IT EN
System

GPU, CPU, overclocking and undervolting

How SkillFishOS controls the BC-250's clocks, voltages and temperatures — with the real numbers measured on the hardware.

On a normal APU you tune clocks through the amdgpu sysfs. On the BC-250 that doesn’t work: control goes through the SMU (System Management Unit) and needs dedicated tools. SkillFishOS bundles them all, pre-configured with safe profiles and a thermal-protection system.

⚠️ Silicon lottery. Every number on this page is measured on our BC-250. Each card is different: one may take a deeper undervolt, another less. That’s why SkillFishOS always boots in the Stock profile and lets you climb using the Tuner, which validates each preset on your card with an automatic test and rollback.

The three profiles

The Tuner exposes four presets. The ISO boots in Stock; the others are one click away after the test.

ProfileCPUGPUNotes
Stock (ISO default)3500 MHz1500 MHzMaximum compatibility on any BC-250
Performance3700 MHz · ~1106 mV2000 MHzBalanced and undervolted
Turbo3900 MHz · ~1199 mV2230 MHzHigh boost, validated under the 85 °C cap
Crazy4.0 GHz · ~1224 mV2230 MHzValidated maximum (~83 °C under stress)

All profiles honour the same 85 °C thermal cap and keep the fan on auto.

The GPU SMU governor

GPU clocks are driven by the cyan-skillfish-governor (written in Rust), a system service configured in /etc/cyan-skillfish-governor/config.toml. It defines frequency/voltage safe-points: 350 MHz / 700 mV at idle, and the profile value under load (e.g. 1500/900 in Stock, 2230/1000 in Turbo).

The standard amdgpu sysfs (power_dpm_force_performance_level, pp_dpm_sclk) does not control the BC-250 — only the SMU governor does. The GPU only ramps to its boost clock under real graphics saturation.

CPU overclocking and undervolting

The CPU (6× Zen 2 “Oberon” cores) is handled by a one-shot service bc250-smu-oc.service that applies the values from /etc/bc250-smu-oc.conf via the bc250_smu_oc project. It shows as inactive after applying — that’s normal (it’s one-shot).

What we measured pushing our card:

  • 3700 MHz (Performance preset) undervolted to ~1106 mV (scale −16);
  • 3900 MHz (Turbo preset) at ~1199 mV (scale −24);
  • 4.0 GHz (Crazy preset) validated at ~1224 mV (scale −36) for 120 s of sustained stress, peaking at 83 °C — the usable maximum on this sample;
  • Hard Vid ceiling: 1.325 V (never exceeded).

Undervolting isn’t about “pushing” — it’s about doing the same work with less heat and less power: at a given frequency, lowering the voltage until it stays stable drops the temperature and leaves thermal headroom for the rest of the APU.

CPU↔GPU thermal coupling

CPU and GPU share the same die and the same power budget. Under mixed load (a demanding game: CPU + GPU together) the APU self-protects and the CPU spontaneously drops to ~3450 MHz to stay within budget and under 85 °C. This is not a defect: it’s the chip protecting itself by shedding the least-useful clocks. For the same reason, a CPU undervolt leaves more thermal “room” for the GPU, and vice-versa.

The 40 Compute Units — live

The BC-250 has 40 CUs (20 WGP, 1 WGP = 2 CU), but the driver enables 24 by default. SkillFishOS routes them up to 40 at runtime, no reboot: the system boots at the driver baseline (24 CU) and a service brings it to 40 at startup; from the Tuner you adjust the count live with a grid of squares and 24/32/40 presets. The first 24 CUs are driver-locked and always on.

With all 40 CUs enabled the GPU measures 11385 GFLOPS FP32 (vkpeak) cold, versus ~6141 for a 24-CU baseline: +85%. Under sustained stress (hot) it settles around 10214 GFLOPS. Measured memory bandwidth (clpeak) is ~350–367 GB/s.

🔬 Silicon lottery. On salvaged/“discard” chips some CUs may be marginal. The Tuner has a “CU test” that stresses each pair and flags GPU faults/hangs, so you can confirm your chip sustains all 40 CUs. (Mechanism via umr, writing the WGP masks — credit to bc250-cu-live-manager, clean-room reimplementation.)

Thermal protection — the 85 °C cap

The thermal ceiling is 85 °C, enforced on two levels:

  1. SMU side: the max_temperature value in the config makes the chip reduce clocks before crossing 85 °C (avoiding hard throttling);
  2. system side: a thermal-guard watchdog that, if temperature exceeds the cap, steps clocks down 100 MHz at a time until it’s back in range.

Things to know about the stock cooler (see also BC-250 hardware for 3D-printable cases and recommended fans):

  • the stock heatsink is marginal: “back-to-back” benchmark comparisons are skewed by heat-soak — let the card cool for a few minutes between runs;
  • only the GPU edge sensor exists; there is no VRAM temperature sensor;
  • memory bandwidth is healthy but mclk is not adjustable.

A real case: CPU-bound games

Some titles — like Black Myth: Wukong in gameplay — are CPU/draw-call bound: FPS barely depend on resolution or GPU clock. There, CPU overclocking and good cooling help instead. For upscaling, FSR 4 is not available (it’s RDNA 4 hardware); use gamescope (FSR1/NIS) or per-game OptiScaler.

When the workload is GPU-bound (e.g. the Wukong benchmark flythrough), the clock matters: in the Tuner you can switch the governor to “Performance”, which holds the GPU at its top safe-point under load (it still idles to 350 MHz). Measured on the Wukong benchmark: 100 → 111 FPS average (+11%), 92 → 102 on the slowest frames. For safety the Tuner caps the GPU at 2200 MHz @ 1000 mV (the stable maximum on stock cooling) with a multi-point voltage curve — pushing 2230 MHz at 1000 mV is undervolted and can hard-freeze the machine.

All of this, without a terminal

Clocks, undervolt, fan and Compute Units are tuned from the Tuner GUI, with the four ready presets and automatic test + rollback if your card can’t hold a value — see Native apps. It’s the recommended way: start at Stock, move to Performance, try Turbo or Crazy, and the Tuner validates everything on your BC-250.

Sources