Hardware & ComponentsLOT 261122 · 3 min read
Bounce House Blower Drops Laptop GPU to 48.8°C, Taped Down for Safety
A modder pushed a bounce house blower through a laptop cooler and cut GPU temps from 84°C to 48.8°C. The FPS gains, however, ran out long before the noise did.
Key takeaways
- Bounce house blower dropped GPU temperature from 84°C to 48.8°C on an i7-10750H / RTX 2070 Max-Q laptop at 90W
- Llano V12 commercial cooler averaged 58.3°C and delivered a 19.4% average gaming performance gain over stock
- Six Arctic P12 fans hit 57.4°C; two 4,000 RPM server fans beat the full-speed Llano V12 by 2.6°C
- Time Spy scores for the top four airflow setups fell within 1% of each other, confirming diminishing returns past the thermal bottleneck

A bounce house blower brought a gaming laptop's GPU from 84°C down to 48.8°C — and the airflow was strong enough that the machine had to be taped to its stand. That is the terminal point of an escalating airflow experiment by modder Trash Bench, who set out to beat a commercial laptop cooler using nothing but moving air.
The test hardware was a laptop with an Intel Core i7-10750H and an RTX 2070 Max-Q running at a 90W power limit. The benchmark game was Cyberpunk, which climbed from 47 FPS to 56 FPS as temperatures fell. Not a bad trade for a device normally employed at children's parties.
Trash Bench was asked to review a laptop cooling stand. He declined, on the grounds that he does not do reviews, and proposed an alternative: test the cooler, then try to beat it with his own builds. The rules were simple. No opening the laptop, no sub-ambient tricks, airflow only.
How good was the commercial cooler?
The baseline device was the Llano V12 laptop cooling fan. At medium settings it averaged 58.3°C on the GPU, a 27-degree drop from the stock 84°C, and improved gaming performance by an average of 19.4%. Running the Llano V12 at 100% produced 56.7°C and considerable noise, with negligible additional performance. Loud, effective, and eventually redundant — a familiar engineering curve.
What did it take to beat it?
Trash Bench worked through three home-built challengers:
- One Noctua NF-A20, a 200mm fan in a 3D-printed duct with foam edging to stop air leaks. It averaged 66.3°C. Respectable for a single fan, though it lost decisively.
- Six Arctic P12 fans at full speed. This beat the medium-speed Llano V12 by 0.9°C at 57.4°C, and narrowly lost to the Llano at full blast.
- Two 4,000 RPM server fans. Louder still, and finally victorious, beating the 100% Llano V12 by 2.6°C.
Then came the blower designed for inflating bounce houses. Trash Bench reused the Llano V12's foam gaskets and 3D-printed an interface at the bottom of the cooler's housing to accept a flexible duct. The blast lifted the laptop off the stand, so tape sealed the gap. The result: an average GPU temperature of 48.8°C, roughly 35 degrees below stock.
Where do the returns diminish?
Not where one might hope. In Time Spy benchmarks, the blower, the full-speed Llano V12, the six Arctic P12s and the server-fan pair all landed within 1% of each other. The Noctua setup trailed the blower by 3%, though it was far quieter. The lesson is unglamorous: once the thermal bottleneck disappears, another bottleneck takes over, and no quantity of rented party equipment will move it.
For buyers, the practical takeaway is that the 19% performance gain arrives well before the industrial hardware does. The remaining 30-odd degrees of cooling bought almost nothing but bragging rights — and, presumably, a small weather system in the room.
Expect the next round of extreme-airflow experiments to chase whatever bottleneck reveals itself after this one, though the tape budget may need to grow first.
Tom's Hardware (Source)
More from Nathan Brooks
Show full bio
Senior reporter covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.
1,201 articles