power consumption at ISC

(watt figures are "average watts")

baytech-a2	1410W	85.1F	20080705 11pm	off: java-stable, hammer4
baytech-a2	1458W	85.1F	20080705 11pm	off: java-stable, hammer4
baytech-a2	1859W	87.8F	20080705 11pm	off: java-stable
						(hammer4 just on)
baytech-a2	1808W	90.5F	20080705 11pm	off: java-stable
baytech-a2	1557W	90.5F	20080705 11pm	off: hammer4 (java-stable
						just on)
baytech-a2	1517W	90.5F	20080705 11pm	off: hammer4
baytech-a2	1480W	90.5F	20080705 11pm	off: hammer4

baytech-a2	1603W	85.1F	20090718 9pm	off: hammer4

(figure 3 Celesticas * 350W, 2 PowerEdge 2850 * 150W, 200W for cs1 and netgear?)

(max with hammer4: 1900W)

(20100111 measured with hammer1-3 gone: 500W)

------------

baytech-a3	 653W	89.6F	20080705 11pm	off: netra2, deo-ultra5,
						panther*, hammer5, hammer7,
						ref4, netra6-9
baytech-a3	 730W	93.2F	20080705 11pm	flip netra2
baytech-a3	 740W	93.2F	20080705 11pm	flip netra9

(figure 4 netras * 75W, 350W for danner, hammer6, rockport)

baytech-a3	1074W	92.3F	20090718 9pm	off: netra2458, deo-ultra5,
						panther*, ref4; on: fif0, fif1,
						hammer5, hammer7

(figure 2 hammers @ 120W: gives 2 soekris * 40W)

(max with all netras: 1370W)

baytech-a3	1090W	96.8F	20091014 3pm	off: netra3, deo*, panther*,
						hammer5; ref4 disconnected

(20100111 measured switch, danner, all netras, hammer5-7, rockport, fif0/1: 1390W)

------------

baytech-b2	 553W	88.7F	20080705 11pm	off: hammer9-11
baytech-b2	 514W	88.7F	20080705 11pm	off: hammer9-11, thunder1
baytech-b2	 499W	88.7F	20080705 11pm	off: hammer9-11, thunder1
baytech-b2	 486W	88.7F	20080705 11pm	off: hammer9-11, thunder1

(figure 5 thunders * 50W = 250W, hammer8 = 300W)

baytech-b2	 798W	95.0F	20090717 9pm	on: thunder1235, hobson123, dgs; off: hammer8-11, thunder4
baytech-b2	 686W	95.0F	20090717 9pm	on: thunder1235, hobson12, dgs; off: hammer8-11, thunder4
baytech-b2	 558W	95.0F	20090717 9pm	on: thunder1235, hobson1, dgs; off: hammer8-11, thunder4
baytech-b2	 481W	95.9F	20090717 9pm	on: thunder1235, hobson1, dgs; off: hammer8-11, thunder34

(figure 3 dell 1750s * 130W = 390W)
(XXX this indicates thunders are 75W?)

(calculated max with 3 1750s, 5 thunders, hammer8 fully going: 1000W)

baytech-b2	1330W	99.5F	20090810 4pm	on: thunder12345, hobson12369, dgs, enet2; off: hammer*, hobson34578
baytech-b2	1540W	99.5F	20090810 4pm	on: thunder12345, hobson123679, dgs, enet2; off: hammer*, hobson3458

Xi systems: 200-220W

(calculated max with 3 1750s, 4 Xi, 5 thunders, hammer* fully going: 2080W)
(calculated max with 4 1750s, 5 Xi, 5 thunders, hammer* gone: 1620W)

(20100111 measured thunder3-4, hobson1356789, 2 switches, all under
  load: 1650W-1857W; 1785W nominal)
(that's 2 * 75W, 3 * 130W, 4 * 220W = 1420W, leaves 230W for switches?
 can't be right.)

(20100111 toggle on hobson2: 1874W-2013W; 1960W nominal)

1960-1785 = 175

now try 2 * 75W, 4 * 175W, 4 * 220W = 1730, leaves 235W for switches,
now less believable than before???

now try 2 * 75W, 4 * 175W, 4 * 250W = 1850, leaves 110W for switches,
much more believable.

I'm not willing to cycle hobsons unless someone is on-site to press
the button -- just in case.  (perhaps this is only a problem with
the 1750s?)  So, measure power later.

calculated max with 5 thunders, 4 1750s, 5 Xi, switches: 2435

so, partition.

thunders, switches, hobson987: 1160
hobson65321: 1025

------------

Celestica		300-350W
hammer rackmount	120W
PowerEdge 1750		130W XXX now believed to be 175W under load
PowerEdge 2850		150W
netra			75W
thunder			50-75W
Xi			220W

------------

minimum: 680W
usual total: 4217W
max total: 5350W

XXX MCL 20100731

a website http://www.youtube.com/watch?v=jErq4gWOOSM measures 2950s
at 240W idle, 315W fully loaded.

assuming we scrap thunderN and hammerN, and give 4x Xi systems to secteam,
we have:

(2*pe2950)-(5*thunder)-(3*hammer)
-105

net savings of 105W

takes us down from current theoretical max of 4235, to 4130