That test is nonsense as well as what the guy above me said. The 580 pulls 250W, and
it goes down lower after each card due to SLi Scaling (GPU usage). You don't just add up numbers on top off each. It goes like 250W-225W-200W at stock
We went over that at eVGA half a year ago, that setup doesn't pull anywhere
near 2kW. He only has a mild OC on Gulftown, all under air with an SSD. Either his load meter is horribly inaccurate or it's doctored (div. is calibrated) to show higher numbers (to market the HCP-1200). Look at here, Guru3D Nehalem 580 Tri-SLi setup with CCFL's, WC'ing and such peak
ed to 993W from
the wall; while running FurMark:
Do the AC-DC convertion, shave off 100W, an another 50W for FurMark, and one more 50W for their extra's. In reality (games - no halfway load) 580 Tri-SLi will pull about 750W. Also, see this post. 480's which pull 50W more. So you can easily expect a 1350W PSU to handle that setup.
If you're trying to say that 3 stock clocked GTX480s are pulling almost 1500w I can tell you that you're not even close to being correct. you're either measuring power usage at the socket and not accounting for efficiency in the AC to DC conversion our your setup its reading horribly inaccurately.
I used a Kilawatt to test my system and gaming it never goes over 900watts.
Windows 7 Ultimate 64bit,
M/B, Asus RE3, EK-FB,
CPU, i7 980x, WC Corsair H50,
RAM, Corsair CMG6GX3M3A2000C7,
GPU 1, Asus GTX 480, SLI WC EK-FC480,
GPU 2, Asus GTX 480, SLI WC EK-FC480,
GPU 3, EVGA GTX 480, SLI, WC EK-FC480
GPU 4, EVGA GTX 260 Physx,
PSU,Corsiar 1000HX x 2
HD, VelociRaptor 150gb Raid0, WD Caviar Blue 1tb"
Edited by Johnny Doe - 11/3/11 at 6:06am