AMD Ryzen Threadripper 9970X vs AMD Ryzen Threadripper PRO 9995WX Comparison

FAQ

Q: Which processor has more cores and threads?

A: The AMD Ryzen Threadripper PRO 9995WX has 96 cores and 192 threads, while the AMD Ryzen Threadripper 9970X has 32 cores and 64 threads. Both are unlocked and built on the Zen 5 architecture.

Q: What is the difference in clock speeds between the two?

A: Both share the same 5.40 GHz boost clock. The base clock differs: the 9970X runs at 4.00 GHz, while the 9995WX runs at 2.50 GHz.

Q: How do their average benchmark scores compare?

A: The 9995WX has an average benchmark score of 412068, placing it in the 100th percentile of all CPUs. The 9970X has an average score of 279778, placing it in the 99th percentile. The 9995WX scores about 47% higher on average.

Q: What are the closest rivals for each processor according to the database?

A: For the 9995WX, the nearest rival is the AMD EPYC 9745 (avg score 425973, 3.3% higher), followed by the AMD EPYC 9655P (397773, 3.6% lower). For the 9970X, the nearest rival is the Intel Xeon 6780E (280438, 0.2% lower), followed by the AMD EPYC 9565 (285471, 2% higher).

Q: Which processor has more L3 cache?

A: The 9995WX has 384 MB of L3 cache, while the 9970X has 128 MB. Both have 64 KB of L1 and 1 MB of L2 per core.

Q: Do both processors support ECC memory?

A: Yes, both support ECC memory and DDR5 memory. However, the 9995WX uses an eight-channel memory bus with 409.6 GB/s bandwidth, while the 9970X uses a quad-channel bus with 204.8 GB/s bandwidth.

The Verdict

The data is unambiguous. The AMD Ryzen Threadripper PRO 9995WX wins all 11 head-to-head benchmark comparisons against the AMD Ryzen Threadripper 9970X. There is no recorded category where the 9970X takes the lead.

The 9995WX is the choice for workloads that demand maximum parallel throughput. Its 96 cores and 192 threads, combined with 384 MB of L3 cache and an eight-channel memory bus, make it the clear pick for heavy server or workstation rendering, simulation, encryption, and data compression tasks. The database places it in the 100th percentile of all CPUs, and its average benchmark score of 412068 reflects that position.

The 9970X is the choice for users who need the same Zen 5 architecture and the same 5.40 GHz boost clock but with fewer cores, lower memory bandwidth, and a smaller cache footprint. Its average score of 279778 still lands in the 99th percentile, so it remains a high-end desktop processor. But the recorded data shows it trails the 9995WX by margins ranging from a negligible 0.3% in single-threaded tests to a massive 158.6% in integer math.

Given that both share the same TDP of 350 W and the same socket, the decision comes down to core count and memory infrastructure. The 9995WX is the dominant part for multi-threaded, memory-intensive workloads. The 9970X is the more modest option for single-socket desktop systems that do not require the full 128 PCIe Gen 5 lanes or eight-channel memory.

Head-to-Head Benchmarks

The head-to-head data shows a consistent pattern of dominance. The 9995WX wins every recorded test, but the margins vary significantly by workload type.

The largest delta is in integer math. The 9995WX scores 1203634, while the 9970X scores 465378, a difference of 158.6%. This is the single biggest gap in the entire comparison and reflects the 3x difference in core count.

Data encryption shows the second-largest margin at 138.2%. The 9995WX records 206662, versus 86765 for the 9970X. Physics performance follows closely, with the 9995WX at 16860 and the 9970X at 6835, a 146.7% advantage. Floating point math also shows a large gap: 725066 against 309719, or 134.1% ahead.

Find prime numbers is another strong win for the 9995WX: 1476 versus 615, a 140% delta. Random string sorting shows 418973 versus 191445, a 118.8% advantage. Data compression comes in at 111.8% ahead, with scores of 3723652 and 1757998 respectively. Extended instructions show a 90.1% lead, with 270647 against 142342.

The multithread test is the narrowest of the multi-threaded wins, at 59.4%. The 9995WX scores 171200, while the 9970X scores 107399. This smaller margin is notable because it suggests the 9970X scales reasonably well in this specific workload despite having fewer cores.

The single-threaded tests are effectively a tie. The 9995WX scores 4542, and the 9970X scores 4530, a delta of just 0.3%. This confirms that per-core performance is nearly identical, as expected from two Zen 5 parts with the same boost clock.

Overall, the data shows that the 9995WX is roughly 1.1x to 1.6x faster in most multi-threaded workloads, while matching the 9970X in single-threaded performance.

Specification Differences

The two processors differ across nearly every major specification except for TDP, socket, and boost clock.

Cores and threads: The 9995WX has 96 cores and 192 threads. The 9970X has 32 cores and 64 threads. This is a 3x difference in core count.

Base clock: The 9995WX runs at 2.50 GHz. The 9970X runs at 4.00 GHz. The higher base clock on the 9970X partially compensates for its lower core count in lightly threaded workloads.

Transistors: The 9995WX has 99,780 million transistors. The 9970X has 33,260 million. The die size is 12x 70.6 mm² for the 9995WX and 4x 70.6 mm² for the 9970X, reflecting the different number of chiplets.

Cache: The 9995WX has 384 MB of L3 cache. The 9970X has 128 MB. Both have 64 KB of L1 and 1 MB of L2 per core.

Memory bus: The 9995WX uses eight-channel memory with 409.6 GB/s bandwidth. The 9970X uses quad-channel memory with 204.8 GB/s bandwidth. Both support DDR5 and ECC.

PCIe: The 9995WX has Gen 5 with 128 lanes (CPU only). The 9970X has Gen 5 with 80 lanes (CPU only).

Market segment: The 9995WX is classified as Server/Workstation. The 9970X is classified as Desktop.

Release date: The 9995WX was released on 2025-07-22. The 9970X was released on 2025-07-29, one week later.

Launch MSRP: The 9995WX launched at $11700. The 9970X launched at $2499.

Architecture Differences

Both processors share the same Zen 5 architecture, the Shimada Peak codename, the 4 nm process node from TSMC, and the AMD Socket sTR5. Both are unlocked and have no integrated graphics.

The fundamental architectural difference is scale. The 9995WX uses 12 chiplets, each measuring 70.6 mm², for a total of 99,780 million transistors. The 9970X uses 4 chiplets of the same size, totaling 33,260 million transistors. This means the 9995WX is essentially three 9970X-class dies combined into a single package.

The memory architecture follows the same scaling pattern. The 9995WX doubles the memory channels to eight and doubles the bandwidth to 409.6 GB/s. The PCIe lane count also increases from 80 to 128 lanes on the 9995WX. These differences are not merely quantitative; they change the class of machine each processor can support.

The L3 cache difference is also architectural in nature. The 9995WX has 384 MB, exactly three times the 128 MB of the 9970X. This aligns with the three-fold increase in chiplets and cores.

The base clock difference is the only architectural parameter where the smaller chip has an advantage. The 9970X runs at 4.00 GHz base, while the 9995WX runs at 2.50 GHz. At boost, both reach 5.40 GHz, which explains why their single-threaded scores are nearly identical (4542 versus 4530).

Neither processor features 3D V-Cache, and both have the same per-core L1 and L2 sizes. The architecture is identical; the implementation scales differently.

Where Each One Wins

The 9995WX wins every benchmark category in the head-to-head comparison, so the question is not which one wins a given test, but rather which workloads benefit most from the 9995WX and which can accept the 9970X.

The 9995WX is the clear winner for any workload that scales with core count and memory bandwidth. The largest deltas are in integer math (158.6% ahead), physics (146.7%), find prime numbers (140%), and data encryption (138.2%). These are compute-heavy, parallel workloads where the 96-core count and 384 MB of L3 cache provide a decisive advantage. The eight-channel memory bus with 409.6 GB/s bandwidth also helps in data compression (111.8% ahead) and random string sorting (118.8% ahead), both of which are memory-intensive.

The 9970X wins only in the sense of efficiency per chiplet. Its 4.00 GHz base clock is higher, and its single-threaded score of 4530 is within 0.3% of the 9995WX. For workloads that are primarily single-threaded or lightly threaded, the 9970X will feel identical to the 9995WX. The 9970X also supports the same 5.40 GHz boost clock, so any task that relies on boost clocks rather than sustained multi-core throughput will see no penalty.

For a desktop system with 80 PCIe Gen 5 lanes and quad-channel memory, the 9970X is sufficient. It occupies the 99th percentile of all CPUs and has an average benchmark score of 279778, which puts it within 2.5% of its nearest rival, the AMD EPYC 9555P.

For a server or workstation with 128 PCIe Gen 5 lanes and eight-channel memory, the 9995WX is the required part. It sits in the 100th percentile and its nearest rival, the AMD EPYC 9745, is only 3.3% ahead, so it is competitive with top-tier server parts.

In summary, the 9995WX wins all multi-threaded and memory-heavy workloads by margins of 59.4% to 158.6%. The 9970X matches it in single-threaded performance and offers the same boost clock, but the core count, cache size, and memory bandwidth differences are decisive in any parallel workload.