Intel Core Ultra 5 235HX vs Intel Xeon Phi 7290 Comparison

Intel
INTEL

Intel Core Ultra 5 235HX

CORE STATE Arrow Lake-HX
CORE SPECS 14 Cores / 14 Threads
CLOCK SPEED 2.9 Base / 5.1 GHz Turbo
CACHE 24 MB (shared)
MAX TDP 55W
ARCHITECTURE Arrow Lake
nm
PROCESS 3 nm
LAUNCH DATE 2025
VS
Intel
INTEL

Xeon Phi 7290

CORE STATE Knights Landing
CORE SPECS 72 Cores / 288 Threads
CLOCK SPEED 1500 Base / 1700 GHz Turbo
CACHE —
MAX TDP 245W
ARCHITECTURE Knights Landing
nm
PROCESS 14 nm
LAUNCH DATE 2016

PERFORMANCE BENCHMARKS

cinebench_cinebench_r15_multicore
3,500
1,528
cinebench_cinebench_r15_singlecore
494
215
cinebench_cinebench_r20_multicore
14,587
6,368
cinebench_cinebench_r20_singlecore
2,059
898
cinebench_cinebench_r23_multicore
34,731
15,163
cinebench_cinebench_r23_singlecore
4,903
2,140
passmark_data_compression
426,417
564,535
passmark_data_encryption
32,697
12,505
passmark_extended_instructions
34,808
41,517
passmark_find_prime_numbers
361
114
passmark_floating_point_math
129,819
47,417
passmark_integer_math
97,177
126,922
passmark_multithread
40,849
17,839
passmark_physics
2,550
2,257
passmark_random_string_sorting
50,929
68,593
passmark_single_thread
4,683
485
passmark_singlethread
4,683
485

Analysis: Intel Core Ultra 5 235HX vs Intel Xeon Phi 7290

The Intel Xeon Phi 7290 and the Intel Core Ultra 5 235HX represent two very different approaches to processor design, separated by nearly a decade of architectural evolution. The Xeon Phi 7290, a Knights Landing part from 2016, is a server and workstation processor built around massive parallelism with 72 cores. The Core Ultra 5 235HX, a 2025 mobile processor on Arrow Lake-HX, is designed for high single-thread performance and efficiency. The benchmark data shows a clear split: the Core Ultra 5 dominates most workloads, while the Xeon Phi 7290 holds specific advantages in a few specialized tasks. The following analysis breaks down where each processor wins, the architectural reasons behind those results, and what the recorded scores mean for potential users.

Where Each One Wins

The benchmark results indicate a decisive overall victory for the Intel Core Ultra 5 235HX, which wins 13 of the 17 head-to-head tests. Its most significant advantages appear in single-threaded and lightly threaded workloads. In the PassMark single-thread test, the Core Ultra 5 scores 4683 compared to the Xeon Phi 7290’s 485, a delta of -89.6% in favor of the newer chip. This pattern repeats across all Cinebench single-core tests: R15 single-core (494 vs 215, -56.5%), R20 single-core (2059 vs 898, -56.4%), and R23 single-core (4903 vs 2140, -56.4%). For any application that relies on per-core performance, such as typical desktop responsiveness, gaming, or lightly threaded productivity, the Core Ultra 5 is the clear choice.

The Core Ultra 5 also wins in multi-threaded Cinebench tests, despite having only 14 cores and 14 threads versus the Xeon Phi’s 72 cores and 288 threads. In Cinebench R23 multi-core, the Core Ultra 5 scores 34731 while the Xeon Phi 7290 scores 15163, a delta of -56.3%. The same pattern appears in R15 multi-core (3500 vs 1528) and R20 multi-core (14587 vs 6368). The PassMark multithread test also favors the Core Ultra 5 (40849 vs 17839, -56.3%). This is a striking result, showing that the newer processor’s higher clock speeds and modern architecture more than compensate for its lower core count in these rendering workloads.

The Core Ultra 5 additionally wins in PassMark data encryption (32697 vs 12505, -61.8%), floating point math (129819 vs 47417, -63.5%), find prime numbers (361 vs 114, -68.4%), and physics (2550 vs 2257, -11.5%). These wins span a range of computational patterns, from cryptographic operations to scientific floating-point calculations.

The Intel Xeon Phi 7290 wins exactly four tests. Its biggest victories come in data compression (564535 vs 426417, +32.4%), random string sorting (68593 vs 50929, +34.7%), integer math (126922 vs 97177, +30.6%), and extended instructions (41517 vs 34808, +19.3%). These are workloads that can exploit the massive thread count of the Xeon Phi. Integer math and extended instruction tests often scale well with many cores, and the 72-core design provides a clear advantage here. Data compression and string sorting also benefit from the sheer number of parallel operations the Xeon Phi can sustain.

The overall average benchmark score is close: the Xeon Phi 7290 averages 53469, while the Core Ultra 5 averages 52073, a difference of less than 3%. Both processors sit in the 91st percentile of all CPUs in the database. This means that while the Core Ultra 5 wins the majority of tests, the Xeon Phi 7290’s wins are large enough to keep its average score competitive. The Xeon Phi 7290’s nearest rivals include the Intel Core i7-14700F (delta -0.3%), AMD Ryzen 9 7900X (delta +0.3%), Intel Xeon 6505P (delta -0.4%), and AMD EPYC 7313P (delta +0.5%). The Core Ultra 5’s nearest rivals are the AMD EPYC 8124P (delta -0.1%), AMD Ryzen 9 5950X (delta +0.2%), Intel Core i7-14700 (delta -0.4%), and Intel Core i9-13900F (delta +0.7%). Both processors sit in a similar performance tier, but their strengths are distributed very differently across workload types.

Architecture Differences

The architectural differences between these two processors explain nearly all of the benchmark results. The Xeon Phi 7290 uses the Knights Landing architecture, built on a 14 nm process from Intel. It has 72 cores and 288 threads, meaning each core supports four threads. Its base clock is 1500 MHz with a boost clock of 1700 MHz. The processor has a TDP of 245 watts and uses the Intel Socket 3647. The cache hierarchy consists of 32 KB of L1 cache per core and 512 KB of L2 cache per core. There is no L3 cache listed. It supports DDR4 memory with ECC enabled. The transistor count is 8,000 million. This processor is designed for server and workstation use, with a release date of June 2016.

The Core Ultra 5 235HX uses the Arrow Lake architecture, specifically Arrow Lake-HX, built on a 3 nm process from TSMC. It has 14 cores and 14 threads, with no hyperthreading. Its base clock is 2.90 GHz and its boost clock is 5.10 GHz. The TDP is 55 watts, significantly lower than the Xeon Phi. It uses the Intel BGA 2114 socket and is a mobile processor. The cache includes 192 KB of L1 per core, 3 MB of L2 per core, and 24 MB of shared L3 cache. It supports DDR5 memory through a dual-channel interface with 102.4 GB/s of memory bandwidth. ECC is not supported. The processor has a PCIe Gen 5 interface with 20 lanes (CPU only). It also includes integrated graphics: Arc Xe-LPG Graphics with 48 execution units. The transistor count is 17,800 million, and the die size is 243 mm². It was released in January 2025 and is currently marked as active. The multiplier is unlocked.

The clock speed difference is stark. The Core Ultra 5 boosts to 5100 MHz, three times higher than the Xeon Phi’s 1700 MHz boost. This single factor drives the massive single-threaded performance gap. The Xeon Phi’s reliance on four threads per core also appears to be a disadvantage in modern workloads, as the Core Ultra 5’s 14 physical cores outperform the Xeon Phi’s 72 cores in most multi-threaded tests. The manufacturing process difference (3 nm vs 14 nm) and the transistor count (17,800 million vs 8,000 million) indicate that the Core Ultra 5 packs more modern transistors into a smaller process, allowing for much higher efficiency and clock speeds.

The memory architecture also differs. The Xeon Phi uses DDR4 with ECC, which is typical for server parts. The Core Ultra 5 uses DDR5 with a dual-channel bus and a specified memory bandwidth of 102.4 GB/s. The integrated graphics on the Core Ultra 5 is a feature the Xeon Phi lacks entirely. The Core Ultra 5 also supports PCIe Gen 5, while no PCIe information is listed for the Xeon Phi.

The Verdict

Based strictly on the recorded benchmark data, the Intel Core Ultra 5 235HX is the superior processor for the vast majority of workloads. It wins every Cinebench test, both single-core and multi-core, by margins of roughly 56% in the head-to-head comparison. It also wins in data encryption, floating point math, prime number finding, physics simulation, and the PassMark multithread test. If a workload involves rendering, encryption, scientific math, or general productivity, the Core Ultra 5 is the clear pick from the data.

The Intel Xeon Phi 7290 should be chosen only for workloads that specifically match its strengths: data compression, random string sorting, integer math, and extended instruction processing. These are the four tests where it wins, and the margins are substantial (between 19.3% and 34.7%). For a user running highly parallel integer-heavy workloads that can saturate all 288 threads, the Xeon Phi 7290 still offers a measurable advantage.

The average benchmark scores are nearly identical (53469 vs 52073), and both processors hold the same 91st percentile ranking. This means neither is a universal winner in terms of overall performance. However, the Core Ultra 5’s wins are much more numerous and cover a broader range of common tasks. The Xeon Phi’s wins are narrow and specialized.

The Core Ultra 5 also comes with practical advantages that are not directly captured in the benchmark scores. It has a TDP of 55 watts versus 245 watts, making it far more power-efficient. It is a mobile processor, so it can be used in laptops. It includes integrated graphics, which the Xeon Phi does not. It supports DDR5 memory and PCIe Gen 5. These features make it a more versatile part for a typical user.

The Xeon Phi 7290, on the other hand, is a server and workstation part from 2016. Its 245 watt TDP and socket 3647 indicate a platform designed for heavy compute density, not everyday use. For a user with a workload that exactly matches its four winning test categories, it remains a viable option. For anyone else, the data points clearly to the Core Ultra 5.

FAQ

Q: Which processor has more cores?

A: The Intel Xeon Phi 7290 has 72 cores and 288 threads, while the Intel Core Ultra 5 235HX has 14 cores and 14 threads.

Q: Which processor is faster in single-threaded performance?

A: The Intel Core Ultra 5 235HX is much faster. In the PassMark single-thread test, it scores 4683 compared to the Xeon Phi 7290’s 485, a delta of -89.6% in favor of the Core Ultra 5.

Q: Does the Xeon Phi 7290 win any multi-threaded tests?

A: No. The Core Ultra 5 235HX wins all Cinebench multi-core tests, including R23 multi-core (34731 vs 15163, -56.3%) and the PassMark multithread test (40849 vs 17839, -56.3%).

Q: What workloads does the Xeon Phi 7290 excel at?

A: The Xeon Phi 7290 wins in data compression (564535 vs 426417, +32.4%), random string sorting (68593 vs 50929, +34.7%), integer math (126922 vs 97177, +30.6%), and extended instructions (41517 vs 34808, +19.3%).

Q: What is the memory support difference?

A: The Xeon Phi 7290 supports DDR4 memory with ECC. The Core Ultra 5 235HX supports DDR5 memory with a dual-channel bus and 102.4 GB/s of bandwidth, but no ECC.

Q: Which processor has integrated graphics?

A: Only the Intel Core Ultra 5 235HX has integrated graphics, specifically Arc Xe-LPG Graphics with 48 execution units. The Xeon Phi 7290 has no listed integrated graphics.

Head-to-Head Benchmarks

The largest win for the Intel Core Ultra 5 235HX comes in the PassMark single-thread test, where it scores 4683 against the Xeon Phi 7290’s 485, a delta of -89.6%. This is the most lopsided result in the comparison. The same pattern appears in the PassMark singlethread test with an identical score and delta. These results reflect the massive clock speed advantage: the Core Ultra 5 boosts to 5100 MHz, while the Xeon Phi 7290 boosts to only 1700 MHz.

In Cinebench tests, the Core Ultra 5 wins all six tests with consistent deltas around -56%. The R15 multi-core test shows 3500 vs 1528, R15 single-core shows 494 vs 215, R20 multi-core shows 14587 vs 6368, R20 single-core shows 2059 vs 898, R23 multi-core shows 34731 vs 15163, and R23 single-core shows 4903 vs 2140. The consistency of the delta across all Cinebench versions indicates a fundamental performance gap that scales uniformly across generations of the test.

The Core Ultra 5 also wins in PassMark data encryption (32697 vs 12505, -61.8%), floating point math (129819 vs 47417, -63.5%), find prime numbers (361 vs 114, -68.4%), physics (2550 vs 2257, -11.5%), and multithread (40849 vs 17839, -56.3%). The physics win is the narrowest for the Core Ultra 5, with a delta of only -11.5%, suggesting the Xeon Phi’s many cores can compete more closely in this specific test.

The Xeon Phi 7290’s largest win is in random string sorting, where it scores 68593 against 50929, a delta of +34.7%. Its other wins are data compression (564535 vs 426417, +32.4%), integer math (126922 vs 97177, +30.6%), and extended instructions (41517 vs 34808, +19.3%). These wins show that the Xeon Phi’s 288 threads can still deliver strong parallel throughput in specific integer-oriented workloads.

Overall, the head-to-head data shows 13 wins for the Core Ultra 5 and 4 wins for the Xeon Phi. The Core Ultra 5’s wins are concentrated in rendering, encryption, floating-point math, and single-threaded tasks. The Xeon Phi’s wins are in data compression, string sorting, integer math, and extended instructions. The total wins count is 13 to 4, but the average benchmark scores are close (53469 vs 52073), demonstrating that the Xeon Phi’s wins are large enough to keep its overall average competitive.

Specification Differences

The two processors differ in nearly every major specification category. The Xeon Phi 7290 has 72 cores and 288 threads, while the Core Ultra 5 235HX has 14 cores and 14 threads. The Xeon Phi has a base clock of 1500 MHz and a boost clock of 1700 MHz. The Core Ultra 5 has a base clock of 2.90 GHz and a boost clock of 5.10 GHz. The TDP is 245 watts for the Xeon Phi and 55 watts for the Core Ultra 5.

The socket differs: the Xeon Phi uses Intel Socket 3647, while the Core Ultra 5 uses Intel BGA 2114. The architecture is Knights Landing for the Xeon Phi and Arrow Lake for the Core Ultra 5. The process node is 14 nm for the Xeon Phi and 3 nm for the Core Ultra 5. The foundry is Intel for the Xeon Phi and TSMC for the Core Ultra 5. The transistor count is 8,000 million for the Xeon Phi and 17,800 million for the Core Ultra 5. The die size is not listed for the Xeon Phi, but is 243 mm² for the Core Ultra 5.

The cache configuration is very different. The Xeon Phi has 32 KB of L1 per core and 512 KB of L2 per core, with no L3 cache listed. The Core Ultra 5 has 192 KB of L1 per core, 3 MB of L2 per core, and 24 MB of shared L3 cache. Memory support is DDR4 for the Xeon Phi and DDR5 for the Core Ultra 5. The Core Ultra 5 has a dual-channel memory bus with 102.4 GB/s bandwidth, while no memory bus or bandwidth is listed for the Xeon Phi. ECC is supported on the Xeon Phi but not on the Core Ultra 5.

The Core Ultra 5 has a PCIe Gen 5 interface with 20 lanes (CPU only), while no PCIe information is given for the Xeon Phi. The Core Ultra 5 includes integrated graphics (Arc Xe-LPG Graphics 48EU), while the Xeon Phi has none. The market segment is Server/Workstation for the Xeon Phi and Mobile for the Core Ultra 5. The production status is not listed for the Xeon Phi, but is Active for the Core Ultra 5. The release date is 2016-06-19 for the Xeon Phi and 2025-01-12 for the Core Ultra 5. The multiplier is locked on the Xeon Phi and unlocked on the Core Ultra 5. The part numbers are SR2WY for the Xeon Phi and SRVFL for the Core Ultra 5.

DETAILED SPECIFICATIONS

SPECIFICATION
Ultra 5 235HX
Phi 7290
Core Specs
Cores
14
72 +414.3%
Threads
14
288 +1957.1%
Base Clock (GHz)
2.9
1,500 +51624.1%
Boost Clock (GHz)
5.1
1,700 +33233.3%
Frequency (GHz)
2.9
1,500 +51624.1%
Turbo Clock (GHz)
5.1
1,700 +33233.3%
Multiplier
29
15 -48.3%
SMP CPUs
1
1 0.0%
Cache
L1 Cache
192 KB (per core)
32 KB (per core)
L2 Cache
3 MB (per core)
512 KB (per core)
L3 Cache
24 MB (shared)
—
Power
TDP (W)
55
245 +345.5%
PL1
55 W
—
PL2
160 W
—
Architecture
Architecture
Arrow Lake
Knights Landing
Codename
Arrow Lake-HX
Knights Landing
Generation
Ultra 5 (Arrow Lake-HX)
Xeon Phi (Knights Landing)
Process Size
3 nm
14 nm
Transistors
17,800 million
8,000 million
Die Size
243 mm²
—
Foundry
TSMC
Intel
Memory
Memory Support
DDR5
DDR4
Memory Bus
Dual-channel
—
Memory Bandwidth
102.4 GB/s
—
ECC Memory
No
Yes
Platform
Socket
Intel BGA 2114
Intel Socket 3647
Chipsets
WM880, HM870
—
PCIe
Gen 5, 20 Lanes(CPU only)
—
Intel Hybrid
Hybrid Cores
P-Cores: 6 E-Cores: 8
—
E-Core Frequency
2.6 GHz up to 4.5 GHz
—
AI/NPU
NPU
Yes / 13 TOPS
—
Graphics
Integrated Graphics
Arc Xe-LPG Graphics 48EU
—
Other
Market
Mobile
Server/Workstation
Production Status
Active
—
Part Number
SRVFL
SR2WY
Package
FC-BGA
FC-LGA3647
Tj Max
105°C
—
View Core Ultra 5 235HX Details View Xeon Phi 7290 Details