PCBenchmark/Results/Intel Core Ultra 5 245K + Matrox Parhelia 128 MB browser benchmark
PC SCORE36OUT OF 100
SYSTEM GRADEE
Limited quality run 9 October 2026 · 00:02 EEST

Intel Core Ultra 5 245K+Matrox Parhelia 128 MB

A completed baseline run. Review quality and raw metrics before comparing it.

SUITEFULL SYSTEM MODESTANDARD GPU APIWebGPU ENGINE1.3.0 SCORE SCALEVALIDATED R5 DURATION00:56
Run same suite
01 / WORKLOAD BREAKDOWN

Performance by workload.

CPU and GPU use the current adaptive scale for the same workload, mode and graphics backend. When a faster validated device becomes the reference, older comparable scores are projected onto the new scale without changing their raw measurements. VRAM remains a pass/fail diagnostic.

CPU59

Processor throughput

Single thread
789.50 Mops/s
Multi thread
6,820.71 Mops/s
Compression proxy
1,006.7 MB/s
GPU1

Graphics throughput

Estimated shader work
437.0 G shader ops/s
Render throughput
48.9 FPS
Frame time p95
33.40 ms
MEMORY64

System memory

Effective bandwidth
22.29 GB/s
Latency proxy
169.3 ns
Browser RAM class
8 GB class
VRAMPASS

GPU-memory diagnostic

Transfer bandwidth
2.44 GB/s
Tested allocation
384 MB
Validation errors
0 errors
02 / RELATIVE PERFORMANCE

What sits above and below this hardware.

Each ladder uses one consistent source. Validated PCBenchmark medians take priority when the matching workload cohort has enough approved runs; until then, the report shows a clearly marked catalog estimate.

CPU POSITION

Intel Core Ultra 5 245K comparison ladder

ESTIMATE

Specification-based estimateApproximate browser CPU capacity derived from official core, thread and clock specifications for Arrow Lake desktop processors. It is not a measured benchmark.

  1. Intel Core Ultra 7 265KArrow Lake · 20 cores / 20 threads · 2024
    +15.8%higher estimate
  2. Intel Core Ultra 5 245KArrow Lake · 14 cores / 14 threads · 2024
    BASECATALOG REFERENCE

No slower peer is available in this comparison set.

GPU POSITION

Matrox Parhelia 128 MB comparison ladder

ESTIMATE

Comparison unavailable
The catalog does not contain a trustworthy theoretical compute reference for this GPU.

03 / STABILITY81.5%

Sustained-load signal

This signal summarizes frame delivery during the timed graphics phase. It can reveal an inconsistent run, but it is not a temperature sensor.

COMPLETEDYES SAMPLE WINDOW56 SEC DROPPED FRAMES18.52%

Why the run quality was reduced

  • Many preview frames were dropped, a sign of a contended or throttled device.
  • The browser exposed fewer threads than the selected CPU has, so the multi-core test could not use all of them.
  • The browser ran the GPU test on a different adapter than the selected graphics card, often the integrated GPU.
04 / ENVIRONMENTRUN CONTEXT

What the browser disclosed

Selected suite
Full system
Platform family
Windows
Logical processors
10
Browser RAM class
8 GB · rounded/capped
Graphics API
WebGPU
Browser
Chrome 152 · Blink engine
Timer resolution
0.100 ms
Memory working set
256 MB
Run quality
Limited · 41/100
Performance index
35.7 · uncapped, 100 = reference
Leaderboard
Ranked
Score fingerprint
adaptive-v2.0 / 1.3.0
Current adaptive scale
Validated revision 5 / 6 October 2026
i
BROWSER BENCHMARK, NOT HARDWARE ATTESTATION

Read this result in its proper context.

It measures this browser’s execution path under the conditions of this run. Cooling, power mode, drivers, background processes and browser updates can change the outcome. CPU/GPU names may be supplied manually or matched to the catalog from a browser hint. Compare only matching suite, mode, engine version and graphics backend. NOT RUN categories were not measured and are not zero-performance results.

Read methodology & limitations ↗
05 / RAW RESULT DATAInspect JSON +
{
    "test_suite": "all",
    "test_mode": "standard",
    "duration_ms": 56112,
    "system": {
        "cpu_name": "Intel Core Ultra 5 245K",
        "gpu_name": "Matrox Parhelia 128 MB",
        "logical_cores": 10,
        "memory_gb": 8,
        "platform": "windows",
        "gpu_api": "webgpu",
        "browser_engine": "blink",
        "browser_name": "chrome",
        "browser_version": 152,
        "timer_resolution_ms": 0.1,
        "cross_origin_isolated": false,
        "on_battery": false,
        "gpu_vendor": "intel",
        "gpu_architecture": "xe-lpg",
        "gpu_fallback_adapter": false,
        "processor_id": 30,
        "graphics_card_id": 348
    },
    "metrics": {
        "stability": {
            "duration_seconds": 56,
            "completed": true,
            "dropped_frames_percent": 18.515
        },
        "cpu": {
            "single_core_ops_per_second": 789504788.534,
            "multi_core_ops_per_second": 6820713438.332,
            "compression_mb_per_second": 1006.73
        },
        "memory": {
            "bandwidth_gbps": 22.286,
            "latency_ns": 169.25,
            "working_set_mb": 256
        },
        "gpu": {
            "frames_per_second": 48.925,
            "compute_gflops": 436.965,
            "frame_time_p95_ms": 33.4
        },
        "vram": {
            "tested_mb": 384,
            "bandwidth_gbps": 2.444,
            "errors": 0
        }
    },
    "raw_points": {
        "cpu": 35.9525068,
        "gpu": 0.3641375,
        "memory": 0.89241394
    },
    "quality_flags": [
        "dropped_frames",
        "thread_cap",
        "gpu_vendor_mismatch"
    ],
    "original_scores": {
        "cpu": 59,
        "gpu": 1,
        "memory": 64,
        "vram": 0,
        "overall": 36,
        "grade": "E"
    },
    "current_scores": {
        "cpu": 59,
        "gpu": 1,
        "memory": 64,
        "vram": 0,
        "overall": 36,
        "performance_index": 35.7,
        "leaderboard_outlier": false,
        "grade": "E",
        "snapshot_revision": 5
    },
    "runner_version": "1.3.0",
    "workload_signature": "browser-1.3",
    "scoring_version": "adaptive-v2.0"
}
SAME SUITE, FRESH CONDITIONS

Think this system can do better?

Close heavy apps, connect power, keep the tab visible and run the same Full system suite and Standard mode again.

Run it again