SYSTEM ANALYZER

Rate My PC: Intel Core Ultra 5 225T + NVIDIA GeForce RTX 5090 D

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

93 / 100
ULTIMATE READY

Apex Performer

Top 7% 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
89%
VS
GPU
98%
PROCESSOR

Intel Core Ultra 5 225T

30,468 Benchmark Score
Top 11% Market Ranking
View Full Specs →
GRAPHICS CARD

NVIDIA GeForce RTX 5090 D

77,712 Benchmark Score
Top 2% 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

# GPU Analysis

The NVIDIA GeForce RTX 5090 D represents the apex of the GeForce 50-series, built on the Blackwell 2.0 architecture with a 5 nm TSMC process. The GPU packs 92,200 million transistors on a 750 mm² die, yielding a transistor density of 122.9M per mm². This is a massive chip, and the data reflects that scale. The memory subsystem is equally imposing: 32 GB of GDDR7 across a 512-bit bus delivers 1.79 TB/s of bandwidth, a figure that dwarfs most competing solutions. The 1750 MHz memory clock translates to 28 Gbps effective, which is the fastest memory configuration in the fact pack.

Compute resources are staggering. The RTX 5090 D features 21,760 shading units, 680 texture mapping units, and 176 raster output pipelines. The ray tracing hardware consists of 170 RT cores, while the tensor core count sits at 680. These numbers explain the GPU's raw throughput: FP32 performance reaches 104.8 TFLOPS, with FP16 matching that at 104.8 TFLOPS (1:1). The pixel rate is 423.6 GPixel/s, and the texture rate is 1,636.8 GTexel/s. The 3DMark Steel Nomad DX12 score of 14,326 confirms this is a top-tier rendering part, placing it in the 92nd percentile of all GPUs.

What does this mean for rendering workloads? The 32 GB VRAM capacity is not just large; it is transformative for high-resolution texture sets and complex scenes that would otherwise spill into system memory. The 1.79 TB/s bandwidth ensures that those textures stream without stuttering. The 170 RT cores provide dedicated hardware for ray-traced effects, while the 680 tensor cores accelerate AI-driven features like DLSS and denoising. The Vulkan score of 376,915 and OpenCL score of 310,674 in Geekbench indicate strong compute versatility beyond pure graphics. For 3D artists, the FP32 and FP16 parity is notable—it means mixed-precision workloads run at full speed without conversion penalties.

The GPU's 575 W TDP and 950 W suggested PSU are substantial, but the dual-slot form factor (304 mm length, 137 mm height, 48 mm width) keeps physical installation manageable in most full-tower cases. The PCIe 5.0 x16 interface ensures no bandwidth bottleneck for data transfer, and the display outputs (1x HDMI 2.1b, 3x DisplayPort 2.1b) support modern high-refresh monitors. The PassMark G3D score of 44,065 and GPU compute score of 28,396 further corroborate the GPU's dominance, though the DirectX 9/10/11/12 PassMark scores (434, 231, 371, 219) are lower—these are legacy tests that do not scale with the Blackwell architecture's modern design.

# Benchmark Performance

The combined percentile for this CPU+GPU pairing is 87, placing it above the vast majority of desktop configurations. The GPU alone sits at the 92nd percentile of all GPUs, while the CPU holds the 82nd percentile among all CPUs. This creates an interesting dynamic: the GPU is significantly more dominant relative to its peers than the CPU is to its own. The average benchmark score for the CPU is 30,468, while the GPU's average is 77,712.

In Cinebench, the CPU delivers 21,971 in R23 multi-core and 3,101 in single-core. The R20 scores are 9,227 multi-core and 1,302 single-core, while R15 shows 2,214 multi-core and 312 single-core. These numbers indicate a CPU that is strong for productivity but not class-leading. The PassMark multi-thread score of 25,358 and single-thread score of 4,348 tell a similar story. The data compression score of 233,998 is exceptionally high, suggesting the CPU excels at tasks that leverage its cache hierarchy and memory bandwidth.

The GPU's 3DMark Steel Nomad score of 14,326 is the headline graphics result. Combined with the CPU's Cinebench R23 multi-core score of 21,971, the pairing shows a system capable of both heavy compute and high-end rendering. The CPU's nearest rivals—Intel Core i9-11980HK (avg score 30,422, delta 0.2%), AMD Ryzen 5 7640HS (30,390, delta 0.3%), Intel Core i5-13600H (30,548, delta -0.3%), and AMD Ryzen 7 7736U (30,364, delta 0.3%)—are all within a third of a percent of the Core Ultra 5 225T. This suggests the CPU performs at a level comparable to high-end mobile chips from previous generations, but the deltaPct values are so small that they are effectively noise.

The GPU's nearest rivals—AMD Radeon RX 6650M XT (avg score 76,904, delta 1.1%), AMD Radeon RX 6850M XT (78,940, delta -1.6%), NVIDIA Tesla P100 PCIe 12 GB (79,396, delta -2.1%), and NVIDIA Tesla P100 PCIe 16 GB (79,605, delta -2.4%)—show a different pattern. The RTX 5090 D is 1.1% ahead of the RX 6650M XT but trails the RX 6850M XT by 1.6% and the Tesla P100 variants by 2.1-2.4%. These are all mobile or data-center parts, which is unusual for a desktop flagship. The deltaPct values are small, but the fact that a desktop 5090 D is not decisively ahead of mobile GPUs in average benchmark score is noteworthy. The combined picture is a system where the GPU is the dominant component, with the CPU providing adequate but not overwhelming support.

# CPU Analysis

The Intel Core Ultra 5 225T is a 10-core, 10-thread desktop processor from the Core Ultra Series 2, built on the Arrow Lake architecture (Arrow Lake-S specifically) using a 3 nm TSMC process. The die contains 17,800 million transistors on a 243 mm² area. The base clock is 2.50 GHz with a boost clock of 4.90 GHz, and the TDP is rated at 65 W. This is a locked multiplier (multiplierUnlocked: false), meaning overclocking is not an option for users seeking extra performance.

The cache configuration is substantial: 192 KB of L1 per core, 3 MB of L2 per core, and 20 MB of shared L3. The memory support is DDR5 with dual-channel architecture, providing 102.4 GB/s of bandwidth. The integrated graphics are Arc Xe-LPG Graphics with 16 execution units, which is sufficient for basic display output but not for gaming. The CPU connects via PCIe Gen 5 with 20 lanes (CPU only), which is adequate for a single high-end GPU plus an NVMe drive.

The Cinebench R23 multi-core score of 21,971 is respectable for a 10-thread part, but it is not exceptional. The single-core score of 3,101 is more impressive, indicating strong per-thread performance that benefits lightly-threaded applications. The PassMark single-thread score of 4,348 corroborates this. The data compression score of 233,998 is the standout result—it suggests the CPU's memory subsystem and cache design are optimized for workloads that move large amounts of data. The floating-point math score of 82,751 and integer math score of 59,543 are solid but not spectacular.

The nearest rivals provide context. The Core Ultra 5 225T has an average benchmark score of 30,468, which is within 0.3% of the Intel Core i9-11980HK (30,422), AMD Ryzen 5 7640HS (30,390), and AMD Ryzen 7 7736U (30,364). The Intel Core i5-13600H is 0.3% ahead at 30,548. These are all mobile processors, which is telling—the 225T, a desktop chip, matches high-end laptop CPUs from the same era. For real workloads, this means the CPU handles office productivity, software compilation, and moderate content creation without issue. The 65 W TDP is frugal, which is a significant advantage for system builders concerned about thermals and power draw in compact desktop builds. However, the 10 threads limit heavy multi-threaded workloads like 3D rendering or video encoding, where 16+ thread CPUs would pull ahead.

# Upgrade Path and Platform

The Intel Core Ultra 5 225T uses the Intel Socket 1851, which is the platform for Arrow Lake processors. This socket supports DDR5 memory in dual-channel configuration, and the CPU provides 20 PCIe Gen 5 lanes. The platform is current and active, with the CPU released on 2024-12-31 and production status marked as active. This means the upgrade path forward includes other Socket 1851 processors, though the fact pack does not specify which models are available.

The memory support is DDR5 only, with no ECC option. The 102.4 GB/s bandwidth is adequate for the CPU's needs, but users looking to maximize memory performance would need to select high-speed DDR5 kits. The PCIe Gen 5 x16 interface on the GPU side is fully supported, ensuring the RTX 5090 D operates at full bandwidth.

The power requirements are the most critical upgrade consideration. The GPU has a 575 W TDP and the suggested PSU is 950 W. The CPU's 65 W TDP is modest by comparison, but the total system draw must account for both components plus peripherals. The GPU uses a single 16-pin power connector, which is standard for the 50-series. The dual-slot design (304 mm length) fits in most mid-tower and full-tower cases, but compact cases may struggle with the GPU's 48 mm width.

A sensible next upgrade would focus on the CPU if the user's workloads are multi-threaded, since the 10-thread limit is the primary bottleneck. Alternatively, adding more DDR5 memory would benefit data-intensive tasks like the compression workloads where the CPU already scores 233,998. The platform's PCIe Gen 5 support ensures that future storage devices (NVMe Gen 5) will run at full speed. The 950 W PSU provides headroom for additional drives or peripherals, but it is near the minimum for this GPU, so an upgrade to a higher-capacity PSU would be prudent if adding more components. The platform does not support ECC memory, so error-correcting workloads would require a different platform entirely.

# Who Should Build It

This pairing targets users who prioritize graphics performance above all else. The RTX 5090 D's 92nd percentile GPU ranking makes it a fit for 4K gaming and high-end rendering, but the CPU's 82nd percentile ranking means the system is not optimized for CPU-bound tasks. Gamers at 4K resolution will benefit most, as the GPU becomes the limiting factor at higher resolutions, allowing the 10-core CPU to keep up without bottlenecking. Gamers at 1080p or 1440p with high refresh rates may find the CPU limiting in CPU-intensive titles, as the 10 threads and 4.90 GHz boost clock may not sustain frame rates as high as the GPU can render.

Content creators working with video editing will see strong GPU acceleration in effects and encoding, but the CPU's 10 threads will slow timeline rendering and export times compared to higher-core-count alternatives. 3D artists using ray-traced rendering will benefit from the 170 RT cores and 32 GB VRAM, but CPU-based physics or simulation tasks will lag. Software developers compiling large codebases will find the CPU adequate but not fast; the PassMark data compression score of 233,998 suggests good performance for build tools that compress artifacts. Students and small business workstations will be overpowered by this GPU, but the system will handle any office task with ease, and the 65 W CPU TDP keeps power consumption reasonable for a workstation.

The GPU's 32 GB VRAM is a key selling point for machine learning researchers and data scientists who need to load large models into memory, though the lack of ECC memory on the CPU platform may be a concern for long-running compute jobs. The build class is desktop, so this is not a portable solution—it requires a dedicated desk and a 950 W PSU.

# Gaming Performance

The fact pack contains no measured FPS data for this exact CPU+GPU combination. There are no rows in the measuredFpsUltraByGame field, and dataIsMeasured is false. Therefore, all frame rate figures discussed here are estimates derived from the benchmark scores, not direct measurements.

Based on the GPU's 3DMark Steel Nomad score of 14,326 and its 92nd percentile ranking, the RTX 5090 D should deliver exceptional performance at 4K resolution in most games. The 32 GB VRAM ensures that even the most texture-heavy titles run without memory pressure. The 1.79 TB/s bandwidth allows for high frame rates at ultra settings, and the 170 RT cores provide dedicated ray tracing performance. Estimates suggest 4K ultra settings should yield well above 60 FPS in most titles, with many games exceeding 120 FPS at 1440p.

The CPU's 10 threads and 4.90 GHz boost clock are sufficient for most gaming scenarios, but the 82nd percentile ranking relative to all CPUs means that in CPU-bound scenarios—such as open-world games with heavy physics or large numbers of NPCs—frame rates may be lower than the GPU's capabilities would suggest. The single-core Cinebench R23 score of 3,101 is strong, which helps in games that rely on a few threads. The PassMark single-thread score of 4,348 supports this. However, competitive shooters at 1080p with high refresh rates (240 Hz or higher) may see frame rates capped by the CPU rather than the GPU.

These are estimates, not measurements. The absence of measured FPS data means the actual performance could vary by game, driver, and system configuration. The benchmark scores provide a strong foundation for these expectations, but users should treat them as a guide rather than a guarantee.

# FAQ

Q: What is the GPU's memory capacity and bandwidth?

A: The RTX 5090 D has 32 GB of GDDR7 memory on a 512-bit bus, delivering 1.79 TB/s of bandwidth.

Q: How does the CPU compare to its nearest rivals?

A: The Core Ultra 5 225T has an average benchmark score of 30,468, which is within 0.3% of the Intel Core i9-11980HK, AMD Ryzen 5 7640HS, and AMD Ryzen 7 7736U, and 0.3% behind the Intel Core i5-13600H.

Q: What is the combined percentile ranking of this system?

A: The combined percentile is 87, with the GPU at 92 and the CPU at 82 relative to all GPUs and CPUs, respectively.

Q: What is the suggested PSU wattage for this build?

A: The suggested PSU is 950 W, which accounts for the GPU's 575 W TDP and the CPU's 65 W TDP.

Q: Does the CPU support ECC memory?

A: No, the Core Ultra 5 225T does not support ECC memory; it supports DDR5 in dual-channel mode with 102.4 GB/s bandwidth.

Q: What are the CPU's core and thread counts?

A: The CPU has 10 cores and 10 threads, with a base clock of 2.50 GHz and a boost clock of 4.90 GHz.

Q: Is there measured FPS data for this CPU+GPU combo?

A: No, there is no measured FPS data for this exact combination; all frame rate figures are estimates from benchmark scores.

# Build Overview

This is a desktop build combining the Intel Core Ultra 5 225T with the NVIDIA GeForce RTX 5090 D. The CPU is a 10-core, 10-thread Arrow Lake processor on the Intel Socket 1851 platform, while the GPU is a Blackwell 2.0 flagship with 32 GB of GDDR7 memory. The combined percentile rank is 87, placing this system in the upper echelon of desktop configurations.

The GPU is the clear star: the 92nd percentile ranking among all GPUs puts it near the top of the graphics hierarchy, and the 3DMark Steel Nomad score of 14,326 confirms its rendering prowess. The CPU, at the 82nd percentile, is solid but not exceptional—it matches high-end mobile processors from the same generation but does not lead the desktop CPU market. The overall tier is high-end desktop, suitable for 4K gaming and GPU-accelerated workloads, but the CPU's 10-thread limit means it is not a workstation-class machine for heavily multi-threaded tasks.

# Balance and Bottleneck

The data shows a clear imbalance: the GPU is 10 percentile points higher than the CPU (92 vs. 82). This means the GPU is the stronger component relative to its peers, and in most workloads, the CPU will be the limiting factor. In gaming at 4K, the GPU is typically the bottleneck, which is the ideal scenario—the RTX 5090 D will be fully utilized, and the CPU's 10 threads can keep up with frame generation. However, at 1080p or 1440p with high refresh rates, the CPU's single-thread performance (Cinebench R23 single-core: 3,101) may not sustain frame rates that match the GPU's rendering capacity, creating a CPU bottleneck.

In productivity workloads, the bottleneck shifts depending on the task. The CPU's PassMark data compression score of 233,998 is exceptional, suggesting that compression and decompression tasks are CPU-friendly and will not bottleneck. However, the multi-thread score of 25,358 is modest, so heavily threaded tasks like video rendering will see the CPU limit overall throughput, even though the GPU accelerates specific effects. The GPU's compute scores (Geekbench OpenCL: 310,674; Vulkan: 376,915) indicate that compute-heavy tasks like machine learning inference will be GPU-bound, which is beneficial. The 32 GB VRAM ensures that large datasets fit entirely on the GPU, avoiding PCIe transfers that would otherwise bottleneck.

The power draw is another balance consideration. The GPU's 575 W TDP dominates the CPU's 65 W TDP, meaning the PSU (950 W suggested) must prioritize GPU power delivery. This is not a bottleneck per se, but it limits the headroom for additional components. Overall, the system is unbalanced toward the GPU, which is appropriate for graphics-focused builds but suboptimal for CPU-intensive workloads.

# Usage Scenarios

High-refresh gaming: At 4K resolution, the RTX 5090 D's 92nd percentile ranking and 3DMark Steel Nomad score of 14,326 should deliver frame rates well above 60 FPS in most titles, with the 32 GB VRAM preventing any memory-related stuttering. The CPU's single-core score of 3,101 in Cinebench R23 is sufficient to avoid becoming the bottleneck at this resolution.

Streaming: The GPU's tensor cores (680) and NVENC-class hardware (implied by the 50-series architecture) can handle encoding alongside rendering, but the CPU's 10 threads may struggle if the game itself is CPU-intensive. The PassMark single-thread score of 4,348 helps, but streamers should expect some frame drops in CPU-bound games.

Video editing: The GPU accelerates effects, color grading, and export encoding, with the 104.8 TFLOPS FP32 performance handling complex timelines. However, the CPU's 10 threads and 25,358 PassMark multi-thread score will slow timeline scrubbing and preview rendering in multi-track projects.

3D rendering: The 170 RT cores and 32 GB VRAM make this a strong GPU for ray-traced rendering; scenes that fit in VRAM will render quickly. CPU-based simulation or physics will be a bottleneck, as the 10 threads limit parallel computation.

Software development: The data compression score of 233,998 indicates fast build artifact compression and decompression, and the 20 MB L3 cache helps with compile workloads. The 10 threads are adequate for most build systems, but large parallel builds will take longer than on higher-core CPUs.

Student and office work: This system is massively overpowered for document editing, spreadsheets, and web browsing. The CPU's 65 W TDP keeps power consumption low, and the integrated Arc Xe-LPG Graphics 16EU handles display output when the discrete GPU is idle, but the 950 W PSU is overkill for such tasks.