Radeon RX 9070 XT cards use GDDR6 from multiple vendors. The biggest differences between Samsung and Hynix appear in temperature and power—Samsung runs notably cooler. Earlier tests with automatic fan control confirmed this. Today I’m adding a temperature comparison at equal fan speeds and, alongside internal sensor data, surface temperatures measured on the chips. That gap is much smaller than expected.
Operating characteristics comparison: Hynix vs Samsung
In the next set, the fan speeds on the RX 9070 XT with Hynix chips are fixed to values similar to those the card with Samsung chips had after warming up. Doing the opposite—lowering the fan speeds on the card with Samsung chips—wouldn’t be possible without reducing the power limit, as the hotspot would get too high.
The first graph shows frame rates. The performance difference is again more likely a measurement error.
The card with Hynix chips now fares slightly better with clock speeds under improved cooling. But this is more likely due to chip-to-chip variation.
The average CPU utilization and its temperature during the test are shown again for a complete picture.
The fan speeds are almost identical.
The average GPU temperature is now the same, but the hotspot temperature differs significantly. The difference on the memory is smaller; the Hynix chips read 88–89 °C according to the sensors, while the Samsung chips are only at 74 °C.
However, a different picture emerges from the temperatures recorded by the external sensors. The temperature on the chip packages under the smaller part of the heatsink differs by only 1 °C, under the better-cooled part the difference is about 3 °C, and underneath the memory, the difference is less than 1 °C. So externally, it’s far less dramatic than with the internal sensors.
The following values are from the card’s monitoring. The difference in memory chip power draw should only be a few watts in favor of the Samsung chips.
The total power draw of the card and the system again does not differ significantly.
⠀
