Intel Core Ultra 9 290K Plus vs Intel Xeon 6520P Comparison

Head-to-Head Benchmarks

The recorded data shows a clear pattern across the seventeen head-to-head tests between the Intel Xeon 6520P and the Intel Core Ultra 9 290K Plus. The Xeon wins thirteen of these tests, while the Ultra 9 wins four. The margins, however, tell a more nuanced story than the raw win count.

The Xeon 6520P delivers its most dominant victory in PassMark physics, scoring 7209 against the Ultra 9's 3315, a delta of 117.5%. This is the largest single gap in the entire comparison. The PassMark integer math test also heavily favors the Xeon, with a score of 214288 versus 166194, a 28.9% advantage. Extended instructions see the Xeon ahead by 23.3%, scoring 64557 against 52338, and data compression favors the Xeon by 20.5% (841518 versus 698346). Random string sorting also goes to the Xeon with a 20.1% margin (95736 versus 79744). Prime number finding is closer, with the Xeon leading 526 to 503, a 4.6% edge.

The Cinebench suite is remarkably consistent. Across all six tests, the Xeon wins by exactly 3.4%. This includes Cinebench R15 multicore (5392 vs 5214), R15 singlecore (761 vs 736), R20 multicore (22467 vs 21727), R20 singlecore (3171 vs 3067), R23 multicore (53495 vs 51731), and R23 singlecore (7552 vs 7303). The PassMark multithread test also shows a 3.4% delta in the Xeon's favor (62936 versus 60860). This uniformity suggests a systematic advantage rather than workload-specific behavior.

The Ultra 9 290K Plus claims its biggest win in PassMark single-thread performance, scoring 4823 against the Xeon's 3356, a 30.4% advantage. This is the second-largest delta in the entire comparison, behind only the Xeon's physics win. The Ultra 9 also wins PassMark floating-point math by 24.2%, scoring 214760 versus 162862. Data encryption goes to the Ultra 9 as well, 52563 against 45188, a 14% margin. These four wins are concentrated in areas that favor high clock speeds and per-core efficiency rather than raw thread count.

Architecture Differences

The two processors diverge sharply in their underlying designs. The Xeon 6520P uses the Granite Rapids architecture on a 5 nm process node fabricated by Intel, with a die size of 598 mm². The Core Ultra 9 290K Plus uses the Arrow Lake Refresh architecture on a 3 nm node fabricated by TSMC, with a die size of 243 mm². The Ultra 9's die is substantially smaller, and the process node difference explains part of its efficiency profile.

Both chips have 24 cores, but the thread counts differ significantly. The Xeon supports 48 threads via simultaneous multithreading, while the Ultra 9 supports only 24 threads. This means the Xeon can process twice as many threads concurrently. Base clocks differ as well: the Xeon runs at 2.40 GHz, while the Ultra 9 starts at 3.70 GHz. Boost clocks show an even larger gap, with the Xeon reaching 4.00 GHz and the Ultra 9 reaching 5.80 GHz.

Cache layouts are fundamentally different. The Xeon has 112 KB of L1 per core, 2 MB of L2 per core, and a massive 144 MB of shared L3 cache. The Ultra 9 has 192 KB of L1 per core, 3 MB of L2 per core, and only 36 MB of shared L3. The Xeon's L3 cache is four times larger, which helps explain its strong performance in data-heavy workloads. The Ultra 9's larger per-core L1 and L2 caches contribute to its single-thread advantage.

Memory architecture further separates the two. The Xeon uses an eight-channel DDR5 memory bus with 409.6 GB/s of bandwidth. The Ultra 9 uses a dual-channel DDR5 bus with 115.2 GB/s of bandwidth. The Xeon's memory bandwidth is roughly 3.5 times higher, which matters for multi-threaded workloads that saturate memory. Both support ECC memory, but the Xeon's server pedigree is evident in its PCIe lane count: 88 Gen 5 lanes on the CPU, versus 20 Gen 5 lanes for the Ultra 9.

The sockets are incompatible. The Xeon uses Intel Socket 4710, while the Ultra 9 uses Intel Socket 1851. The Xeon has no integrated graphics, whereas the Ultra 9 includes Arc Xe-LPG Graphics 64EU. The Xeon's TDP is 210 watts, while the Ultra 9's TDP is 125 watts. The Xeon is not multiplier-unlocked, but the Ultra 9 is. The Xeon launched on 2025-02-23 with a launch MSRP of $1295; the Ultra 9 has no recorded launch date or MSRP in the database.

FAQ

Q: Which processor has more threads?

A: The Intel Xeon 6520P has 48 threads, while the Intel Core Ultra 9 290K Plus has 24 threads. Both have 24 cores, but the Xeon supports simultaneous multithreading.

Q: How much L3 cache does each CPU have?

A: The Xeon 6520P has 144 MB of shared L3 cache. The Core Ultra 9 290K Plus has 36 MB of shared L3 cache.

Q: What is the memory bandwidth difference?

A: The Xeon 6520P has an eight-channel DDR5 bus with 409.6 GB/s bandwidth. The Core Ultra 9 290K Plus has a dual-channel DDR5 bus with 115.2 GB/s bandwidth.

Q: Which CPU wins in single-threaded PassMark tests?

A: The Core Ultra 9 290K Plus wins PassMark single-thread with a score of 4823, versus 3356 for the Xeon, a 30.4% advantage. The same margin appears in the single-thread duplicate test.

Q: Which CPU wins in Cinebench R23 multicore?

A: The Xeon 6520P scores 53495 in Cinebench R23 multicore, versus 51731 for the Ultra 9, a 3.4% margin. The Xeon wins all six Cinebench tests by exactly 3.4%.

Q: What process nodes do the two chips use?

A: The Xeon 6520P uses a 5 nm node from Intel. The Core Ultra 9 290K Plus uses a 3 nm node from TSMC.

Specification Differences

The two processors differ in nearly every specification category. The Xeon 6520P has 24 cores and 48 threads, while the Ultra 9 290K Plus has 24 cores and 24 threads. The Xeon's base clock is 2.40 GHz, and its boost clock is 4.00 GHz. The Ultra 9's base clock is 3.70 GHz, and its boost clock is 5.80 GHz. The Xeon's TDP is 210 watts, while the Ultra 9's TDP is 125 watts.

The Xeon uses Intel Socket 4710 with a 598 mm² die on Intel's 5 nm process. The Ultra 9 uses Intel Socket 1851 with a 243 mm² die on TSMC's 3 nm process. The Ultra 9 is fabricated by TSMC, while the Xeon is fabricated by Intel. The Ultra 9 has 17,800 million transistors; the Xeon has no recorded transistor count.

Cache sizes differ across all levels. The Xeon has 112 KB of L1 per core, 2 MB of L2 per core, and 144 MB of shared L3. The Ultra 9 has 192 KB of L1 per core, 3 MB of L2 per core, and 36 MB of shared L3. Memory support is DDR5 for both, but the bus widths differ: eight-channel for the Xeon, dual-channel for the Ultra 9. Memory bandwidth is 409.6 GB/s for the Xeon and 115.2 GB/s for the Ultra 9.

PCIe lanes differ substantially. The Xeon has 88 Gen 5 lanes from the CPU, while the Ultra 9 has 20 Gen 5 lanes. The Xeon has no integrated graphics; the Ultra 9 has Arc Xe-LPG Graphics 64EU. The Xeon is not multiplier-unlocked; the Ultra 9 is multiplier-unlocked. The Xeon's market segment is Server/Workstation, while the Ultra 9 is Desktop. The Xeon's part number is SRVNQ; the Ultra 9's part number is unknown.

Where Each One Wins

The Xeon 6520P wins in workloads that scale with thread count, memory bandwidth, and large cache capacity. Its 48 threads give it a natural advantage in parallel processing. The 117.5% physics win indicates strong multi-threaded compute capability. The 28.9% integer math advantage and 23.3% extended instructions win point to heavy computational tasks. Data compression and random string sorting, both with margins above 20%, suggest the Xeon excels in data manipulation and throughput-oriented scenarios. The 144 MB L3 cache and 409.6 GB/s memory bandwidth make it well-suited for large datasets that need to stay close to the compute cores.

The Core Ultra 9 290K Plus wins in single-threaded and latency-sensitive workloads. Its 30.4% PassMark single-thread advantage is substantial. The 24.2% floating-point math win suggests strength in numerically intensive operations that rely on per-core performance. The 14% data encryption advantage indicates efficiency in cryptographic workloads. The higher boost clock of 5.80 GHz and the larger per-core L1 and L2 caches explain these wins. The 3 nm TSMC process also contributes to lower TDP, which may be relevant in thermally constrained environments.

The Cinebench results show the Xeon wins all six tests by exactly 3.4%, including single-core tests. This is notable because the Ultra 9 has a much higher boost clock. The Xeon's consistent edge in Cinebench suggests architectural efficiency that offsets its lower clock speeds. The PassMark multithread test also goes to the Xeon by 3.4%, matching the Cinebench pattern.

The Verdict

The data supports a clear split based on workload type and system context. The Intel Xeon 6520P is the stronger choice for server and workstation environments where multi-threaded throughput, memory bandwidth, and large cache capacity are paramount. Its 48 threads, 144 MB L3 cache, and eight-channel memory architecture provide measurable advantages across most of the recorded tests. The 13 wins out of 17 head-to-head benchmarks, including the dominant physics and integer math results, confirm this direction. The 96th percentile ranking among all CPUs places it among the top tier of recorded processors.

The Intel Core Ultra 9 290K Plus is the better fit for desktop workloads that prioritize single-thread responsiveness and per-core efficiency. Its 30.4% single-thread advantage and 24.2% floating-point win are significant. The 125 watt TDP and unlocked multiplier may appeal to users who want a more power-efficient, overclockable desktop part. The built-in Arc graphics and 3 nm process node differentiate it from the server-focused Xeon. Its 96th percentile ranking matches the Xeon, but the average benchmark score of 84003 is lower than the Xeon's 93786.

Users who need maximum multi-threaded compute, large memory footprints, or extensive PCIe connectivity should select the Xeon 6520P. Users who prioritize single-thread speed, floating-point performance, or a desktop platform with integrated graphics should select the Core Ultra 9 290K Plus. The Xeon's launch MSRP of $1295 provides a reference point for its positioning, while the Ultra 9 has no recorded MSRP in the database. Neither processor is universally superior; the recorded data shows complementary strengths.