The thing that AMD cites for these temperatures being recorded is what they describe as ‘Enhanced thermal monitoring’. This is to say that the RX 5700s have multiple temperature sensors, while cards in the past would typically only have one temperature sensor. The post says that throttling or other responses to heat in older cards would have to be based off a worst case scenario from the limited dataset available, in other words if the temperature is high at the sensor, the designers would assume things were amiss. AMD claims that the network of sensors on the RX 5700s means that the cards can react more precisely, so that the card can run hotter without throttling until it absolutely has to.
The exact phrasing of the AMD quote on this subject is interesting as much for its pugnacious tone as the content.
‘Instead of setting a conservative, ‘worst case’ throttling temperature for the entire die, the RadeonTM 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 RadeonTM RX 5700 series GPUs to offer much higher performance and clocks out of the box, while maintaining acoustic and reliability targets.’
The prioritisation of pursuing faster speeds, even when the temperatures being talked about are breaking into triple digits, suggests that for all the technology that has gone into this generation of cards AMD is not shy about making them work them extremely hard. And that’s fine by us.