TECH NEWS – PowerColor is fed up with more and more people complaining about the video cards reaching 110° and is helping consumers instead.
Nvidia had a different blunder, as their GeForce RTX 4090 GPU had an overheated 12VHPWR connector that could even melt. At least thirty users reported this issue after the release of the high-end graphics card. In the case of AMD, we wrote the other day about reference cards overheating, which degrades performance, and setting the fans to maximum doesn’t help, meaning it could be a manufacturing error.
Three years ago, AMD published an article on their blog titled: “AMD Radeon Community Update: More Control Over GPU Power and Performance, Enhanced Thermal Monitoring, Maximized Performance.” When talking about maximum temperatures in AMD’s Radeon VII GPU enhanced thermal monitoring, they wrote, “Instead of setting a conservative, ‘worst case’ throttling temperature for the entire die, the Radeon RX 5700 series GPUs will continue to opportunistically and aggressively ramp clocks until any one of the many available sensors hits the ‘hotspot’ or ‘Junction’ temperature of 110 degrees Celsius. Operating at up to 110C Junction Temperature during typical gaming usage is expected and within spec. This enables the Radeon RX 5700 series GPUs to offer much higher performance and clocks out of the box while maintaining acoustic and reliability targets.”
On Reddit, the AMD Radeon RX 7900 XTX’s technical engineering lead “Kevin” claims that the company is aware of the issue and is actively investigating. They have a plan “to repro and collect serial numbers.” The current Chinese COVID situation is slowing down the process, but he says that an RMA should be fulfilled in case of a mechanical issue. Firmware upgrades would solve isolated cases where no hardware replacement is needed. Kevin says that 70° towards the edge of the GPU silicon is nominal and acceptable, but he considers 90° abnormal.
Steven from PowerColor also wrote on Reddit about his monitoring of the cases: “If you are dealing with the 110° issue and nobody else is helping you out, it doesn’t matter which AIB you bought it from. Send me an email or direct message since my emails are not working today. Also, please don’t use the chat. That s__t doesn’t work right. I am putting together a list for my AMD contacts of how many issues we are seeing. They asked if it was just a few users. I said it’s more than a handful, so I need to show some backing for my fat mouth. Send me the serial numbers. I am emailing the AMD guys with how many units are affected. PowerColor, Sapphire, Gigabyte, AMD, whatever, send me the serial numbers. Hopefully, AMD will get some news for all of us, but when I hear something of substance, I’ll make a post to provide some level of guidance on how to get this sorted out.”
From uneven cold plates to the use of graphite heat pads, there are many possible theories as a solution. For now (for anyone who would have such a video card and is experiencing such high temperatures), it is best to be patient.
Leave a Reply