To attempt to stress the PSU, the following hardware was used:

• Opteron 165 Dual Core @ 2.7GHz w/ 1.45v
• Sapphire X1900XT
• DFI LanParty UT nF4 Ultra-D
• 2GB G.Skill HZ PC4000
• X-Fi Xtreme Music
• 250GB Samsung 7200RPM PATA SP2514N
• NEC DVDRW 3520A
• 2x 80CFM 120's, 2x30CFM 80's

Readings were taken with a digital multimetre. Full load consisted of two instances of Prime95 (one for each core) and looping 3dmark05.

Here are the results at full load:

• 12v: 11.92
• 5v: 5.13
• 3.3v: 3.36

ATX standards state you are allowed a 5% tolerance either way, so for example the 12v rail can range anywhere from 11.4 to 12.6 and still be in spec.

This PSU had no trouble at all powering the system. In comparison to the X2-Connect, I felt a lot more confident using this particular PSU. It didn't seem to struggle and I would be quite prepared to add, for example, a couple more hard drives and still be confident the PSU would manage. The X2-Connect was indeed tested with an X1800XT rather than an X1900XT as was the case here, but the difference in power consumption between the two is minimal, and given the fact that the X1800 was overclocked and the X1900 is not, this really shouldn't be a factor. No, this PSU just seems to be of better quality.