First I used the multimeter and checked each voltage. The +5v was 5.04v and the +12v was 12.049v. I say .049 because the display bounced from .04 to .05 and back the entire testing time. I left the multimeter connected for further testing.

Next I launched OCCT. OCCT is a benchmark that stresses the CPU and memory for a 30 minute run. The cool thing about OCCT is that it makes a chart of the voltages during the test, accessing the rig sensors through Speedfan. We can use this chart to determine if the voltage dropped during the test run.

While OCCT was benchmarking, I performed other various applications to further stress the power supply, including the acid test, Furbench. Furbench is an OpenGL graphics benchmark, which cranks up the GPU(s) to 100%. With today's trend towards going “green”, processors really don't use that much juice - the GPU is the real power user now.

The only issue with this testing is that for some reason, Speedfan doesn't measure the correct voltage, I guess it doesn't like Vista. But the chart is still useful, as it still measures the PSUs activity; only the baseline is off.



The Real Power Pro 1250 did not budge at all, even when (at 16 minutes into the test) I launched Furbench. This is the first time I have seen this performance on a multiple-railed PSU. All others have dropped voltage at least a little. Once again, my little HD 3850s only use a fraction of the power that an 8800 GTX or Ultra uses, but the test does show the awesome power of the 1250, undaunted by two maxed GPUs.