When the RTX 50 collection launched, reviewers rapidly found that the hotspot temperature was being misreported in customary diagnostics instruments equivalent to HWiNFO or MSI Afterburner. Eventually, folks realized that Nvidia had outright eliminated the choice to watch hotspot temps, nevertheless it looks like the {hardware} was by no means faraway from the GPU. New testing by Brazilian restore specialist Paulo Gomes has revealed that the sensor remains to be current and readable with particular instruments.
In the video, the host reveals a Gigabyte variant of the RTX 5070 Ti that was despatched to him resulting from overheating points. Within Windows, the monitoring instruments confirmed no irregular indicators, because the “average” temperature was reported at 67 to 68 levels Celsius. However, when identified with a specialised device known as “MODS,” the hotspot temperature reached 107 levels Celsius nearly instantly underneath load.
MODS stands for Modular Diagnostics Software, and it is an inside Nvidia device used to check GPUs earlier than they hit the cabinets or throughout the RMA course of. It’s not accessible to the general public and would not work on Windows as a result of the OS retains intercepting calls from the {hardware} monitoring APIs. You want a Linux distribution that boots straight right into a command line, from the place MODS (and MATS, for reminiscence testing) can run as meant.
Some restore outlets have been recognized to get entry to MODS, equivalent to on this case, which unlocks the hidden hotspot temperature sensor on Blackwell gaming GPUs. Keep in mind that Nvidia ships much more comprehensive diagnostic utilities for its server-grade and workstation GPUs that can actively monitor all aspects of the card. It’s unknown why the company decided to keep some sensors locked out of gamers’ reach.
Perhaps we are able to infer the rationale from final yr, when Igor’s Lab examined a number of RTX 50-series GPUs and located a “hotspot issue” affecting all of them. The purpose was poor PCB manufacturing — not utilizing heavy-duty supplies to construct the PCB layers, inflicting sure components of the substrate to warmth up even when the core was comparatively cool. This was exacerbated by Nvidia’s personal tips, which informed AIBs to compensate for perfect circumstances as an alternative of worst-case situations.
Anyhow, as Paulo Gomes and his staff found, the RTX 5070 Ti’s hotspot was hitting 107 levels Celsius, and the cardboard throttled and dropped its clock speeds instantly. Nvidia mandates 107 levels Celsius because the higher restrict for RTX 50-series, so it was clear that the cardboard was slowing down to stop injury. To examine what was truly fallacious, they opened up the cardboard and located poor thermal contact between the cooler and the componentry.
The TIM (thermal interface materials) software was insufficient; the paste had accrued across the perimeter of the core whereas the middle was largely dry. The restore personnel eliminated the outdated materials and changed it with SnowDog Husky paste, which was sufficient to drop the hotspot temperatures to 100 levels Celsius. Now, it was throughout the secure working vary and now not thermal throttling underneath load.
What would’ve been a easy repair on the buyer’s finish was become a restore job solely as a result of Nvidia hid the GPU’s hotspot temperature, actually misreporting the cardboard’s inside situation. Had this RTX 5070 Ti simply run at 107 levels Celsius constantly, the silicon would put on down extremely quick, and the shopper would by no means even know why. Not to say some producers’ insistence on voiding guarantee upon breaking the GPU’s “seal,” which is an unlawful and unenforceable apply within the United States.
Follow Tom’s Hardware on Google News, or add us as a preferred source, to get our newest information, evaluation, & critiques in your feeds.