SmoothFPS

Sample Pro Report

What a SmoothFPS HWiNFO analysis looks like

This is a complete example report from a real-style log — Demo rig — Core i7-13700H · RTX 4060 Laptop · 32 GB DDR5, 41-minute Cyberpunk 2077 session · 2,431 samples · 1.0s interval. Every number below is the kind of evidence your own upload produces: a physics-based bottleneck time-split, decoded throttle causes with percentages, thermal residency, and an AI fix plan. Upload your own log free →

61/ 100

Issues Detected

CPU Thermals: warningCPU Clocks: criticalGPU: okPerformance: warningMemory: okStorage: ok

Bottleneck — measured, not guessed (8 of 12 windows shown)

Per-frame GPU-Busy-gap: 38.4% of frames GPU-bound, 61.6% CPU-bound across 2,431 frames — the GPU sat waiting 4.7ms per frame on average.

0:00
gpu bound · 96 fps
5:07
gpu bound · 92 fps
10:15
cpu bound · 71 fps
15:22
cpu bound · 54 fps
20:30
cpu bound · 48 fps
25:37
cpu bound · 51 fps
30:45
mixed · 74 fps
35:52
gpu bound · 90 fps

The story is instantly visible: minutes 10–30 the session turns CPU-bound and FPS collapses from ~95 to ~50.

CPU throttle causes — decoded

Verdict: Throttled · top cause: Power Limit Exceeded (PL1)

RAPL/PBM PL1 Power Limit
34.2% · 831
Thermal Event
2.1% · 51
Max Turbo Limit (routine)
91.4% · 2,222
Turbo Attenuation (MCT) (routine)
88% · 2,139

Routine turbo-governor flags are separated from real limits — no false alarms from bookkeeping bits.

GPU limit causes

Verdict: No real limits — voltage/utilization caps are normal operation

Reliability Voltage (routine)
76.3% · 1,855
Utilization (routine)
22.9% · 557
Thermal
0.4% · 10

Thermal residency & findings

Time ≥90°C: 14.2% of the session · ≥95°C: 3.1% — judged by residency, so a hot interval can never hide inside a comfortable average.

criticalSEVERE_CLOCK_THROTTLEEffective clock below 1 GHz for 58% of under-load samples.
infoCPU_BOTTLENECKCPU-limited for 61.6% of frames — the GPU sat waiting 4.7ms per frame on average.

AI fix plan (excerpt)

  1. Your CPU is power-limit throttling (the PL1 limiter asserted for 34% of the session) while temperatures are warm but not the root cause — this is a firmware power cap, not a cooling failure.
  2. Set the Windows power mode to "Best performance" and check your laptop maker's tool (Armoury Crate / Legion Vantage …) for a Performance/Turbo profile — your current profile clamps sustained power.
  3. The mid-session FPS drop (minutes 10–30) is the CPU side: the power cap bites exactly when crowd-dense scenes push the CPU — the same minutes where the time-split turns red.
  4. Your GPU is healthy and underused during the throttled windows — fixing the CPU power cap should recover roughly 25–35% average FPS in CPU-bound scenes.

Get this for your own PC

The verdict, health score and measured CPU-vs-GPU split are free. The per-window timeline, the named and ranked causes, and every stutter traced to what asserted at that exact moment come with the report — a one-time purchase you keep forever, no subscription.

Analyze my log — freeSee what a full report costs

New to logging? 2-minute HWiNFO logging walkthrough