SYSTEM ANALYZER

Rate My PC: Intel Core Ultra 5 338H + NVIDIA GeForce RTX 4070 Mobile

Get a comprehensive performance analysis of your gaming rig with detailed benchmarks, bottleneck detection, and upgrade recommendations

92 / 100
ULTIMATE READY

Apex Performer

Top 8% of systems. Capable of 4K Ultra gaming and advanced rendering.

4K 60+ FPSVR ReadyRay Tracing

System Balance Analysis

CPU vs GPU performance ratio
Well Balanced
CPU
90%
VS
GPU
93%
PROCESSOR

Intel Core Ultra 5 338H

33,989 Benchmark Score
Top 10% Market Ranking
View Full Specs →
GRAPHICS CARD

NVIDIA GeForce RTX 4070 Mobile

27,435 Benchmark Score
Top 7% Market Ranking
View Full Specs →

Market Position

How your build compares to others
Budget
0-30
Mid-Range
30-60
High-End
60-85
Enthusiast
85-100
Your Build

Game Performance Benchmarks

Real-world 4K FPS in popular titles
View All Games →

Performance Insights

Tips to maximize your system

Optimal Performance

Your system is in the top tier. You can run any modern game at maximum settings.

4K Gaming Ready

Consider a 4K 144Hz monitor to fully utilize your hardware capabilities.

Compatible Games See what you can play Compare CPUs Find upgrades Compare GPUs Find upgrades

Performance Tiers Explained

90-100

Ultimate

4K Ultra gaming, VR ready, ray tracing enabled, professional workloads

4K 60+ FPS VR Ready
70-89

High-End

1440p Ultra or 4K High settings, excellent for modern AAA titles

1440p Ultra 4K High
50-69

Mid-Range

1080p Ultra or 1440p Medium, great value for most gamers

1080p Ultra 1440p Med
30-49

Entry Level

1080p Medium settings, suitable for eSports and older titles

1080p Med eSports
0-29

Legacy

Basic gaming, older titles, consider upgrading for modern games

720p-1080p Low Older Games

The Intel Core Ultra 5 338H and NVIDIA GeForce RTX 4070 Mobile pairing represents a high-performance mobile computing configuration, placing in the 79th percentile overall among all benchmarked systems. This laptop-class build combines a next-generation Panther Lake processor with an Ada Lovelace graphics processor, creating a platform that is positioned to handle demanding workloads across both CPU-intensive and GPU-accelerated applications. The data indicates a system that sits at a critical intersection of portability and performance, though the absence of directly measured game frame rates requires careful inference from the synthetic benchmark scores.

GPU Analysis — VRAM, bandwidth, clocks, RT/tensor hardware, what the benchmark scores mean for rendering

The NVIDIA GeForce RTX 4070 Mobile is built on the AD106 chip using TSMC's 5 nm process node, with a transistor count of 22,900 million on a 188 mm² die. This translates to a transistor density of 121.8M per mm², a figure that speaks to the architectural efficiency of the Ada Lovelace design. The GPU operates with a base clock of 1395 MHz and a boost clock of 1695 MHz, with memory running at 2000 MHz, which yields 16 Gbps effective speed. This is not a maximalist configuration by desktop standards, but the data suggests a carefully balanced mobile part.

Memory configuration is a critical element here. The GPU is equipped with 8 GB of GDDR6 memory on a 128-bit bus, providing a total bandwidth of 256.0 GB/s. This bandwidth figure is a limiting factor for certain high-resolution textures and compute workloads, yet it remains sufficient for the vast majority of gaming and creative applications at the laptop's target resolutions. The 128-bit bus width is notably narrower than what one might find in higher-tier desktop cards, and the data implies that memory bandwidth could become a constraint in scenarios involving heavy data streaming, such as 4K texture packs or large scene compilation in 3D rendering.

The compute infrastructure is substantial. With 4608 shading units, 144 texture mapping units, and 48 raster operation pipelines, the GPU delivers a pixel rate of 81.36 GPixel/s and a texture rate of 244.1 GTexel/s. The FP32 performance stands at 15.62 TFLOPS, with FP16 performance also at 15.62 TFLOPS due to a 1:1 ratio, a characteristic of the Ada Lovelace architecture that favors consistent throughput across precision levels. RT acceleration is handled by 36 dedicated RT cores, while AI-adjacent tasks are supported by 144 tensor cores. This hardware complement positions the RTX 4070 Mobile as a capable ray tracing solution, though the benchmark scores suggest that its strength lies more in rasterization than in path-traced extremes.

Benchmark results for the GPU paint a clear picture of its standing. The PassMark G3D score is 19587, which places the GPU in the 73rd percentile of all GPUs. This is a high placement, but the nearest rival data reveals the competitive landscape. The AMD Radeon RX 6700 XT scores 27425 on average, a delta of 0%, meaning the RTX 4070 Mobile is functionally equivalent in this metric. More surprisingly, the NVIDIA GeForce RTX 3090 averages 27565, with the RTX 4070 Mobile trailing by only 0.5%. This indicates that the mobile 4070 delivers performance comparable to a previous-generation desktop flagship, at least in synthetic aggregate. The Geekbench OpenCL score of 109197 and Vulkan score of 108367 further corroborate this, showing strong compute capability. The PassMark GPU Compute score of 8399, however, is notably lower relative to the G3D score, suggesting that pure compute workloads may not scale as well as graphics-specific ones. For rendering, the data implies that the RTX 4070 Mobile is well-suited for 1080p and 1440p workloads, with ray tracing effects feasible due to the RT core count, but the memory bandwidth and bus width will be the limiting factors when pushing higher resolutions or complex scenes.

Balance and Bottleneck — which component limits which workload, using percentiles and FPS scaling as evidence

The balance between the Intel Core Ultra 5 338H and the RTX 4070 Mobile is the central question for this platform. The CPU holds an 84th percentile ranking among all CPUs, while the GPU sits at the 73rd percentile. This divergence in percentile positions is the first indicator that the processor may have more headroom relative to the graphics solution in certain workloads. However, the combined system percentile of 79 suggests that neither component is dramatically outclassing the other, creating a relatively harmonious pairing.

The benchmark data provides evidence for workload-specific bottlenecks. The CPU's PassMark Multithread score of 28717 and Single Thread score of 4180 indicate a strong processor that can feed a GPU efficiently in most gaming scenarios. The GPU's PassMark G3D score of 19587, when compared to its nearest rival at 0% delta, shows that it is a mid-to-high-tier performer. In gaming, the CPU is unlikely to be a limiting factor in most titles, given its 84th percentile standing, which exceeds the GPU's 73rd percentile. This suggests that frame rates will be GPU-bound in the majority of cases, meaning the RTX 4070 Mobile will be the primary determinant of FPS, especially at higher resolutions where pixel throughput becomes the constraint.

Yet, there are workloads where the CPU could become the bottleneck. The CPU's PassMark Physics score of 2697 is relatively low, and multi-threaded integer operations score 64934, which, while respectable, may lag in physics-heavy simulations or certain strategy games that rely on CPU threading. The PassMark Data Encryption score of 21367 and Data Compression score of 276539 indicate that the CPU handles these tasks well, but the GPU's compute score of 8399 is significantly lower than its graphics score. For compute-intensive tasks like machine learning inference or certain rendering pipelines, the GPU may not accelerate as much as expected, leaving the CPU to shoulder more of the burden.

The absence of measured FPS data for this exact combination is a notable gap. Since no measuredFpsUltraByGame rows exist, any FPS scaling discussion must be framed as estimations based on the synthetic scores. The data implies that in GPU-bound scenarios, the system will scale with GPU clock and memory bandwidth, while in CPU-bound scenarios, the 338H's 4.70 GHz boost clock and single-thread score of 2044 in Cinebench R23 will be the determining factor. The balance is tipped toward the GPU for gaming, but the CPU is not weak enough to create a severe imbalance, making this a well-matched pair for most mixed-use cases.

CPU Analysis — cores, clocks, architecture, what the benchmark scores mean for real workloads

The Intel Core Ultra 5 338H is a 12-core, 12-thread processor based on the Panther Lake architecture, built on Intel's 3 nm process node. It is part of the Core Ultra Series 3, with a base clock of 1.90 GHz and a boost clock of 4.70 GHz. The lack of hyperthreading, as evidenced by the equal core and thread counts, is a notable design choice that affects multi-threaded scaling. The cache hierarchy includes 192 KB of L1 per core, 2.5 MB of L2 per core, and 18 MB of shared L3 cache. This cache configuration is generous for a mobile part, particularly the L2 allocation, which should help with data locality in single-threaded tasks.

The CPU's benchmark scores reveal its performance tier. In Cinebench R23, it scores 16331 in multi-core and 2044 in single-core. These are strong numbers for a 25-watt TDP part, indicating that Intel has achieved significant efficiency gains with the Panther Lake architecture. The PassMark Single Thread score of 4180 is competitive, and the MultiThread score of 28717 places it in the upper echelon of mobile processors. The average benchmark score of 33989 puts it in the 84th percentile of all CPUs, a remarkable achievement for a processor with a 25-watt thermal envelope.

The nearest rivals provide context. The Intel Core Ultra 7 165H averages 34083, with the 338H trailing by only 0.3%, making them effectively equivalent in aggregate performance. The Intel Core i7-12800HX averages 33875, with the 338H leading by 0.3%. This is particularly interesting because the i7-12800HX is a higher-power HX-series part, yet the 338H matches it in average score, underscoring the architectural efficiency of Panther Lake. The AMD EPYC 4244P, a server chip, averages 34220, with the 338H trailing by 0.7%, and the Intel Xeon 6353P averages 33844, with the 338H leading by 0.4%. This suggests that the 338H, despite being a mobile chip, competes with desktop and server processors in synthetic benchmarks, proof of its IPC improvements.

Real-world workload implications are clear. The single-core performance of 2044 in Cinebench R23 suggests excellent responsiveness in daily tasks, web browsing, and office applications. The multi-core score of 16331 indicates that the processor can handle video encoding, software compilation, and 3D scene rendering, though the 12-thread limit may constrain highly parallel workloads compared to 16-thread or 24-thread competitors. The PassMark Floating Point Math score of 84067 and Integer Math score of 64934 show balanced arithmetic capabilities, while the Extended Instructions score of 23906 and Find Prime Numbers score of 304 indicate that AVX-512 and other instruction sets are handled competently. The Data Encryption score of 21367 and Data Compression score of 276539 suggest that the CPU is well-suited for file archiving and encryption tasks. The Random String Sorting score of 34082 further indicates strong memory subsystem performance. Overall, the 338H is a high-efficiency processor that delivers desktop-class performance in a mobile form factor, with the caveat that its thread count is lower than some rivals.

Who Should Build It — target users and industries (gamers at specific resolutions, content creators, developers, students, small business workstations) tied strictly to the measured performance

The target audience for this build is defined by the measured performance of both the CPU and GPU. Given the GPU's 73rd percentile ranking and the CPU's 84th percentile ranking, this system is positioned for users who need a balance of CPU and GPU power without sacrificing portability. Gamers at 1080p and 1440p resolutions are the most obvious beneficiaries. The GPU's PassMark G3D score of 19587, which is on par with the desktop RX 6700 XT and RTX 3090 in aggregate, implies that it can handle demanding titles at high settings in these resolutions, provided the memory bandwidth of 256.0 GB/s is sufficient for the game's texture requirements.

Content creators, particularly video editors and 3D artists, will find value in this configuration. The CPU's Cinebench R23 multi-core score of 16331 and PassMark Multithread score of 28717 suggest that rendering times will be reasonable for shorter projects. The GPU's 36 RT cores and 144 tensor cores enable hardware-accelerated ray tracing and AI-driven features like DLSS, which can speed up render previews and final output. The FP32 performance of 15.62 TFLOPS is adequate for GPU-accelerated rendering in applications like Blender or Octane, though the compute score of 8399 is a limiting factor for longer, more complex renders.

Software developers will appreciate the CPU's single-thread performance of 4180 in PassMark, which translates to fast compilation times for single-threaded build steps, and the multi-thread score of 28717 for parallel builds. The 18 MB of L3 cache and 2.5 MB of L2 per core are beneficial for code and data locality. Students and small business users who rely on office productivity, spreadsheets, and web-based applications will find the system more than capable, with the CPU's single-core strength ensuring snappy UI responsiveness. The 25-watt TDP of the CPU, combined with the GPU's 115-watt TDP, suggests a total system power draw that is manageable for a laptop, making it suitable for users who need to work on the go. However, the data does not support recommending this build for users who require extreme multi-threaded throughput, as the 12-thread limit may be a bottleneck for heavy server-style workloads.

Upgrade Path and Platform — socket, memory support, PCIe, PSU headroom from suggestedPsu/tdp, what a sensible next upgrade looks like

The upgrade path for this platform is constrained by its mobile form factor and the specifics of its hardware interfaces. The CPU uses the Intel BGA 2540 socket, which is a soldered, non-upgradeable connection. This means that the processor is permanently attached to the motherboard, and any CPU upgrade would require replacing the entire system or motherboard, which is not a practical option for most laptop users. The memory support is limited to LPDDR5X in a dual-channel configuration, providing a memory bandwidth of 136.5 GB/s. This is soldered memory in most implementations, so upgrading RAM capacity or speed is typically not feasible.

The PCIe interface for the CPU is Gen 5 with 4 lanes, which is reserved for the CPU's direct connections, likely to the GPU or a high-speed NVMe SSD. The GPU, however, uses a PCIe 4.0 x8 interface, which is standard for mobile graphics. This is a potential bottleneck for data transfer between the CPU and GPU, but in practice, the bandwidth available on PCIe 4.0 x8 is sufficient for gaming and most creative workloads. The lack of a suggested PSU figure in the data, combined with the CPU's 25-watt TDP and the GPU's 115-watt TDP, implies that the system is designed to operate within a specific power envelope typical of a laptop charger. There is no headroom for adding a discrete graphics card or expanding storage beyond what the motherboard supports.

A sensible next upgrade, given these constraints, would be to focus on external peripherals or storage. The PCIe Gen 5 lanes could support a high-speed external NVMe enclosure, though the practical benefit would be limited by the interface's 4-lane allocation. Alternatively, users might consider an external GPU enclosure, but the GPU's PCIe 4.0 x8 interface already occupies the primary connection, making this impractical. The data suggests that this is a closed platform with limited internal upgradability, and the most logical upgrade path is to invest in a new system with a higher-tier processor or GPU when performance demands exceed this build's capabilities. The platform's strengths lie in its integrated design and efficiency, not its expandability.

Usage Scenarios — grounded in the scores: high-refresh gaming, streaming, video editing, 3D rendering, software development, student and office work. One short paragraph per scenario, citing the numbers that support the verdict

High-refresh gaming: The RTX 4070 Mobile's PassMark G3D score of 19587, which is 0% behind the desktop RX 6700 XT, suggests that it can drive high-refresh-rate panels at 1080p in most titles. The CPU's single-thread score of 4180 ensures that frame pacing is consistent, but the GPU is the limiting factor, and users should expect to adjust settings to maintain 144 Hz or higher in the latest AAA games.

Streaming: The CPU's 12 threads and PassMark Multithread score of 28717 provide enough headroom for software encoding at moderate bitrates, but the GPU's NVENC capabilities, supported by its 144 tensor cores, are more efficient. The data implies that streamers should use GPU-based encoding to minimize CPU load, reserving the processor's 4.70 GHz boost clock for game logic and overlays.

Video editing: The CPU's Cinebench R23 multi-core score of 16331 and the GPU's FP32 performance of 15.62 TFLOPS combine to accelerate both timeline scrubbing and export rendering. The 8 GB of VRAM on the GPU is sufficient for 1080p and 1440p timelines, though 4K projects with heavy effects may exceed this capacity, leading to slower performance.

3D rendering: The CPU's average benchmark score of 33989, which is 0.3% ahead of the Core i7-12800HX, indicates solid CPU-based rendering performance. For GPU rendering, the RTX 4070 Mobile's 36 RT cores and 15.62 TFLOPS are capable, but the PassMark GPU Compute score of 8399 is lower than its graphics score, suggesting that render times will be longer than for a more compute-optimized card.

Software development: The CPU's PassMark Single Thread score of 4180 and Integer Math score of 64934 indicate fast compilation for single-threaded and multi-threaded build steps. The 18 MB of L3 cache helps with code reuse, and the overall 84th percentile ranking means that developers will rarely wait on the processor for builds, though large projects with many parallel tasks may expose the 12-thread limit.

Student and office work: The CPU's single-core Cinebench R23 score of 2044 ensures that web browsing and document editing feel responsive. The integrated Arc B370 graphics, combined with the discrete GPU, provides ample display output capability, and the 25-watt TDP means long battery life for all-day use on campus or in the office.

Gaming Performance — measured FPS by game and resolution from measuredFpsUltraByGame (or, if dataIsMeasured is false, frame expectations qualitatively from the benchmark scores and say the figures are estimates)

There are no measured FPS rows for this exact CPU-GPU combination in the data. The FACT PACK contains no measuredFpsUltraByGame entries, meaning any frame rate figures would be speculative and are not provided. Instead, gaming performance must be estimated from the synthetic benchmark scores, and the following analysis should be treated as an estimation rather than a definitive measurement.

The GPU's PassMark G3D score of 19587, which is statistically tied with the desktop RX 6700 XT (0% delta) and within 0.5% of the RTX 3090, provides a baseline for expectations. In aggregate, this suggests that the RTX 4070 Mobile can match a prior-generation desktop mid-range card. For 1080p gaming at ultra settings, this implies that most titles should run at high frame rates, likely exceeding 60 FPS in many cases, with esports titles potentially reaching 144 FPS or higher. At 1440p, the 256.0 GB/s memory bandwidth becomes a more significant factor, and frame rates are expected to be lower but still playable in the majority of games, with some demanding titles requiring a reduction to high settings.

The CPU's performance is not expected to be a limiting factor in most games. The 84th percentile ranking and single-thread score of 4180 indicate that the 338H can feed the GPU adequately, avoiding CPU-bound frame drops in most scenarios. However, in games that are heavily dependent on multi-threaded CPU performance, the 12-thread limit might cause occasional stutters, particularly in large-scale simulations or strategy titles. The absence of measured data is a significant caveat, so these figures should be interpreted as directional guidance rather than precise predictions. Users seeking exact frame rates should look for independent testing of this specific configuration.

FAQ

Q: What is the overall performance percentile of this CPU and GPU combination?

A: The combined system percentile is 79, placing it in the upper tier of all benchmarked systems. The CPU scores in the 84th percentile, while the GPU scores in the 73rd percentile.

Q: How does the Intel Core Ultra 5 338H compare to its nearest CPU rival, the Core Ultra 7 165H?

A: The two processors are effectively equivalent in aggregate performance. The Core Ultra 7 165H averages 34083, while the 338H averages 33989, a delta of only -0.3%.

Q: What is the memory bandwidth of the RTX 4070 Mobile and how does it affect performance?

A: The GPU has 8 GB of GDDR6 memory on a 128-bit bus, providing a bandwidth of 256.0 GB/s. This is a potential constraint for high-resolution textures and heavy data streaming workloads.

Q: Does the RTX 4070 Mobile support ray tracing and AI acceleration?

A: Yes, it has 36 RT cores for ray tracing and 144 tensor cores for AI tasks. It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What is the CPU's boost clock and how does it contribute to single-threaded performance?

A: The CPU has a boost clock of 4.70 GHz, which contributes to its PassMark Single Thread score of 4180 and Cinebench R23 single-core score of 2044.

Q: What is the process node for the CPU and GPU?

A: The CPU is built on Intel's 3 nm process, while the GPU is built on TSMC's 5 nm process.

Q: Is there a measured FPS data for this specific build?

A: No, there are no measured FPS rows for this exact combination. All gaming performance conclusions are estimated from synthetic benchmark scores.

Benchmark Performance — exact CPU and GPU scores, percentile positions, and what the combined picture is

The CPU's benchmark results are led by a Cinebench R23 multi-core score of 16331 and a single-core score of 2044. In Cinebench R20, it scores 10213 multi-core and 1441 single-core, while Cinebench R15 yields 2504 multi-core and 305 single-core. The PassMark suite shows a MultiThread score of 28717 and Single Thread score of 4180. Additional PassMark scores include Floating Point Math at 84067, Integer Math at 64934, Extended Instructions at 23906, Data Encryption at 21367, Data Compression at 276539, Random String Sorting at 34082, Physics at 2697, and Find Prime Numbers at 304. The average benchmark score for the CPU is 33989, placing it in the 84th percentile of all CPUs. Its nearest rival, the Core Ultra 7 165H, averages 34083, a delta of -0.3%.

The GPU's benchmark results include a PassMark G3D score of 19587 and a GPU Compute score of 8399. Geekbench OpenCL scores 109197, while Vulkan scores 108367. PassMark DirectX tests score 116 for DirectX 10, 179 for DirectX 11, 85 for DirectX 12, and 223 for DirectX 9. The G2D score is 763. The average benchmark score for the GPU is 27435, placing it in the 73rd percentile of all GPUs. Its nearest rival, the RX 6700 XT, averages 27425, a delta of 0%.

The combined picture shows a CPU that outperforms its GPU in relative percentile terms, with an 11-point gap between the 84th and 73rd percentiles. This suggests that the system is slightly CPU-strong, but the GPU is by no means a weak link, as its aggregate score is comparable to a desktop RTX 3090. The system's overall 79th percentile ranking indicates that this is a high-performance laptop configuration, but not a flagship one, as there are faster mobile parts available.

Build Overview — what this CPU+GPU pairing is, its class (desktop/laptop from buildClass), and overall tier from the percentiles

This build is a laptop-class configuration, as indicated by the buildClass field. It pairs the Intel Core Ultra 5 338H, a 12-core, 12-thread Panther Lake processor with a 25-watt TDP, with the NVIDIA GeForce RTX 4070 Mobile, an Ada Lovelace GPU with a 115-watt TDP. The combination is designed for high-performance mobile computing, targeting users who require desktop-class performance in a portable form factor.

The overall tier of this system is upper-mid-range to high-end, based on the combined percentile of 79. The CPU's 84th percentile ranking places it among the top processors, outperforming the Core Ultra 7 165H and Core i7-12800HX in aggregate score. The GPU's 73rd percentile ranking places it in the upper-mid tier of graphics cards, matching a desktop RX 6700 XT and nearly matching an RTX 3090 in synthetic benchmarks. This pairing is not the absolute fastest available, but it offers a strong balance of CPU and GPU performance for a wide range of tasks.

The system is further characterized by its connectivity options, including PCIe Gen 5 with 4 lanes for the CPU and PCIe 4.0 x8 for the GPU. The integrated Arc B370 graphics provides a fallback for low-power tasks, while the discrete GPU handles demanding workloads. The platform supports LPDDR5X memory with a bandwidth of 136.5 GB/s, and the CPU is built on Intel's 3 nm process, indicating a focus on efficiency. This is a modern, well-rounded laptop platform that excels in multi-threaded productivity and 1080p-to-1440p gaming, with the caveat that its 12-thread CPU and 8 GB GPU memory are not at the extreme high end. It is a sensible choice for users who want a portable system that can handle almost any task without making major compromises.