AMD Ryzen Threadripper 2990WX vs Intel Xeon Gold 6326 Comparison

The Intel Xeon Gold 6326 and the AMD Ryzen Threadripper 2990WX occupy different corners of the processor market, yet the recorded benchmark data places them close in raw compute output. The Xeon Gold 6326, a 16-core Ice Lake-SP server part, wins all six head-to-head Cinebench comparisons against the 32-core Zen+ Threadripper, though by margins of only about 2 percent in each test. The Threadripper counters with a much larger core count, a higher boost clock, and an unlocked multiplier, but the data shows it does not convert those advantages into a multi-threaded victory in the recorded tests. The Xeon wins every round, yet the Threadripper remains the more flexible platform for specific workflows. The verdict splits cleanly: choose the Xeon for consistent single-thread and multi-thread Cinebench leadership, choose the Threadripper when raw core count, higher clocks, and overclocking headroom matter more than the measured benchmark deltas.

The Verdict

The data supports two distinct buying decisions. For a server or workstation where the workload is heavily threaded but also sensitive to single-thread performance, the Xeon Gold 6326 is the safer pick. It wins all six recorded Cinebench tests, including both single-core and multi-core variants, with deltas ranging from 1.8 to 2 percent. The largest margin is 2 percent in the multi-core R15, R20, and R23 tests, and the single-core tests show similar 1.8 to 2 percent edges. The Xeon also posts a higher average benchmark score of 8071 against the Threadripper's 7578, a difference of roughly 6.5 percent across the recorded suite. Its percentile ranking among all CPUs is 64, one point above the Threadripper's 63. For a user running Cinebench-style rendering workloads, the Xeon is the consistent winner, but the margin is narrow enough that real-world differences may be small.

The Threadripper 2990WX, however, is not without appeal. It offers 32 cores and 64 threads, exactly double the Xeon's 16 cores and 32 threads, and it boosts to 4.20 GHz versus the Xeon's 3.50 GHz. It also has an unlocked multiplier, so users who overclock can potentially push it past its stock scores. The launch MSRP was $1799. The data shows it loses each head-to-head test by roughly 2 percent, but its platform attributes suggest it may handle other tasks better than the recorded Cinebench results indicate. For users who prioritize core count and clock speed over the specific benchmark outcomes in this database, the Threadripper is the more capable hardware on paper, even if the measured scores do not reflect that advantage in these tests.

Architecture Differences

The two processors come from different design philosophies. The Xeon Gold 6326 uses Intel's Ice Lake-SP architecture on a 10 nm process node, built by Intel. It has 16 cores and 32 threads, with a base clock of 2.90 GHz and a boost clock of 3.50 GHz. Its cache layout is per-core: 64 KB of L1 per core, 1 MB of L2 per core, and a shared 24 MB L3 cache. Memory support is DDR4 over an eight-channel bus, yielding a memory bandwidth of 204.8 GB/s. It supports ECC memory, which is a typical requirement for server and workstation reliability. The PCIe interface is Gen 4 with 64 lanes available from the CPU. The socket is Intel Socket 4189, the market segment is Server/Workstation, and it was released in April 2021. The multiplier is locked.

The Threadripper 2990WX uses AMD's Zen+ architecture, codename Colfax, on a 12 nm process node built by GlobalFoundries. It has 32 cores and 64 threads, with a base clock of 3.00 GHz and a boost clock of 4.20 GHz. It contains 19,200 million transistors across a die size of 4x 213 mm². Its cache differs significantly: 96 KB of L1 per core, 512 KB of L2 per core, and a larger shared 64 MB L3 cache. Memory support is DDR4 over a quad-channel bus, giving a memory bandwidth of 93.9 GB/s, less than half the Xeon's bandwidth. It does not support ECC memory. The PCIe interface is Gen 3 with 60 lanes. The socket is AMD Socket SP3r2, the market segment is Desktop, and it was released in August 2018. The multiplier is unlocked, and the part number is YD299XAZUIHAF.

The architectural contrast is stark: the Xeon trades core count for memory bandwidth, ECC support, and a newer process node, while the Threadripper offers double the cores, a larger L3 cache, and higher clocks on an older, larger process. The Xeon's eight-channel memory bus gives it a theoretical bandwidth of 204.8 GB/s, more than double the Threadripper's 93.9 GB/s. That bandwidth advantage likely helps the Xeon close the gap against a chip with twice as many cores. The Threadripper's 64 MB L3 cache is nearly triple the Xeon's 24 MB, which could benefit workloads with large working sets, but the recorded Cinebench results do not show that advantage materializing.

Head-to-Head Benchmarks

The head-to-head data covers six Cinebench tests, and the Xeon Gold 6326 wins every one. In Cinebench R15 multi-core, the Xeon scores 2812 against the Threadripper's 2758, a 2 percent lead. In R15 single-core, the Xeon scores 396 against 389, a 1.8 percent lead. In R20 multi-core, the Xeon scores 11719 against 11495, a 1.9 percent lead. In R20 single-core, the Xeon scores 1654 against 1622, a 2 percent lead. In R23 multi-core, the Xeon scores 27904 against 27370, a 2 percent lead. In R23 single-core, the Xeon scores 3939 against 3864, a 1.9 percent lead.

The pattern is consistent: the Xeon wins by roughly 2 percent in every test, regardless of whether the workload is single-threaded or multi-threaded. This is notable because the Threadripper has twice the cores and a higher boost clock. The fact that the Xeon still leads in multi-core tests suggests that memory bandwidth, cache efficiency, or the newer architecture compensates for the core count disadvantage. The Threadripper's only additional recorded benchmarks are Geekbench scores, which are not part of the head-to-head set. In Geekbench, the Threadripper scores 11821 multi-core and 1303 single-core. Those scores are not compared directly to the Xeon in the head-to-head table, but they contribute to the Threadripper's average benchmark score of 7578, which trails the Xeon's 8071.

The wins break down as 6 for the Xeon and 0 for the Threadripper. No test in the head-to-head set favors the AMD part. The narrow margins, all under 2.1 percent, mean the Threadripper is not far behind, but the direction is uniform. Users expecting a 32-core chip to dominate a 16-core chip in multi-threaded Cinebench will be surprised by the data. The Xeon's eight-channel memory bandwidth and higher per-core cache allocation appear to be decisive factors in these tests.

FAQ

Q: Which processor wins the most head-to-head benchmarks?

A: The Intel Xeon Gold 6326 wins all six recorded head-to-head tests, with deltas of 1.8 to 2 percent. The AMD Ryzen Threadripper 2990WX wins none.

Q: How big is the performance gap in multi-core tests?

A: In Cinebench R23 multi-core, the Xeon scores 27904 against the Threadripper's 27370, a 2 percent lead. Similar 1.9 to 2 percent margins appear in R15 and R20 multi-core tests.

Q: Does the Threadripper's 32-core count give it an advantage?

A: The recorded data shows no advantage in Cinebench tests. Despite having double the cores, the Threadripper trails the Xeon by about 2 percent in every multi-core test.

Q: What about single-core performance?

A: The Xeon leads in all three single-core tests: R15 396 vs 389, R20 1654 vs 1622, and R23 3939 vs 3864. The margins are 1.8 to 2 percent.

Q: Which processor has higher memory bandwidth?

A: The Xeon Gold 6326 has an eight-channel memory bus with 204.8 GB/s bandwidth, while the Threadripper 2990WX has a quad-channel bus with 93.9 GB/s.

Q: Does the Threadripper support ECC memory?

A: No, the Threadripper 2990WX does not support ECC memory. The Xeon Gold 6326 does support ECC.

Where Each One Wins

The Xeon Gold 6326 wins in every recorded benchmark category. Its strengths are single-thread performance, multi-thread Cinebench scores, memory bandwidth, and ECC support. The data shows it leads by 2 percent in R15 and R20 multi-core, by 1.9 percent in R20 single-core, and by 2 percent in R23 multi-core. Its average benchmark score of 8071 is higher than the Threadripper's 7578, and its percentile ranking of 64 is one point higher. For users running Cinebench-style rendering, video encoding, or other workloads that resemble the recorded tests, the Xeon is the better performer. Its eight-channel memory bus at 204.8 GB/s also makes it a stronger candidate for memory-bandwidth-sensitive tasks like large-scale data processing or virtualization, though those specific workloads are not in the recorded data.

The Threadripper 2990WX wins on hardware attributes, not benchmark scores. It has 32 cores and 64 threads, double the Xeon's counts. Its boost clock of 4.20 GHz is 0.70 GHz higher than the Xeon's 3.50 GHz. Its base clock of 3.00 GHz is also higher than the Xeon's 2.90 GHz. It has a 64 MB L3 cache, nearly triple the Xeon's 24 MB. It has an unlocked multiplier, allowing overclocking beyond stock settings, which the locked Xeon cannot match. Its market segment is Desktop, so it may fit into consumer or prosumer builds where the Xeon's Server/Workstation socket would not. The Threadripper also has a 250 W TDP versus the Xeon's 185 W, indicating higher power draw, but with the unlocked multiplier, users can trade power for clock speed. For workloads not well represented by Cinebench, such as those that scale with core count and cache size, the Threadripper's architecture may be the better fit, even though the recorded data does not show it winning any test.

Specification Differences

The two processors differ in nearly every major specification. The Xeon has 16 cores and 32 threads, while the Threadripper has 32 cores and 64 threads. The Xeon's base clock is 2.90 GHz, the Threadripper's is 3.00 GHz. The Xeon boosts to 3.50 GHz, the Threadripper to 4.20 GHz. TDP differs: 185 W for the Xeon, 250 W for the Threadripper. The sockets are different: Intel Socket 4189 versus AMD Socket SP3r2. The architecture is Ice Lake-SP on a 10 nm Intel process, versus Zen+ (Colfax) on a 12 nm GlobalFoundries process. The Threadripper lists 19,200 million transistors and a die size of 4x 213 mm², while the Xeon has no transistor or die size data recorded.

Cache configurations diverge significantly. The Xeon has 64 KB of L1 per core, 1 MB of L2 per core, and 24 MB of shared L3. The Threadripper has 96 KB of L1 per core, 512 KB of L2 per core, and 64 MB of shared L3. Memory support is DDR4 for both, but the bus differs: eight-channel for the Xeon, quad-channel for the Threadripper. Memory bandwidth is 204.8 GB/s for the Xeon and 93.9 GB/s for the Threadripper. ECC memory is supported on the Xeon but not the Threadripper. PCIe versions and lane counts differ: Gen 4 with 64 lanes for the Xeon, Gen 3 with 60 lanes for the Threadripper. The Xeon has an integrated graphics field that is null, as does the Threadripper. The Xeon's multiplier is locked, the Threadripper's is unlocked. The market segments differ: Server/Workstation for the Xeon, Desktop for the Threadripper. The release dates are 2021-04-05 for the Xeon and 2018-08-12 for the Threadripper. The Threadripper has a launch MSRP of $1799, while the Xeon has no recorded launch MSRP. The average benchmark score is 8071 for the Xeon and 7578 for the Threadripper, and the percentile rankings are 64 and 63, respectively. The Threadripper's part number is YD299XAZUIHAF; the Xeon has no recorded part number.