Kill-A-Watt: AMPs draw on PE2950r3

[Brian A. Seklecki (Mobile)]

On Thu, 2008-02-21 at 13:42 -0600, Jeff Macfarland wrote:
> These are *not* scientific. For loading, I started bonnie and/or prime95
> torture test, waited a minute and looked at the meter.
> 
> prime95: ./mprime -t -b8
> 
> bonnie++: ./bonnie++ -d `pwd` -u nobody -r 4096 -s 8192
> Two instances of bonnie++ were running in parallel as there are two VDs.
> 
> Setup:
> 
> 2950
> 2x Intel(R) Xeon(R) CPU E5430  @ 2.66GHz
> Dual PS
> 16G (8x2GB) ECC @ 667Mhz
> Drac5
> Perc 6/i
> 6x ST3300655SS (300GB 15k 3.5" SAS)

Wow... you near 2.4 amps, which is only .3 to .4 amps more than we were
seeing.

Not quite the stratification we were expecting with additional RAM, CPU,
Disk.

This is an Energy-Saver or non-ES model?

~BAS


> I used a dual ported power cable, so all readings are accounting for
> both PS units.
> 
> 
> 
> No load
> 
> Off:     0.25 amp, 0.35 PF
> Idle: 172.8 Volts, 1.70 amps, 0.56 PF
> 
> Full load
> 
> bonnie++: 172.8 Volts, 1.88 amps, 0.57 PF
> prime95:  172.8 Volts, 2.37 amps, 0.58 PF
> bonnie + prime95: 172.8 Volts, 2.38 amps, 0.58 PF
> 
> Brian A. Seklecki wrote:
> > Slightly OT, but there has been recent discussion about the 'adaptive
> > energy smart power' and AMD64 "green" models.
> > 
> > We recently used a kill-a-watt(1) to measure AMP draw on a 'entry-level'
> > PE2950 (9th gen, 3rd revision).
> > 
> > This unit had 4 SAS drives (instead of the usual 6), PERC5 w/256, 4G of
> > ECC RAM, 1x physical CPU w/ 4 cores (Xeon E5310 @ 1.60GHz), and dual
> > power supplies, DRAC5, no add-on HBAs, 4x Addon Intel Gig 82571EB.
> > 
> > Here are the result (courtesy of Bill Moran):
> > 
> > Total average power draw at idle: 1.8 amps.
> >       * Total average power draw at full load: 2.1 amps.
> >       * When both power supplies are in use, the power usage is
> >         distributed between them nearly evenly.
> >       * The system uses approximately .1 amp less power when running off
> >         a single PSU than when running on both, during both idle and
> >         loaded tests.
> > 
> > So 1.8->2.1 amps for an entry-level PE2950.  With I'd be curious to see
> > what a loaded 2850 or 2950 looks like (2x quad-core 3.0ghz 4560, 32TB
> > ECC, 1TB RAID, etc.).  Then also compare similar Disk/RAM spec'd to
> > AMD64 equivalents.  
> > 
> > ~BAS
> > 
> > 
> > (1).  The kill-a-watt isn't going to be used in UL-certification tests
> > any time soon, but its relatively accurate for measuring Amps, Watts,
> > KWH usage, as well as finding Volt and HZ spikes.
> > 
> > http://digitalfreaks.org/~lavalamp/cp/displayimage.php?album=2&pos=315
> > http://digitalfreaks.org/~lavalamp/cp/displayimage.php?album=2&pos=314
> > 
> > 
> > 
> 




IMPORTANT: This message contains confidential information and is intended only for the individual named. If the reader of this message is not an intended recipient (or the individual responsible for the delivery of this message to an intended recipient), please be advised that any re-use, dissemination, distribution or copying of this message is prohibited. Please notify the sender immediately by e-mail if you have received this e-mail by mistake and delete this e-mail from your system.