Volumeshader_bm GPU Test
Volumeshader_bm GPU test: WebGL fractal stress test, right in your browser.
WebGL fractal stress test · rendered at a fixed 1920×1080 so scores stay comparable across devices.
Why this tester
Nothing to install
Pure WebGL — open the page and run the heavy volumeshader_bm fractal stress test. No downloads, no disk space, no software.
Live FPS monitor
A real-time frame-rate counter sits in the top corner of the test view, showing exactly how the GPU holds up under load.
Scale the load to the limit
Customize the fractal parameters to ramp the load from a light warm-up to a worst-case power draw, probing the limits of cooling and PSU headroom.
How to start the test
- 01 · Start the test
Click "Start test" to open the parameter page. On a low-end GPU, do not raise the defaults carelessly.
- 02 · Adjust view & parameters
In the fullscreen view, drag with the left button to rotate, right button to pan, wheel to zoom; double-click for the quick settings panel.
- 03 · Watch the FPS, then exit
Keep an eye on the live FPS counter. When you are done, press "Exit test" to return to the page.
How to read your score
This pressure test rewards stability rather than a single high FPS: hold the same parameters for a few minutes and a flat, stable curve means your cooling and PSU are up to the load; a steadily falling frame-rate usually means you hit the thermal limit and downclocked. As a rough reference, a mainstream card sits around 30–60 FPS on the default fractal — the more high-end, the closer it gets to full rate. If you are well below that peer level, check drivers, aging thermal paste and case airflow before blaming the hardware.
- Load ceiling
- Maxing out fractal complexity matches a classic burn-in tool. Confirm PSU headroom and case airflow before running.
- A / B benchmark
- Run the same parameters before and after a driver update or a repaste — performance changes become obvious.
Parameter guide
The parameter page only supports numeric edits. Defaults represent a standard load — here is what each value does:
- Iterations: for(int i=0;i<5;i++)
- Fractal complexity and detail. Suggested i<3 (easy), i<5 (standard), i<8 (heavy), i<12 (extreme).
- Angle scale: *8.0
- Angle transform magnitude. Suggested *4.0 (mild), *8.0 (standard), *12.0 (intense), *16.0 (extreme).
- Power exponent: pow(b,8.0)
- Scaling strength and recursion depth. Suggested pow(b,4.0) light, pow(b,8.0) standard, pow(b,12.0) strong, pow(b,16.0) extreme.
- Break threshold: if(b>6.0)
- Boundary of the fractal. Suggested b>4.0 small, b>6.0 standard, b>8.0 large, b>10.0 extra large.
- Return value: return 4.0-
- Surface density and visibility. Suggested 2.0- dense, 4.0- standard, 6.0- sparse, 8.0- very sparse.
GPU terms explained
- WebGL
- The browser-native 3D graphics API. Volumeshader_bm is a WebGL renderer — no plugins needed.
- FPS
- Frames per second. A weak GPU drops FPS sharply and the animation visibly stutters — the most direct number for comparisons.
- Fractal rendering
- Math-generated infinite detail computed in real time. The huge floating-point and ray-marching workload is a brutal GPU workout.
- Core temperature & power
- A stress test pushes the GPU to full load — power draw maxes out and temperatures climb, the perfect probe for cooling and PSU headroom.
Use cases
Second-hand GPU check
Run it for 15 minutes and watch for artifacts, driver crashes or black-screen reboots — classic mining-card symptoms.
Cooling limit test
Raise the load to reach thermal equilibrium quickly and judge your case airflow and cooler headroom.
Phone SoC comparisons
A favorite with reviewers to compare peak performance and thermal throttling across phones.
Overclock stability
After an OC tune, verify clocks and voltage quickly — faster than a full 3DMark loop.
Troubleshooting guide
Black screen / reboot mid-test
Usually an over-current protection trip (undersized PSU) or core overheating. Check PSU wattage and clean dust.
Snow, mosaic or colored stripes
Classic VRAM failure or a cracked BGA — common on ex-mining cards. Stop testing and contact support.
Render freezes, system alive
A driver crash: unstable clocks or a driver bug. Reset clocks and cleanly reinstall drivers with DDU.
FPS absurdly low
Check the monitor cable is plugged into the graphics card (not the motherboard iGPU), and that the browser uses the dGPU on laptops.
Frequently asked questions
Can the test damage my GPU?
Generally no. Modern cards throttle or shut down on over-temperature and respect power limits. Still, extreme parameters plus poor cooling can accelerate aging — use judgment.
Why do some call this the "toadstool" (毒蘑菇) test?
The original project is volumeshader_bm. The fractal looks mushroom-like and the load is "poisonous" for GPUs, so Chinese users gave it the informal nickname 毒蘑菇 (toadstool) — the official name is Volumeshader_bm GPU Test.
Can I run it on a phone?
Yes, if the browser supports WebGL — but it will heat the SoC heavily, and older phones may freeze or reboot.
The render is completely black — why?
Check that hardware acceleration is enabled in your browser settings; WebGL cannot render properly without it.
Can I lower the load via parameters?
Yes. In the parameter page, change the iteration count in for(int i=0;i<5;i++) from 5 to 2 or 3 for a much lighter load — or raise it for the extreme version.
What a graphics card is
Form factor: discrete vs. integrated
- Discrete GPU (dGPU): has its own VRAM and cooling, delivers the performance that gaming, design and AI workloads need.
- Integrated GPU (iGPU): built into the CPU and shares system memory. Weaker but lower-power and cheap — fine for office work and light entertainment.
By brand camp
- NVIDIA (GeForce): strong driver and game ecosystem, an edge in ray tracing and large-model AI.
- AMD (Radeon): known for value and generous VRAM.
- The two camps differ in architecture, software ecosystem and exclusive features.
Connection and the core metrics
Connection & power
- PCIe slot: the card hooks into the motherboard via PCIe; the PCIe version (4.0, 5.0) and lane width (x16, x8) shape data throughput.
- Power cables: power comes from the slot or extra PCIe cables; high-end cards often need multiple 8-pin / 12V-2x6 connectors.
Core metrics
- FPS: frames rendered per second, the most direct readout of game smoothness.
- VRAM: stores textures and geometry; more capacity (8/12/16/24 GB) survives higher resolution and quality. Running short causes stutter and drops.
Vendor "theoretical" numbers often flatter. Measuring real FPS, VRAM use and temperature across workloads is the directest way to judge a card.
Inside the card
From the inside out, each part has a job
- GPU die — “the brain”: does all the geometry and parallel compute. Architecture (Ada Lovelace, RDNA) and node decide efficiency and headroom.
- VRAM chips — “the warehouse”: type (GDDR6, GDDR6X) and bus width decide throughput.
- Cooler — “the AC”: fan, fin stack, vapor chamber / heatpipes carry heat away and decide how long the card stays at high clocks.
Size and fit
- PCB length, thickness (how many slots) and the power connectors must match your case and PSU — or it does not fit or starves for power.
Architecture and VRAM
GPU core and generations
- Each new architecture improves node, stream processors (cores) and efficiency, then splits the same die into high-to-entry SKU zoomed.
- RT cores: dedicated ray-tracing horsepower.
- Tensor cores: run DLSS, FSR and other AI upscaling / ML workloads.
Rating VRAM
- Capacity: how much geometry and texture it can hold; too little causes stutter and late-loading textures.
- Bus width / bandwidth: the “lane width” for data; wider and faster is better.
- Type: GDDR6X pushes higher transfer rates than regular GDDR6.
Real VRAM usage, clock swings and thermal throttling are invisible to the eye — monitor them to find your card's true behavior instead of trusting the spec sheet.
How a GPU works, and its kinds
How it works
- The CPU feeds render commands to the GPU, which chews a huge amount of pixels, vertices and textures in parallel stream processors, then pushes through rasterization, ray tracing and AI upscaling to the display.
By target market
- Gaming cards: consumer segment, value for money.
- Workstation / pro cards: stable drivers and the studio, modeling and editing on a certified stack.
- AI / datacenter cards: aimed at training and inference, tuned for capacity and parallel compute.
Thermals and power
Thermal wall and throttling
- TDP: the cooling requirement under full load; higher means stricter demands on the PSU and case airflow.
- When the die passes the vendor threshold, the card drops clocks to protect itself — a well-cooled card can hold markedly higher FPS than an identical hot one.
Overclocking
- Manually raising core / memory clocks or voltage squeezes out more performance within cooling and power limits, but adds heat, noise and wear headroom.
Common faults, self-check
Typical symptoms
- Artifacts / black-screen reboot: usually overheating, unstable power, or a failing VRAM.
- FPS unusually low: maybe a stale driver, background load, or it never hit its power budget.
- Fan noise or no spin: many cards have a semi-mode that deliberately stops the fan at idle — that is normal.
Self-check approach
- Run a stress + benchmark loop and watch temperature, clock curve, VRAM use and FPS stability to tell cooling, power, driver and hardware apart quickly.
Daily care
Clear dust
Blow out the fans and fins with compressed air now and then, so heat does not build up.
Air flow
Check the case airflow after long runs and add a case fan if the interior is crowded.
Seat the cables
Make sure the power cables are fully seated — a loose one can cause flickers and freezes.
Update drivers
Keeping drivers current fixes known bugs and usually brings game-ready optimizations.
Buying guide
Use case & game type
1080P daily gaming is fine on a mid-range card; 2K high-quality wants mid-to-high compute; 4K or pro creation needs a flagship with plenty of VRAM. Also weigh heavy ray tracing and whether DLSS/FSR-style upscaling matters — both steer the model choice. For a well-rounded decision, cross-reference the graphics card leaderboards below.
Desktop GPUs Laptop GPUs Blender rank Geekbench rank Time Spy rank Trending GPUs
VRAM size
8 GB is enough for 1080p; at 1440p+, or with memory-hungry AAA titles, choose 12 GB plus to avoid a bottleneck.
Cooling & power
Confirm the PSU holds the peak draw and the case fits the card — two easy "gotcha" details before you build.
Brands & budget
The chips come from NVIDIA or AMD; boards are made by ASUS, MSI, Gigabyte, Colorful and more. Entry fits 1080p, flagship for 4K and pro work. Bench before you buy.
Meet Volumeshader_bm
What it is
- A WebGL-based in-browser rendering stress test. The on-screen organic, mushroom-shaped fractal, spinning in bright colors, prompted Chinese fans to nickname it “毒蘑菇” (toadstool), which is why its official name is Volumeshader_bm.
- Popularized by short video platforms, it became a quick way to check a phone or PC's graphics power.
Zero install, zero skill
- Nothing to download; open the browser and go. Any device with WebGL — phone, tablet or desktop — can use it.
- That convenience sets it apart from most professional GPU benchmarks.
Format and how it scores
What you see
- A colorful, draggable, spinning shape in the middle, whose rotation frame-rate is computed live to gauge the machine.
Core metric: FPS and smoothness
- Whether the shape spins and drags smoothly, without stutter or frame drops, is the most direct judge of rendering ability.
WebGL tuning varies by browser and OS, so the same device can score differently in different environments. Verify across scenarios with a tool instead of a single vibe — more objective.
How it rates performance
Volume shading
- It runs a volume-shader render workload, stressing the card on complex light, and shadow. Volume shading needs a lot of parallel math for light travel, scattering and layered materials — a perfect stress vector.
Dragging re-renders live
- The GPU edits light angle, shadow and material in real time while you drag, and tries to stay at high FPS.
- Fast cards keep it fluid; the slower drop frames, tear, or briefly freeze.
Its limit-load value
Stacking the GPU full
- The test keeps the GPU fully loaded — it can hang a weak machine and visibly heat a device over a long run.
- That extreme pressure is exactly the point: paired with a monitoring tool, frame time, temperature and throttling expose hidden cooling and stability problems faster than daily use.
Where it fits
Across devices
- Covers just about any modern browser: Android, iPhone, Windows or tablet. Very portable.
- Mobiles throttle on heat and battery — it may look weaker and run hot, so mind thermals and power on phones.
Two use directions
- Casual fun: a feel for your device's graphics muscle.
- Quick self-check: whether new hardware is underperforming, or you suspect thermal throttling — a quick pressure gauge.
Notes on use and risk
Before you run
- Close other GPU-heavy apps — tabs videos and other load skew the result.
- Close out of background memory to get a cleaner read.
- On a weak or old device, expect potential stutter — that's low frame rate, not a bug. Nothing is guaranteed, so run only if you can risk a crash.
While it runs
- Save any open document first, in case the high load freezes the machine.
- If it gets very hot or the fan screams, stop, let it cool, and decide whether to continue.
- Before blaming the part, cross-check with a monitor whether the fault is thermal, driver, background load or the hardware itself.
Turning the result into a useful signal
Set realistic expectations
- Treat it as a quick first impression, not a replacement for a professional grading.
- A real purchase decision also weighs budget, compatibility, brand and the specific application — not just a single score.
Pair with other tools
- First use Volumeshader_bm for a quick stress and a first judgment. If something looks clearly off — heavy stutter, rapid overheating, unexpected reboots — reach for a deeper benchmark for exact FPS, the thermal curve and stability, forming a fuller picture.