SYSTEM ANALYZER

Rate My PC: Intel Core Ultra 5 235A + NVIDIA GeForce RTX 5070 SUPER

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

86 / 100
HIGH-END

Power Build

Top 14% of systems. Excellent for 1440p Ultra or 4K High gaming.

1440p Ultra4K High

System Balance Analysis

CPU vs GPU performance ratio
Well Balanced
CPU
94%
VS
GPU
79%
PROCESSOR

Intel Core Ultra 5 235A

48,201 Benchmark Score
Top 6% Market Ranking
View Full Specs →
GRAPHICS CARD

NVIDIA GeForce RTX 5070 SUPER

2,690 Benchmark Score
Top 21% 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

Strong Performance

Excellent for 1440p gaming. Most games will run at high/ultra settings smoothly.

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

This pairing combines Intel's Core Ultra 5 235A, a 14-core Arrow Lake desktop processor sitting in the 90th percentile of all CPUs, with NVIDIA's GeForce RTX 5070 SUPER, a Blackwell 2.0 card whose single recorded benchmark result places it in the 18th percentile of all GPUs in the database. The combined build percentile is 54, and the build class is desktop. Note up front: no measured FPS rows exist for this exact CPU+GPU combination in the database, so all frame-rate discussion below is estimated from benchmark scores rather than direct game testing.

GPU Analysis

The RTX 5070 SUPER is built on TSMC's 5 nm process, packing 31,100 million transistors onto a 263 mm² die (the GB205 chip), for a transistor density of 118.3M per mm². Memory is the headline spec: 18 GB of GDDR7 on a 192-bit bus, running at 1750 MHz (28 Gbps effective) and delivering 672.0 GB/s of bandwidth. That bandwidth figure matters enormously for rendering workloads — large textures, 4K framebuffers, and scene assets all live or die by memory throughput, and 672 GB/s is the kind of number that keeps texture streaming smooth in dense scenes.

Compute resources include 6400 shading units, 200 TMUs, 80 ROPs, 50 RT cores, and 200 tensor cores. Raw output numbers: 32.15 TFLOPS FP32 (with FP16 at a 1:1 ratio, meaning no half-rate penalty), 201.0 GPixel/s pixel fill rate, and 502.4 GTexel/s texture fill rate. The 200 tensor cores are what enable NVIDIA's AI-accelerated rendering stack — denoisers in offline renderers and AI upscaling in games both lean on this hardware. The 50 RT cores handle ray-triangle intersection, and the DirectX 12 Ultimate (12_2) feature level with Vulkan 1.4 and OpenGL 4.6 support confirms current-generation API coverage.

Clocks are 2325 MHz base and 2512 MHz boost, on a 275 W board fed by a single 16-pin power connector, in a dual-slot card measuring 245 mm long, 115 mm tall, and 40 mm wide. Display output is 1x HDMI 2.1b plus 3x DisplayPort 2.1b over the PCIe 5.0 x16 bus.

The benchmark picture is more complicated. The single recorded GPU benchmark is a 3DMark Steel Nomad DX12 score of 2690, which puts the card in the 18th percentile against all GPUs in the database. The nearest rivals by that metric are a strange crowd: NVIDIA Quadro K1100M at 2664 (1% behind), NVIDIA GeForce GT 1030 at 2662 (1.1% behind), Intel Arc Pro B50 at 2660 (1.1% behind), and NVIDIA GeForce GT 440 at 2645 (1.7% behind). Take that at face value and the data shows the 5070 SUPER scoring roughly level with entry-level and legacy parts in this specific test — a result that sits oddly against the hardware specs above but is what the database records. For rendering conclusions, the honest read is this: the VRAM capacity, bandwidth, and tensor/RT hardware say "capable modern accelerator," while the one composite score says "bottom-tier relative to the full GPU field." Treat any frame-rate estimate derived from these scores as conservative.

Benchmark Performance

Starting with the CPU numbers, since that's where this build is strong. Cinebench R15: 3289 multi-core, 464 single-core. Cinebench R20: 13705 multi-core, 1934 single-core. Cinebench R23: 32633 multi-core, 4607 single-core. The R23 multi-core figure is a solid mid-range rendering result — enough to complete production render jobs without long waits, though not competitive with higher-core-count flagships. The single-core scores (464/1934/4607 across the three Cinebench generations) indicate snappy per-thread responsiveness.

Passmark tells a consistent story: 38392 multithread, 4557 single thread, 2437 physics, 88626 integer math, 118778 floating point math, 393800 data compression, 30136 data encryption, 31625 extended instructions, 49489 string sorting, and 392 on prime finding. The single-thread score of 4557 against a multithread of 38392 reflects a 14-core/14-thread design without hyperthreading — the ratio is exactly what you'd expect from that topology.

The CPU's average benchmark score is 48201, placing it in the 90th percentile of all CPUs. Its nearest rivals are remarkably tightly clustered: AMD Ryzen AI Max PRO 380 at 48171 (0.1% delta), Intel Core Ultra 5 245HX at 48287 (-0.2% delta, i.e., the rival is fractionally ahead), AMD EPYC 4345P at 48470 (-0.6%), and AMD Ryzen 9 3900 at 47918 (0.6% behind). In practical terms, this CPU trades blows within a fraction of a percent of all four — expect near-identical application performance to any of them.

The GPU's average score is 2690 (the single Steel Nomad result), in the 18th percentile. The combined build percentile of 54 lands between the two component percentiles, and it tells you the shape of this system: the processor is the star, the graphics card is the anchor dragging the composite down. Any workload that scales with the CPU will feel like a 90th-percentile machine; any workload that scales with the GPU will feel like an 18th-percentile one.

CPU Analysis

The Core Ultra 5 235A is an Arrow Lake-S desktop part from Core Ultra Series 2, manufactured on TSMC's 3 nm node, with 17,800 million transistors on a 243 mm² die. It runs 14 cores and 14 threads — no hyperthreading — with a 3.40 GHz base and 5.00 GHz boost. Cache: 192 KB L1 per core, 3 MB L2 per core, and 24 MB of shared L3. The 5.00 GHz boost is what drives that strong single-thread performance, and the 3 MB of per-core L2 is generous, which helps latency-sensitive workloads like compilation and game logic.

Memory support is DDR5 in a dual-channel configuration with 102.4 GB/s of bandwidth. ECC is not supported. Platform connectivity is PCIe Gen 5 with 20 CPU lanes, on Socket 1851 — so Gen 5 NVMe and Gen 5 graphics are both directly CPU-attached. TDP is 65 W, which is low for a 14-core desktop part; thermals should be manageable with a capable air cooler. The multiplier is locked, so no overclocking. Integrated graphics are Arc Xe-LPG with 24 EUs — enough for display output and basic desktop work, not for gaming. Launch MSRP was $269. The part is in active production, part number SRWPN.

What do the scores mean for real workloads? The R23 multi score of 32633 says this CPU handles moderate multi-threaded rendering, video encodes, and batch processing competently. The single-thread numbers (4607 R23, 4557 Passmark) say interactive work — IDE responsiveness, browser tabs, DCC application UIs — will feel fast. The 90th-percentile overall ranking with an average score of 48201 confirms it as an upper-tier mainstream CPU, and the rival clustering (all four nearest competitors within 0.6%) means buyers choosing between these parts are splitting hairs.

FAQ

Q: How many cores and threads does the Core Ultra 5 235A have?

A: 14 cores and 14 threads, with no hyperthreading. Each core gets 192 KB of L1 and 3 MB of L2 cache, plus 24 MB of shared L3.

Q: How much VRAM does the RTX 5070 SUPER include, and what type?

A: 18 GB of GDDR7 on a 192-bit bus, running at 1750 MHz (28 Gbps effective) for 672.0 GB/s of memory bandwidth.

Q: Are there measured game FPS figures for this exact CPU+GPU pairing?

A: No. The database contains no measured FPS rows for this combination, so all frame-rate discussion is estimated from the benchmark scores, not direct testing.

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

A: Its average benchmark score of 48201 sits within 0.6% of all four nearest rivals: the AMD Ryzen AI Max PRO 380 (0.1% delta), Intel Core Ultra 5 245HX (-0.2%), AMD EPYC 4345P (-0.6%), and AMD Ryzen 9 3900 (0.6% behind it). Performance differences between these chips are negligible.

Q: What power and cooling considerations does this build have?

A: The CPU is rated at 65 W TDP and the GPU at 275 W, connected via a single 16-pin power connector. The GPU is dual-slot, 245 mm long.

Q: Can the CPU be overclocked?

A: No — the multiplier is locked. The CPU boosts to 5.00 GHz from a 3.40 GHz base on its own.

Q: What memory and platform features does the CPU support?

A: DDR5 dual-channel at up to 102.4 GB/s, PCIe Gen 5 with 20 CPU lanes, and Intel Socket 1851. ECC memory is not supported.

Usage Scenarios

High-refresh gaming: Estimated from benchmark scores only (no measured FPS data exists for this pairing), the picture is mixed. The CPU's single-thread strength (4607 R23) can push high frame rates, but the GPU's 18th-percentile standing and its single Steel Nomad score of 2690 — level with cards like the GT 1030 in that metric — suggests GPU-limited frame rates in demanding titles. Expect the processor to rarely be the limiter and the graphics card to set the ceiling.

Streaming: The 14 cores handle encoding and compositing alongside gameplay, and the 393800 data compression and 88626 integer math scores support background-task throughput. The GPU's 200 tensor cores can assist with AI-enhanced encoding features where supported. Streaming is more CPU-bound than pure gaming, which favors this build's balance.

Video editing: DDR5 bandwidth of 102.4 GB/s and the R23 multi score of 32633 make timeline scrubbing and export passes workable for standard projects. GPU-accelerated effects draw on the 672 GB/s of VRAM bandwidth and 32.15 TFLOPS of FP32, but the card's low composite percentile tempers expectations for heavy GPU effect stacks.

3D rendering: Here the split matters most. CPU rendering benefits from the 90th-percentile processor — offline renders on 14 cores will complete in reasonable time. GPU/path-traced rendering leans on the 50 RT cores and 18 GB of GDDR7, which gives headroom for large scenes, but the single benchmark result cautions against expecting top-tier viewport or render performance relative to the broader GPU field.

Software development: A clear strength. The 5.00 GHz boost, 3 MB per-core L2, and 4557 Passmark single-thread score translate to fast compile times and responsive IDEs. The PCIe Gen 5, 20-lane platform supports fast NVMe storage directly off the CPU.

Student and office work: More than sufficient. The integrated Arc Xe-LPG 24EU graphics can drive displays without the discrete card, the 65 W TDP keeps the system quiet, and daily multitasking is well served by the single-thread speed and dual-channel DDR5.

Gaming Performance

To repeat the key caveat: there are no measured FPS rows for this exact combination — the database's measured FPS data for this pairing is empty, and pair ranking by FPS is null. Every statement in this section is an estimate derived from the benchmark scores, not from tested games.

Working from the data available: the CPU posts strong gaming-relevant numbers, with a Passmark physics score of 2437, single-thread score of 4557, and R23 single-core of 4607 — figures consistent with feeding a graphics card at high frame rates in CPU-bound scenarios. The GPU side is the constraint. Its sole recorded benchmark, 2690 in 3DMark Steel Nomad DX12, places it 18th percentile overall, and its nearest rivals in that metric — the Quadro K1100M, GT 1030, Arc Pro B50, and GT 440 — are all clustered within roughly 1 to 1.7% of it. If that composite score is taken as representative, frame-rate expectations should be modest at high settings: playable results at lower resolutions with settings reduced, and GPU-bound performance at ultra presets likely below what the hardware specification sheet (18 GB GDDR7, 32.15 TFLOPS) would imply. The spec sheet and the score disagree; without measured game data, the score is the more defensible anchor, and buyers should treat ultra-setting frame estimates with skepticism.

Who Should Build It

This configuration suits users whose work is CPU-dominant. Developers and students get a 90th-percentile processor with strong single-thread speed, a Gen 5 platform, and integrated graphics as a fallback — a solid foundation for a workstation or study machine. Content creators doing CPU-based video editing and rendering will extract real value from the 14 cores and the 32633 R23 multi-core score, and the 18 GB of GPU VRAM with 672 GB/s bandwidth helps with large asset working sets. Small businesses building office workstations benefit from the 65 W TDP, dual-channel DDR5, and everyday responsiveness. Gamers, by contrast, should look carefully: this is not a balanced gaming build. Anyone targeting high-refresh or high-resolution gaming as the primary use case will find the GPU side of this pairing underpowered relative to what the CPU can feed it, based strictly on the recorded percentile data.

Balance and Bottleneck

The imbalance is stark and quantified: CPU at the 90th percentile, GPU at the 18th, combined at 54. The GPU is the bottleneck in any graphics-limited workload, and it is a deep one. In games at higher settings and resolutions, the RTX 5070 SUPER's recorded performance level — with its nearest rivals in the benchmark being entry-level and legacy parts — will hold frame rates well below what the Core Ultra 5 235A could otherwise drive. There is essentially no scenario in this build where the CPU limits gaming first; the single-thread and physics scores are far ahead of the GPU's composite standing.

Flip to CPU-bound workloads — compiling, encoding, CPU rendering, compression (393800 Passmark), and integer math (88626) — and the graphics card sits mostly idle, letting the processor perform at its full 90th-percentile level. That asymmetry defines the build: a strong computing machine with a graphics card that, per the database's single recorded benchmark, does not match it. Since no measured FPS scaling data exists for this pairing, the bottleneck conclusion rests entirely on the percentile gap between the two components — but that gap is wide enough that the direction of the imbalance is not in doubt.