The measured values are really surprising and the question is what is behind them. In light of the numbers from the previous test, this result is several levels lower. Indeed, if we take the GTX 1060 as a reference point with 100% performance, then the GT 1030 reaches 53% and the MTT S80 even only 39%. It would be even slower than the already very slow GT 1030. It would be even more interesting in terms of efficiency. The GT 1030 has a TDP of 30W (and a gross performance of 1.13 TFLOPS), the GTX 1060 has 120 W (and a gross performance of 4.4 TFLOPS). The MTT S80 is known for consuming 110 W even without load, and here we reached 252 W. With less power than the GT 1030, it basically consumes an order of magnitude more electricity. The strange thing is that, despite the consumption, the temperature of the card was only 44 °C. The cooling seems to be very powerful with three coolers, but it’s still interesting.
As for the power supply, the MTT S80 does have an 8-pin connector, but it is an EPS 12V, i.e. the same one that powers the processor. The manufacturer supplies an adapter that combines two 8-pin PCIe connectors and can supply up to 300 W to the card.
