Basics PSU System

The battle of graphics card against power supply – power consumption and peak loads demystified

Let’s start with the card that also appeared at the power supply manufacturer and in various requests to speak, e.g. at Reddit, namely the AMD Radeon RX Vega. I even measured two cards and in the end I fairly decided on the “worse” version with the slightly higher peak load (AMD RX Vega64 reference card).

The also tested Gigabyte RX Vega64 was only marginally “better”. But all this does not influence the result anyway. A RX Vega56 was, by the way, once again far below the values listed below!

Measurements under maximum gaming load

The game with the highest load peaks and also the longest load conditions in the maximum range is, who is surprised, Witcher 3 in Ultra-HD. This is what emerged after the first comparison measurement with 10 games. The average power consumption over a measuring room of one second is approx. 250 watts, which is fully in line with the norm.

But now we come to the measurements. What the end customer understands best is the power consumption, i.e. what effectively goes over the counter in energetic form. From a total of 100 automatically logged measurements for gaming (the other 100 are for the stress test) over 60 ms each (6000 measuring points in intervals of 10 each μs) I filtered out the ones with the highest single load >= 1 ms and only cut out a time window of 20 ms with 2000 measuring points (zoom).

For better orientation, I have set the auxiliary lines so that they show a 1 ms grid. One box corresponds to one millisecond. The longest continuous peak load here was about 370 watts over 1.7 ms. The highest peak load in this measurement was approx. 420 watts in 0.3 ms. So a RX Vega64 doesn’t generate more as long as the power supply is constant and the voltage remains constant. The yellow curve represents the total power consumption of the card on the two 12V rails.

 

Let us therefore take a closer look at the currents and voltages. If you exclude the voltage fluctuations from the curve calculation, you can see that the currents flow much more “evenly” (graphic below). The jittering thus comes from the voltage multiplied in order to show the power consumption (graphic above). The supply voltage fluctuates between 11.8 and 12.3 volts, which is a really good value considering the high resolution.

Load peaks of more than 600 watts / 50 amps over 10 ms belong therefore already something into the country of the imagination, because more than 40 amps with a RX Vega64 (not Vega56) was in none of the measurements (here approx. 38 amps). The measurement published by the power supply manufacturer with 102 Ampere for 2 smaller cards should never have been done directly on the graphics card itself. Then the curves would look different. The actual construction site is therefore not the AMD graphics card.

Measurements during the stress test

The reason why graphics cards get hotter in the stress test, even though the power consumption as an average value often does not even increase significantly over a long second, is due to the type of loads generated. This can be imagined as touching a hot surface with a finger for a short time. A periodic, very short touch is much easier to survive than a permanent touch on a not so hot surface!

The same applies to the loads. I still measure about 250 watts as average value, but the intervals of the load peaks are much longer! Whereby the longest interval with 237 watts over 7.7 ms is even below the TBP (Typical Board Power)! The highest interval here is about 8 ms and sometimes up to 338 watts in peak, but then drops to 242 watts within this interval. Here the arbitrator of Powertune restrictively intervenes already after approx. 8 ms and cuts the power consumption.

 

The tensions even fluctuate somewhat more strongly. By the way, I never measure more than 28 amperes for the maximum current over all the individual tests, even if I shift the measurement intervals to under 1 μs to the nanosecond range.

 

Interim conclusion

I would really like to know where the over 50 amps for the much weaker RX Vega56 were really measured and in what application. I talked to board developers from the board partners and AMD themselves, not just now and today. Such extreme outliers and peak values have never been observed in such long intervals, which makes me – I really don’t want to accuse the engineers of anything without exact knowledge of the circumstances – doubt the result.

What you have to consider when choosing and dimensioning a power supply, however, is a good and practical secondary side, which should produce at least 400 watts over 2 to 3 ms and approx. 350 watts over 10 ms, in order to be completely safe. But then the 550 Watt power supply should have been enough. So I will have to say something about it in the summary after we have also looked at the other test candidates.

 

Danke für die Spende



Du fandest, der Beitrag war interessant und möchtest uns unterstützen? Klasse!

Hier erfährst Du, wie: Hier spenden.

Hier kannst Du per PayPal spenden.

About the author

Igor Wallossek

Editor-in-chief and name-giver of igor'sLAB as the content successor of Tom's Hardware Germany, whose license was returned in June 2019 in order to better meet the qualitative demands of web content and challenges of new media such as YouTube with its own channel.

Computer nerd since 1983, audio freak since 1979 and pretty much open to anything with a plug or battery for over 50 years.

Follow Igor:
YouTube Facebook Instagram Twitter

Werbung

Werbung