Ryzen Threadripper 9970X, a 7.2% advantage.

Q: How large is the multithreaded performance gap?

A: The Threadripper 9970X leads by 89.7% in the PassMark multithread test, scoring 107399 versus 56602. In data compression, the lead grows to 192%.

Q: Does the Intel Core Ultra 9 285 support ECC memory?

A: Yes, both processors support ECC memory. The AMD Ryzen Threadripper 9970X and the Intel Core Ultra 9 285 both list ECC memory support as true.

Q: What is the memory bandwidth difference?

A: The Threadripper 9970X uses quad-channel DDR5 with 204.8 GB/s of bandwidth. The Core Ultra 9 285 uses dual-channel DDR5 with 102.4 GB/s. The AMD part has exactly double the bandwidth.

Q: Which processor has more PCIe lanes?

A: The AMD Ryzen Threadripper 9970X provides 80 Gen 5 lanes from the CPU, while the Intel Core Ultra 9 285 provides 20 Gen 5 lanes. The Threadripper has four times the lane count.

Q: Does the Intel part include integrated graphics?

A: Yes, the Core Ultra 9 285 includes Arc Xe-LPG Graphics 64EU. The Threadripper 9970X lists no integrated graphics.

Specification Differences

The two processors differ across nearly every major specification field.

• Cores: AMD has 32, Intel has 24.
• Threads: AMD has 64, Intel has 24.
• Base clock: AMD 4.00 GHz, Intel 2.50 GHz.
• Boost clock: Intel 5.60 GHz, AMD 5.40 GHz.
• TDP: AMD 350 W, Intel 65 W.
• Socket: AMD Socket sTR5, Intel Socket 1851.
• Architecture: Zen 5 (Shimada Peak) versus Arrow Lake (Arrow Lake-S).
• Process node: TSMC 4 nm for AMD, TSMC 3 nm for Intel.
• Transistors: 33,260 million for AMD, 17,800 million for Intel.
• Die size: 4x 70.6 mm² for AMD, 243 mm² for Intel.
• L1 cache: 64 KB per core for AMD, 192 KB per core for Intel.
• L2 cache: 1 MB per core for AMD, 3 MB per core for Intel.
• L3 cache: 128 MB for AMD, 36 MB for Intel.
• Memory bus: Quad-channel for AMD, dual-channel for Intel.
• Memory bandwidth: 204.8 GB/s for AMD, 102.4 GB/s for Intel.
• PCIe lanes: 80 Gen 5 for AMD, 20 Gen 5 for Intel.
• Integrated graphics: None for AMD, Arc Xe-LPG Graphics 64EU for Intel.
• Multiplier unlocked: Yes for AMD, no for Intel.
• Release date: AMD 2025-07-29, Intel 2024-12-31.
• Launch MSRP: AMD $2499, Intel $579.
• Percentile vs all CPUs: AMD 99th, Intel 95th.
• Average benchmark score: AMD 279778, Intel 75488.

The specification sheet reads like two different product categories. The Threadripper is a workstation-class part with massive core counts, cache, memory bandwidth, and PCIe expansion. The Core Ultra 9 is a high-end desktop part with a much lower TDP, higher boost clock, and integrated graphics.

Head-to-Head Benchmarks

The head-to-head results are one-sided, but the individual deltas show where each architecture's strengths lie.

The largest AMD win is in extended instructions, at 213.8% ahead. This test exercises SIMD and specialized instruction paths, and the Threadripper's 32 Zen 5 cores with 64 threads simply overwhelm the Intel part's 24 threads. Data compression follows at 192%, and integer math at 182.3%. These are all workloads that scale almost linearly with core count, so the thread advantage translates directly into score.

Random string sorting shows a 159.9% lead. This is a memory-access-heavy test, and the Threadripper's quad-channel 204.8 GB/s bandwidth versus the Intel part's 102.4 GB/s explains much of the gap. Physics simulation is 90% ahead, and the multithread score is 89.7% higher. Floating-point math is 58.8% ahead, a smaller margin than the integer tests, but still a decisive win. Data encryption is 84.8% ahead, and prime number generation is 34% ahead.

The Intel Core Ultra 9 285 wins only the single-thread tests, and by a slim 7.2% margin. Its boost clock of 5.60 GHz and the smaller 3 nm process likely contribute, but the score difference is small relative to the multithreaded gaps. In absolute terms, AMD's single-thread score of 4530 is within reach of Intel's 4881, so the Intel part's single-thread advantage does not compensate for its multithreaded deficit.

The average benchmark scores reinforce the gap: the Threadripper averages 279778 across the recorded benchmarks, while the Core Ultra 9 averages 75488. The nearest rivals for the Threadripper are server-class parts: the Intel Xeon 6780E at 280438 (0.2% ahead), the AMD EPYC 9565 at 285471 (2% ahead), the Intel Xeon 696X at 286102 (2.2% ahead), and the AMD EPYC 9555P at 287066 (2.5% ahead). The Core Ultra 9's nearest rivals are EPYC and Ryzen PRO parts in the 75373 to 75738 range, all within 0.3% of its score. The Threadripper sits in a much higher performance tier.

The Verdict

The data points to a clear split by workload type. For anyone running heavily threaded, compute-intensive tasks: rendering, simulations, data compression, encryption, scientific computing, the AMD Ryzen Threadripper 9970X is the correct choice. It wins 9 of 11 head-to-head tests, with leads ranging from 34% to 213.8% in the categories that matter for multithreaded throughput. Its 64 threads, 128 MB of L3 cache, quad-channel memory bandwidth, and 80 PCIe lanes make it a workstation engine.

The Intel Core Ultra 9 285 is the pick for single-thread-dominated workflows. It wins the single-thread test by 7.2%, has a higher boost clock, and draws far less power on paper at 65 watts versus 350 watts. It also includes integrated graphics, which the Threadripper lacks. For a desktop user who needs strong single-core performance and does not need 64 threads, it is the sensible part.

The percentile ranking summarizes the tier difference: the Threadripper sits at the 99th percentile of all CPUs, the Core Ultra 9 at the 95th. The average benchmark score gap is 279778 versus 75488. The Threadripper is not just faster; it competes in a different performance class, among server and workstation processors. The Core Ultra 9 competes in the high-end desktop class. Choose based on whether the workload scales with cores, because the benchmark results are unambiguous about which part wins where.