AMD Ryzen Threadripper 2970WX vs Intel Xeon Gold 5317 Comparison

The AMD Ryzen Threadripper 2970WX and Intel Xeon Gold 5317 represent two very different philosophies for high-core-count computing. The 2970WX is a 24-core desktop part built on AMD’s Zen+ architecture, while the 5317 is a 12-core server processor on Intel’s Ice Lake-SP design. Both land at the 62nd percentile among all CPUs, and their average benchmark scores are nearly identical — 6760 for the AMD part versus 6710 for the Intel part, a 0.7% gap. However, the head-to-head benchmark data tells a one-sided story: the Threadripper wins all six tests by a consistent margin of roughly 13.4%. This comparison hinges on core count, memory architecture, and platform features rather than any subtle performance parity.

FAQ

Q: Which processor has more cores and threads?

A: The AMD Ryzen Threadripper 2970WX has 24 cores and 48 threads. The Intel Xeon Gold 5317 has 12 cores and 24 threads, exactly half of each.

Q: How do their single-core benchmark scores compare?

A: The Threadripper wins every single-core test. In Cinebench R23 single-core, it scores 3713 versus the Xeon’s 3275, a 13.4% advantage. The same delta appears in Cinebench R15 single-core (374 vs 330).

Q: What are the key architectural differences in memory support?

A: The Threadripper uses a quad-channel DDR4 memory bus with 93.9 GB/s bandwidth. The Xeon Gold 5317 uses an eight-channel DDR4 bus with 187.7 GB/s bandwidth — exactly double the bandwidth of the AMD part.

Q: Do both processors support ECC memory?

A: No. The Intel Xeon Gold 5317 supports ECC memory, while the AMD Ryzen Threadripper 2970WX does not. This is a critical distinction for error-sensitive workloads.

Q: Which processor has a higher boost clock?

A: The AMD part boosts to 4.20 GHz, while the Intel part boosts to 3.60 GHz. Both have the same 3.00 GHz base clock.

Q: How do their PCIe lanes differ?

A: The Threadripper provides 60 PCIe Gen 3 lanes from the CPU. The Xeon Gold 5317 provides 64 PCIe Gen 4 lanes from the CPU, offering both more lanes and a newer generation.

Architecture Differences

The two CPUs are built on fundamentally different foundations. The Threadripper 2970WX uses AMD’s Zen+ architecture on a 12 nm process from GlobalFoundries, with a die size of 4x 213 mm² and a total of 19,200 million transistors. It is part of the Colfax codename and the 2000 series, designed for the desktop market with an AMD Socket SP3r2. The Xeon Gold 5317 uses Intel’s Ice Lake architecture on a 10 nm process, codenamed Ice Lake-SP, and fits into Intel Socket 4189. It targets the server/workstation segment.

Cache layouts differ substantially. The Threadripper allocates 96 KB of L1 cache per core and 512 KB of L2 per core, with a large 64 MB L3 cache. The Xeon uses 64 KB of L1 per core and a much larger 1 MB of L2 per core, but its L3 is only 18 MB and shared. This means the AMD part has over three times the total L3 cache, which can benefit working sets that fit in cache. The Intel part compensates with double the memory bandwidth (187.7 GB/s versus 93.9 GB/s) thanks to its eight-channel memory controller.

The Threadripper’s 24 cores are unlocked for overclocking (multiplier unlocked: true), while the Xeon’s 12 cores are locked (multiplier unlocked: false). The AMD part also has a higher TDP at 250 W versus 150 W for the Intel part, reflecting its larger core count and higher boost clocks. Neither processor includes integrated graphics. The release dates are far apart: the Threadripper launched on 2018-10-01 with a launch MSRP of $1299, while the Xeon launched on 2021-04-05 with no listed MSRP.

Head-to-Head Benchmarks

The head-to-head results are unambiguous: the AMD Ryzen Threadripper 2970WX wins all six benchmark tests, and each win comes at exactly 13.4% or 13.3% — a remarkably consistent margin. Starting with Cinebench R15 multicore, the Threadripper scores 2651 against the Xeon’s 2338, a 13.4% lead. The single-core version of that test shows 374 versus 330, also a 13.3% advantage. Moving to Cinebench R20, the multicore result is 11048 for AMD versus 9743 for Intel (13.4% delta), and single-core is 1559 versus 1375 (13.4%). Cinebench R23 multicore follows the same pattern: 26305 versus 23199, a 13.4% win. The single-core R23 result is 3713 versus 3275, again 13.4%.

The consistency of the delta across all tests suggests the advantage is structural rather than workload-specific. The Threadripper’s 24 cores versus 12 cores gives it a clear edge in multicore throughput, but its single-core wins are more surprising given the Xeon’s newer 10 nm process. The Threadripper’s higher boost clock of 4.20 GHz versus 3.60 GHz likely explains the single-core superiority. Notably, the Threadripper also has a Geekbench multicore score of 7195 and a single-core score of 1236, while the Xeon lacks Geekbench results entirely.

The average benchmark scores paint a closer picture. The Threadripper averages 6760, while the Xeon averages 6710. This is a 0.7% difference, and the nearest rival data shows the Xeon Gold 6154 (avg 6803) and Core i7-12700E (avg 6776) sitting just above the Threadripper, while the Xeon Gold 5317 sits just below. The deltaPct values in the rival lists confirm the near-tie: the Threadripper is -0.2% versus the i7-12700E and +0.7% versus the Xeon Gold 5317, while the Xeon is -0.7% versus the Threadripper.

Specification Differences

The specification sheets reveal where these CPUs diverge on paper. Core and thread counts are the most obvious: 24 cores and 48 threads for AMD versus 12 cores and 24 threads for Intel. Base clocks match at 3.00 GHz, but boost clocks differ — 4.20 GHz for the Threadripper versus 3.60 GHz for the Xeon. TDP is another major split: 250 W for AMD versus 150 W for Intel, a 100 W difference that has cooling and power-supply implications.

Memory support shows the Xeon’s server pedigree. The Threadripper has a quad-channel DDR4 bus with 93.9 GB/s bandwidth and no ECC support. The Xeon has an eight-channel DDR4 bus with 187.7 GB/s bandwidth and ECC support. That is double the memory bandwidth and a feature that matters for data integrity in long-running server workloads. PCIe support also favors Intel: the Xeon has 64 lanes of Gen 4, while the Threadripper has 60 lanes of Gen 3. Newer PCIe Gen 4 doubles the per-lane bandwidth, so the Xeon’s I/O capability is substantially higher despite the similar lane count.

Cache organization differs significantly. The Threadripper has 96 KB L1 per core, 512 KB L2 per core, and 64 MB of L3. The Xeon has 64 KB L1 per core, 1 MB L2 per core, and 18 MB of shared L3. The Threadripper’s total cache is larger, but the Xeon’s per-core L2 is double. Process node and foundry also differ: 12 nm GlobalFoundries for AMD versus 10 nm Intel for Intel. The Threadripper is unlocked for overclocking, the Xeon is not. The Threadripper has a launch MSRP of $1299 and a part number (YD297XAZUHCAF), while the Xeon has neither a launch MSRP nor a listed part number.

The Verdict

The data is clear: the AMD Ryzen Threadripper 2970WX outperforms the Intel Xeon Gold 5317 in every benchmark test included here. The Threadripper wins all six head-to-head tests with a consistent 13.4% margin, including both multicore and single-core workloads. Its 24 cores versus 12 cores and higher boost clock (4.20 GHz versus 3.60 GHz) give it a decisive advantage in raw compute. The average benchmark scores confirm this is not a fluke: 6760 versus 6710, a 0.7% edge for AMD.

However, the Xeon Gold 5317 is not without merit. Its eight-channel memory bus delivers 187.7 GB/s, exactly double the Threadripper’s 93.9 GB/s, and it supports ECC memory. Its 64 PCIe Gen 4 lanes outclass the Threadripper’s 60 Gen 3 lanes. It also draws 100 W less power (150 W versus 250 W), making it easier to cool and potentially more economical in dense server environments. The Xeon’s 10 nm process versus 12 nm also suggests better architectural efficiency per core, though this does not translate into benchmark wins.

For a pure performance-per-core and performance-per-watt perspective, the Xeon looks better on paper, but the benchmark results speak differently. The Threadripper’s massive core-count advantage and higher boost clock overcome the Xeon’s architectural improvements. The verdict is straightforward: pick the Threadripper if you need maximum compute throughput in applications that scale across cores. Pick the Xeon if you need ECC memory, higher memory bandwidth, PCIe Gen 4, or lower power draw in a server context.

Where Each One Wins

AMD Ryzen Threadripper 2970WX wins in: All CPU-bound benchmarks. It takes every Cinebench test — R15, R20, and R23 — in both multicore and single-core variants, with margins ranging from 13.3% to 13.4%. This makes it the better choice for rendering, video encoding, software compilation, and any workload that leverages many threads. Its 64 MB L3 cache provides ample capacity for large datasets. The unlocked multiplier allows overclocking, potentially extending its lead further. Its desktop market segment and 2018 release date position it as a workstation-class part for content creation.

Intel Xeon Gold 5317 wins in: Memory bandwidth and I/O throughput. Its eight-channel DDR4 bus delivers 187.7 GB/s, double the Threadripper’s 93.9 GB/s, which is critical for memory-intensive tasks like large database operations or in-memory analytics. ECC memory support protects against data corruption. PCIe Gen 4 with 64 lanes enables faster NVMe storage and high-bandwidth accelerators. Its 150 W TDP makes it more suitable for multi-socket servers or power-constrained environments. The 10 nm process and 18 MB shared L3 suggest better per-core efficiency, even if that does not show up in these benchmarks. For server workloads where data integrity, memory capacity, and I/O density matter more than raw core count, the Xeon is the appropriate pick.