AMD Ryzen Threadripper 9980X vs AMD Ryzen Threadripper PRO 9985WX Comparison

The AMD Ryzen Threadripper 9980X and AMD Ryzen Threadripper PRO 9985WX are two 64-core, 128-thread Zen 5 processors built on the same Shimada Peak design, yet the benchmark data separates them into distinct personalities. Both chips share identical core counts, clock speeds, and cache hierarchies, but their platform positioning—one as a Desktop part, the other as a Server/Workstation part—creates measurable differences in memory bandwidth, PCIe lanes, and workload-specific performance. The data shows the PRO 9985WX winning 11 of 17 head-to-head benchmarks, while the 9980X takes 6, but the margins and the nature of those wins tell a more nuanced story than a simple tally.

Where Each One Wins

The PRO 9985WX dominates the Cinebench suite, which stresses sustained multi-threaded rendering and single-core responsiveness. Across all six Cinebench tests—R15, R20, and R23, both single-core and multi-core—the PRO chip wins every time with consistent deltas between -1.7% and -1.8%. This is a narrow but uniform margin, indicating that the PRO 9985WX’s eight-channel memory bus provides a slight edge in memory-hungry rendering workloads, even at identical core counts and clocks. The 9980X, by contrast, never wins a single Cinebench test.

The 9980X fights back in PassMark’s computational tests. Its most decisive victory comes in extended instructions, where it beats the PRO 9985WX by 23.5% (252,783 vs 204,637). Data compression also swings heavily toward the 9980X, with a 12.4% advantage (3,204,633 vs 2,851,752). Floating-point math favors the 9980X by 10.1% (592,543 vs 538,413), and integer math shows a smaller but still positive 1.4% edge (900,760 vs 888,032). The single-thread PassMark test is essentially a tie, with the 9980X ahead by 0.1% (4,591 vs 4,586).

The PRO 9985WX’s wins in the remaining PassMark tests are mostly narrow, but one stands out dramatically: physics. The PRO chip scores 19,284 versus the 9980X’s 9,793, a delta of -49.2% from the 9980X’s perspective—meaning the PRO is roughly twice as fast. This is the largest single margin in the entire comparison and suggests a fundamental advantage in physics simulation that cannot be explained by core count or clock speed alone. The PRO also wins find prime numbers by a wide margin (1,196 vs 837, a -30% delta), data encryption by 5.4% (169,124 vs 159,952), and multithread by 1.8% (156,305 vs 153,564). Random string sorting and the remaining Cinebench tests are all PRO wins with deltas under 2%.

Architecture Differences

Both processors are built on the same 4nm TSMC process with 66,520 million transistors across 8x 70.6 mm² dies. They share the Zen 5 architecture, the Shimada Peak codename, and the AMD Socket sTR5. L1 cache is 64 KB per core, L2 is 1 MB per core, and L3 is 256 MB shared. Both support DDR5 memory with ECC enabled and have unlocked multipliers.

The critical architectural split lies in the memory controller and PCIe implementation. The 9980X is configured with a quad-channel memory bus, delivering 204.8 GB/s of bandwidth. The PRO 9985WX, on the other hand, uses an eight-channel memory bus, doubling the bandwidth to 409.6 GB/s. This is the single largest specification difference between the two chips, and it likely explains the PRO’s consistent wins in Cinebench multithreaded tests, which are notoriously memory-bandwidth-sensitive.

PCIe connectivity also differs. The 9980X offers Gen 5 with 80 lanes (CPU only), while the PRO 9985WX doubles that to Gen 5 with 128 lanes. For workloads that rely on many GPUs, NVMe drives, or high-speed networking cards, the PRO’s additional lanes provide expansion headroom the 9980X cannot match. The market segment classification reinforces this—the 9980X is a Desktop part, while the PRO 9985WX is aimed at Server/Workstation environments.

Release dates differ by one week, with the PRO 9985WX launching earlier. The 9980X has a launch MSRP of $4999, while the PRO 9985WX has a launch MSRP of $7999. Both are active production parts, and both rank in the 100th percentile versus all CPUs.

Head-to-Head Benchmarks

The Cinebench results are remarkably consistent. In Cinebench R15 multicore, the PRO 9985WX scores 13,392 against the 9980X’s 13,157, a -1.8% delta. Single-core R15 shows the same pattern: 1,890 vs 1,857, a -1.7% delta. R20 multicore delivers 55,800 vs 54,822 (-1.8%), and R20 single-core 7,877 vs 7,739 (-1.8%). R23 multicore follows with 132,859 vs 130,529 (-1.8%), and R23 single-core 18,756 vs 18,427 (-1.8%). The uniformity of these margins—nearly every one at -1.8%—suggests a systematic advantage from the PRO’s doubled memory bandwidth rather than any per-core performance difference.

The PassMark results break the pattern. Data compression is a clear 9980X win: 3,204,633 vs 2,851,752, a 12.4% delta. Extended instructions show an even larger gap: 252,783 vs 204,637, a 23.5% delta. Floating-point math favors the 9980X at 592,543 vs 538,413 (10.1%), and integer math is close but positive at 900,760 vs 888,032 (1.4%). The single-thread test is a near dead heat, with the 9980X at 4,591 and the PRO at 4,586.

The PRO 9985WX’s PassMark wins include several narrow margins and one outlier. Physics is the outlier: 19,284 vs 9,793, a -49.2% delta. Find prime numbers shows a -30% delta (1,196 vs 837), data encryption a -5.4% delta (169,124 vs 159,952), multithread a -1.8% delta (156,305 vs 153,564), and random string sorting a -1.6% delta (352,614 vs 347,099). The PRO’s wins in physics and prime number calculation are far larger than its Cinebench margins, indicating that specific algorithmic patterns—likely involving memory access patterns or instruction-level optimizations—benefit disproportionately from the PRO’s platform.

Specification Differences

The two chips share most core specifications, so the differences are concentrated in platform-level features. The memory bus is the clearest divider: the 9980X uses quad-channel memory with 204.8 GB/s bandwidth, while the PRO 9985WX uses eight-channel memory with 409.6 GB/s. PCIe lanes also differ: 80 lanes on the 9980X versus 128 lanes on the PRO 9985WX, both Gen 5. The market segment is Desktop for the 9980X and Server/Workstation for the PRO 9985WX. The part numbers differ (100-000001593 for the 9980X, 100-000000722 for the PRO), and the release dates are one week apart, with the PRO launching first. The launch MSRP differs as well: $4999 for the 9980X and $7999 for the PRO 9985WX.

Everything else—cores, threads, base clock of 3.20, boost clock of 5.40, TDP of 350, socket, architecture, process node, transistor count, die size, cache hierarchy, ECC support, and unlocked multiplier—is identical between the two.

FAQ

Q: Which processor has more memory bandwidth?

A: The AMD Ryzen Threadripper PRO 9985WX has an eight-channel memory bus delivering 409.6 GB/s, while the AMD Ryzen Threadripper 9980X has a quad-channel bus at 204.8 GB/s.

Q: Which chip wins the Cinebench R23 multicore benchmark?

A: The PRO 9985WX wins with a score of 132,859 versus the 9980X’s 130,529, a delta of -1.8%.

Q: Does the 9980X win any benchmark by a large margin?

A: Yes, the 9980X wins PassMark extended instructions by 23.5% (252,783 vs 204,637) and data compression by 12.4% (3,204,633 vs 2,851,752).

Q: What is the largest single benchmark margin in the comparison?

A: PassMark physics, where the PRO 9985WX scores 19,284 against the 9980X’s 9,793, a -49.2% delta.

Q: Do both processors have the same number of cores and threads?

A: Yes, both have 64 cores and 128 threads, with identical base and boost clocks of 3.20 GHz and 5.40 GHz.

Q: What is the PCIe lane difference between the two?

A: The 9980X offers Gen 5 with 80 lanes (CPU only), while the PRO 9985WX offers Gen 5 with 128 lanes (CPU only).

The Verdict

The benchmark data indicates that the AMD Ryzen Threadripper PRO 9985WX is the better choice for rendering workloads, physics simulation, and prime number calculations. Its Cinebench sweep—winning all six tests with consistent -1.8% margins—and its dominant physics performance (19,284 vs 9,793) point to a platform advantage that matters in sustained, memory-intensive tasks. The eight-channel memory bus and 128 PCIe lanes reinforce this positioning for workstation and server environments.

The AMD Ryzen Threadripper 9980X is the better pick for data compression, extended instruction workloads, and floating-point math. Its 23.5% lead in extended instructions and 12.4% lead in data compression are substantial, and its 10.1% advantage in floating-point math shows strength in scientific and analytical computations. The 9980X also matches the PRO in single-thread performance (4,591 vs 4,586), meaning day-to-day responsiveness is essentially equivalent.

The decision hinges on workload type. For users prioritizing rendering, physics, or memory-bandwidth-hungry tasks, the PRO 9985WX’s data justifies its higher launch MSRP of $7999. For those focused on compression, encryption-adjacent instruction workloads, or floating-point-heavy analysis, the 9980X at a launch MSRP of $4999 provides competitive performance with a significant price gap, though the data does not comment on value. The PRO wins more benchmarks (11 vs 6), but the 9980X wins the ones where it counts by larger margins. Both sit at the 100th percentile versus all CPUs, so neither is a weak choice—the data simply shows two different strengths from the same silicon.