Skip to main content

AMD Radeon R9 380X Nitro Launch Review

Power Usage Results

Measuring power consumption at idle can be difficult due to the system’s load spontaneously changing. For this reason, we use a longer observation and pick the most representative two-minute time interval for our reading. We then average the results from this interval.

Please note that the minimum and maximum results for the different rails in the tables below don’t necessarily add up to the overall total for that category. The reason for this is that the extremes on the individual rails don’t always occur all at the same time.

Idle Power Consumption

At idle, AMD's Radeon R9 380X consumes approximately 13W, and this number stays essentially the same when two monitors are connected.

MinimumMaximumAverage
PCIe0W21W8W
Motherboard 3.3V0W1W1W
Motherboard 12V0W15W5W
Graphics Card Total1W29W13W
Image 1 of 7

Image 2 of 7

Image 3 of 7

Image 4 of 7

Image 5 of 7

Image 6 of 7

Image 7 of 7

Gaming Power Consumption (Metro: Last Light Loop)

The gaming loop runs at Ultra HD, since that resolution imposes the highest power consumption. Sapphire's R9 380X comes in only six watts above MSI’s R9 380, which is a smaller delta than the performance increase of approximately nine percent might have suggested. Tonga XT turns out to be a bit more efficient across several different benchmark runs.

The table paints a fairly decent picture:

MinimumMaximumAverage
PCIe73W212W141W
Motherboard 3.3V1W3W2W
Motherboard 12V24W78W48W
Graphics Card Total102W287W191W

The motherboard slot never hits its output ceiling. Even the spikes almost never exceed the slot’s maximum. This is perfectly done.

Pulling out a snapshot of just one moment in a game shows how power is distributed across the rails:

The overall distribution can be seen here in its usual gallery format:

Image 1 of 7

Image 2 of 7

Image 3 of 7

Image 4 of 7

Image 5 of 7

Image 6 of 7

Image 7 of 7

Full Load Power Consumption

Even when AMD's Radeon R9 380X is pushed as hard as possible via FurMark, its power supply stays stable. The new card does exhibit the same behavior we observed from the Nano, though: once it really gets going, it doesn’t stop for anything. The 252W it consumed during the torture test almost matched the 390(X).

MinimumMaximumAverage
PCIe244W186W
Motherboard 3.3V1W2W2W
Motherboard 12V23W94W64W
Graphics Card Total94W340W252W 

We put together all of the individual full-load diagrams as well:

Image 1 of 7

Image 2 of 7

Image 3 of 7

Image 4 of 7

Image 5 of 7

Image 6 of 7

Image 7 of 7