Intel Core i7-1165G7

The Intel Core i7-1165G7 is a mobile processor from the Tiger Lake-U generation, built on Intel's 10 nm process and featuring 4 cores and 8 threads. Its benchmark data places it in the 70th percentile of all CPUs, with an average benchmark score of 9787, indicating a solid mid-range position within the mobile landscape.

Benchmark Performance

The benchmark suite reveals a processor that scales predictably with thread count, but with notable headroom in multi-threaded workloads. In 3DMark testing, the chip scores 811 in single-thread, 1342 in 2-thread, 2015 in 4-thread, 2500 in 8-thread, 2617 in 16-thread, and 2645 in max-thread tests. The near-linear scaling from 8-thread (2500) to max-thread (2645) suggests that the 4-core/8-thread configuration is fully utilized by the 16-thread test, with only marginal gains beyond that point. This pattern indicates efficient thread scheduling and resource allocation across the available cores.

Cinebench results reinforce this behavior. The R23 multi-core score of 8460 versus a single-core score of 1194 yields a multi-to-single ratio of roughly 7.1x, which is strong for a quad-core part. The R20 results (3553 multi-core, 501 single-core) and R15 results (852 multi-core, 120 single-core) follow the same trajectory, showing consistent performance across different rendering workloads. These scores position the 1165G7 as a capable performer for content creation tasks that can leverage multiple threads, though the absolute numbers trail higher-core-count competitors.

PassMark tests provide a broader view. The multi-thread score of 9958 and single-thread score of 2763 show a ratio of 3.6x, which is lower than the Cinebench ratio, reflecting that PassMark's workload mix is less parallelizable. Specific sub-tests illustrate this: floating-point math scores 19830, integer math scores 33349, data compression hits 103497, and data encryption reaches 5561. The find-prime-numbers score of 46 is notably low, indicating weak performance in highly sequential integer loops. The extended-instructions score of 7323 suggests decent SIMD throughput, which benefits media encoding and scientific workloads.

Relative to its nearest rivals, the 1165G7 is marginally behind the Intel Core Ultra 5 245HX by 0.6%, with average scores of 9787 versus 9849. It leads the Intel Core i5-1035G1 by 1%, the Intel Core i5-1135G7 by 1.7%, and the AMD EPYC 7402P by 2.7%. These deltas are small, meaning the 1165G7 sits in a tightly contested performance band where generational and architectural differences produce only minor score variations.

Platform and Compatibility

The processor uses the Intel BGA 1449 socket, making it a soldered mobile part with no upgrade path for end users. It belongs to the Tiger Lake-U family, codenamed Tiger Lake, and is built on the 10 nm process node with a die size of 144 mm². The architecture is Willow Cove-U, representing Intel's Core i7 generation for thin-and-light laptops. The production status is end-of-life, with a release date of September 2020, meaning it is no longer actively manufactured.

Memory support includes DDR4 and LPDDR4X in a dual-channel configuration. ECC memory is not supported, which is typical for consumer mobile processors. The lack of a listed memory bandwidth figure prevents direct throughput comparisons, but dual-channel operation is the standard for this class of chip. PCIe support is Gen 4 with 4 lanes available from the CPU only, which provides adequate bandwidth for a discrete GPU or fast NVMe storage, though the lane count is limited compared to desktop platforms.

The integrated graphics are Iris Xe-LP Graphics G7 with 96 execution units. This is a high-end integrated GPU for the Tiger Lake generation, offering substantial graphics horsepower for a chip without a discrete GPU. The socket and platform are designed for ultraportable and mainstream laptops, not desktop systems, so cooling and power delivery are dictated by the laptop chassis rather than aftermarket motherboard choices. The multiplier is locked, preventing overclocking.

How It Compares

Against the Intel Core Ultra 5 245HX, the 1165G7 trails by 0.6% in average score (9787 vs 9849). This negligible gap is remarkable given the generational difference, suggesting that for the tested workloads, the older Tiger Lake architecture holds its own against a newer, presumably higher-tier part. The delta is within run-to-run variance, making these two effectively equivalent in raw benchmark throughput.

Versus the Intel Core i5-1035G1, the 1165G7 leads by 1% (9787 vs 9691). This is a modest advantage, reflecting the architectural improvements from Ice Lake (1035G1) to Tiger Lake (1165G7). The higher boost clock of 4.70 GHz on the 1165G7, combined with the Willow Cove cores, provides a consistent but small edge across the benchmark suite.

The Intel Core i5-1135G7 is the closest rival at 1.7% behind (9620 vs 9787). Both are Tiger Lake parts, but the i7-1165G7 has a higher base clock (2.80 GHz) and likely better binning, resulting in a measurable lead. The 1.7% delta is consistent across multi-threaded and single-threaded tests, indicating a clock-for-clock advantage rather than a workload-specific benefit.

The AMD EPYC 7402P, a server-class part, sits 2.7% behind (9529 vs 9787). This is surprising given the EPYC's enterprise positioning, but the benchmark average likely reflects the 1165G7's superior single-thread performance and the EPYC's lower per-core clocks in these specific tests. The 1165G7's mobile efficiency does not translate to a deficit here; it actually outpaces the server chip in average score.

Who Should Consider It

For gaming workloads, the 1165G7's integrated Iris Xe-LP Graphics G7 with 96 EU is a strong integrated solution, but the 3DMark scores (811 single-thread, 2645 max-thread) suggest it can handle esports titles and older games at playable frame rates, not demanding AAA releases. The single-thread 3DMark score of 811 is competitive for a mobile part, which helps in games that rely on a few fast cores. The 4-core/8-thread configuration is sufficient for modern game physics and AI, but the lack of more cores may bottleneck heavily threaded titles.

Content creation is where the 1165G7 shows its strength. The Cinebench R23 multi-core score of 8460, combined with floating-point math at 19830, indicates solid rendering and simulation performance for a thin-and-light laptop. The data compression score of 103497 is particularly strong, benefiting file archiving and compression workflows. The extended-instructions score of 7323 supports video encoding and image processing, making this a viable choice for photo editing and light video work on the go.

Office and productivity tasks are handled comfortably. The PassMark single-thread score of 2763 is among the higher results for this class, ensuring snappy response in word processors, spreadsheets, and web browsers. The integer math score of 33349 handles typical office calculations and database operations without strain. The 70th percentile overall ranking means it outperforms a majority of CPUs, so it will not feel sluggish in everyday use. However, the end-of-life status and limited 4-lane PCIe Gen 4 may deter those needing future expansion.

Power and Thermals

The TDP is rated at 28 W, placing the 1165G7 in the upper tier of ultraportable processors. This TDP class typically requires a capable cooling solution, such as a dual-heat-pipe cooler with a dedicated fan, to sustain boost clocks under load. The 10 nm process helps with efficiency, but 28 W is a design point that allows for performance bursts without excessive heat in thin chassis.

The base clock of 2.80 GHz and boost clock of 4.70 GHz indicate a wide frequency range, which is common for mobile parts that need to balance performance and thermals. Sustained multi-threaded workloads will likely push the power envelope to its limit, requiring the cooling system to dissipate 28 W continuously. Single-threaded tasks can boost higher but draw less total power, making thermal management more straightforward.

Given the end-of-life production status, the 1165G7 is no longer found in new systems, but used laptops with this chip will have aging thermal solutions. The 28 W TDP is manageable for a laptop with adequate ventilation, but users should expect fan noise under sustained load. The lack of a multiplier unlock means no user-level overclocking to increase power draw, which helps keep thermals predictable.

Single-Thread vs Multi-Thread Behavior

The split between single-thread and multi-thread performance is the defining characteristic of the 1165G7. In Cinebench R23, the single-core score of 1194 is strong for a 2020 mobile chip, while the multi-core score of 8460 shows efficient scaling across all 8 threads. The ratio of 7.1x indicates that the processor extracts near-optimal performance from its 4 cores and 8 threads, with minimal overhead from thread synchronization.

PassMark tells a different story. The single-thread score of 2763 is excellent, but the multi-thread score of 9958 yields a ratio of only 3.6x. This discrepancy arises because PassMark's workloads include many single-threaded components that do not benefit from additional cores. The find-prime-numbers score of 46 is a clear example: this highly sequential algorithm runs at roughly the same speed regardless of thread count, dragging down the multi-thread aggregate.

For real-world applications, this means the 1165G7 excels in workloads that are inherently single-threaded, such as web browsing, office applications, and light coding. Multi-threaded workloads that are well-parallelized, like video rendering in Cinebench, see substantial gains. However, workloads with mixed parallelism, such as compression or encryption, will show variable performance. The data encryption score of 5561 and data compression score of 103497 indicate that these tasks scale reasonably but not perfectly, reflecting the limitations of a 4-core design in latency-sensitive operations. The single-thread 3DMark score of 811 versus max-thread 2645 further underscores this: gaming physics and logic benefit from the high single-thread score, while scene rendering scales with threads. Users should match their primary workloads to the chip's strengths in fast single-thread execution and efficient, if not massive, multi-thread scaling.