It is true and AMD "fixed" it with a driver hack. https://www.custompcreview.com/news/amd-aims-reduce-rx480-power-consumption-upcoming-graphics-driver/30719/ h

It is true and AMD "fixed" it with a driver hack. https://www.custompcreview.com/news/amd-aims-reduce-rx480-power-consumption-upcoming-graphics-driver/30719/ https://www.custompcreview.com/news/amd-aims-reduce-rx480-po... The driver do two things to "fix" the problem: 1. Draw more power from the cable (it is what the card should have done in first place). 2. They put back on the card the behaviour the 380X have, that I am trying to get rid of, of clamping down hard on the TDP and throttling heavily. And yes, the claims of this making the card faster ARE true, and also applies to older cards, AMD cards draw so much power and make so much heat, that if you UNDERCLOCK them, they can get faster, because sometimes the stable speeds with less throttling is a bigger benefit than the amount that you underclocked the card. That said, last I checked (this was 2 months ago), the driver changes were Windows only, some guys mining bitcoin on Linux didn't noticed and happily melted their riser cables. I really don't like nVidia behaviour as company, pulling shady tactics left and right, but AMD frankly shocked me with their shoddy engineering. (also see: recently Linux kernel devs REFUSED AMDGPU driver patch... AMD was warning in February that the patch they were attempting to do was crap, and they still went ahead with it, there was even some passive-agressive personal insults exchanged on official Linux kernel mailing list).