My ASUS TUF RTX 4080 SUPER OC has recently become much hotter and louder despite running at factory settings in a TUF GT502 V2 case. It previously stayed at or below about 65°C during gaming, but current HWiNFO readings show GPU core temperatures around 80–84°C and hotspot temperatures between 105–111°C. Memory temperatures remain around 66–70°C, while fan speeds reach roughly 2,400–3,300 RPM. Thermal limiting was reported during Hogwarts Legacy and Cyberpunk 2077, and the results were reproduced in several games plus a short FurMark 2 test at 2560×1440. ASUS has offered warranty service but said they would not install a phase-change thermal pad such as PTM7950, following their standard repair procedure. They have not specified which conventional thermal paste they would use. I am comfortable working on electronics, but opening the card could void the remaining warranty. I am considering either sending it to ASUS or replacing the GPU thermal interface material myself with genuine Honeywell PTM7950. I am also unsure about the correct thickness, whether replacing the material could affect contact with the existing VRAM and VRM pads, how difficult PTM7950 is to install, and whether high-quality paste replaced periodically would be a reasonable alternative. Genuine PTM7950 is difficult to obtain locally, so reliable sourcing and the correct sheet size are also concerns.
4 Answers
Be very careful not to damage the original memory and VRM pads. Photograph their locations, measure them if possible, and have compatible replacements available before opening the cooler. Manufacturers may use several different thicknesses in different areas, so a generic pad kit is not necessarily suitable. Ideally, get the exact specifications for this board revision from ASUS or a detailed teardown of the same model. For the GPU die itself, PTM7950 is generally a stronger long-term option than ordinary paste, but a quality paste applied correctly is still a sensible fallback if warranty preservation matters more.
PTM7950 can work extremely well on GPUs and often lasts longer than ordinary paste because it does not pump out as easily. It needs heat cycles before reaching its best performance, so do not judge it immediately after installation. Keep the material cold for a short time before cutting and fitting it, and handle it with gloves or finger cots and clean tweezers. At room temperature it becomes much more difficult to position cleanly. An 80 × 40 mm sheet should provide plenty of material for the GPU die, assuming the product is genuine.
If the card is still under warranty, sending it to ASUS is the safest choice, even if their standard repair uses conventional paste. A 105–111°C hotspot with thermal limiting is worth having documented and inspected by the manufacturer. DIY work may produce better temperatures, but it can complicate or eliminate your warranty coverage. If the warranty has expired, follow a teardown guide for this exact card model before attempting PTM7950.
That is my main hesitation. I am comfortable doing the repair, but I do not know whether ASUS would reject future warranty work as soon as they see that the card was opened.
A successful PCM or thermal-pad replacement can make a dramatic difference; one owner reported reducing core temperature from about 84°C to 68°C and hotspot temperature from 105°C to 84°C. However, results depend on mounting pressure, pad thickness, cooler condition, and the original defect. Since your temperatures used to be much lower, an uneven mount, degraded interface, or cooler-pressure problem may be involved, which is another reason to let ASUS inspect it while coverage remains active.
The thickness is critical. Do not assume a 0.25 mm sheet is automatically correct for every part of the card. The GPU interface and the surrounding VRAM/VRM pads serve different purposes, and changing the wrong pad thickness can prevent the heatsink from making proper die contact.