BM Seer Unofficial thoughts from an anonymous Sun employee

Chopped configs and SPECpowerssj2008

Wednesday Jan 02, 2008

Timothy Prickett Morgan in the IT Jungle missed the point a bit when writing about SPECpower_ssj2008. It is not that you can't use power supply ratings to determine power usage. This difference really proves that these are chopped configs that are being used just for making numbers on SPECpower_ssj2008. Vendors who are trying to build power-efficient systems don't make mistakes like using small configs on power supplies that are too big. Power supplies are most efficient when the watts drawn are near the power supply ratings. If HP thought these small-memory chopped configs were typical they would have put smaller power supplies in them.

Have power supplies that are ONLY at 22% load (269/1200) is smoking gun aimed right at showing this is not typical config...

    "it is no surprise at all that the best metrics for this are coming from streamlined two-socket machines so far. An HP ProLiant DL160 G5 with two quad-core Xeon E5450 processors running at 3 GHz with 16 GB of main memory and a single 80 GB 7.2K SATA drive was able to max out at 308,022 ops/sec on the test; the average idle power draw on this machine was 160 watts and when running full speed (99.2 percent of CPU capacity), it drew 269 watts. This machine had a 1200 watt power supply, which tells you how useless these maximum draw ratings on power supplies can be in determining power usage."
http://www.itjungle.com/bns/bns121807-story01.html

Disclosure statement

SPECpower_ssj2008: HP Proliant DL160 G5 (2-chip QC Xeon E5450 3GHz), 698 overall ssj_ops/watt. SPECpower_ssj2008:HP Proliant DL580 G5 (4-chip QC Xeon L7345 1.86GHz), 546 overall ssj_ops/watt, 359523 ssj_ops and 387 watt at 100% target load, 255512 ssj_ops and 359 watt at 70% target load, and 71409 ssj_ops and 294 watt at 20% target load. Results from www.spec.org as of 12/11/07.

[7] Comments
Like this post? del.icio.us | furl | slashdot | technorati | digg
Posted at 02:00PM Jan 02, 2008 in Wattage-Power
Tags:
Comments:

Given that even Sun's own "SWaP" "Fair Use" rules call for actual metered power:

http://www.sun.com/servers/coolthreads/swap/fair_use.jsp

why all the concern about trying to defend what can best/most diplomatically be described as "guessing" systems' power consumption by taking a fraction of the rating of the power supply?

Also, there are a couple URL's out there which suggest that having a larger capacity power supply isn't necessarily such a bad idea:

http://www.pcpower.com/technology/myths/
http://www.wikihow.com/Buy-a-Power-Supply

This one wasn't quite as strong, but still intriguing:

http://www.anandtech.com/casecoolingpsus/showdoc.aspx?i=3186&p=2

I wouldn't know an "80plus" power supply if it reared-up and shocked me on my backside, nor which systems do or do not have such things, but seems that an "80plus" certified power supply is at least 80 percent efficient even at 20% of its rating:

http://www.80plus.org/80what.htm

Posted by rick jones on January 02, 2008 at 08:47 PM PST #

Sure always best to use measured data, but HP doesn't show measured
data on 32GB or 64Gb configs that they use in other SPEC benchmarks.
The truth is every vendor knows that memory draws a lot of watts.

HP only wants to show tiny 8GB and 16GB configs, smoking gun...

HP uses lots of 32GB and 64GB configs on benchmarks, clearly shown at: http://blogs.sun.com/bmseer/entry/big_memory_used_on_benchmarks

But never publishes measured wattage on these reasonable configs. Maybe I'll get a power meter and just measure it for you all...

Posted by BM Seer on January 03, 2008 at 09:39 AM PST #

Looking at SPEC results further...

Sun only wants to show 16 DIMM configurations on T5200, smoking gun...

There must be a design flaw in the memory subsystem such that adequate performance is achieved only when all 16 slots are populated.

Not true? Prove it with a SPECpower_ssj2008 result!

The other vendors are being energy efficient by using only the memory necessary for the workload. For example, HP uses 16GB for SPECjbb2005 and SPECpower_ssj2008, so they are simply optimizing intelligently. Sun, on the other hand, is trying to hide a design flaw by using 64GB for a benchmark that only needs 4GB when using a single jvm.

Not true? Prove it by publishing a SPECpower_ssj2008 result.

Posted by Carly on January 03, 2008 at 03:29 PM PST #

Gosh it is so much simplier than that...
Carly :) just publish measured wattage with every SPEC benchmark.

How come the only SPECpower_ssj2008 HP Proliant DL580 G5 is a 16GB 4-chip QC Xeon L7345 1.86GHz... and not the 32GB-64GB with 2.6GHz to 3GHz Xeons used on the other benchmarks?

Why can't we see the measured watts on these HP configurations used
in other benchmarks?
64GB: HP DL580 G5 (4 quad-core Xeon 2.933GHz)
64GB: HP DL585 G2 (4 dual-core Opteron 3GHz)
32GB: HP DL380 G5 (2 quad-core Xeon 2.66GHz)
48GB: HP rx6600 (4 dual-core Itanium2 1.6GHz)
32GB: HP DL380 (2 quad-core Xeon 3GHz)
32GB: HP BL685c (4 dual-core Opteron 3GHz)
32GB: HP BL460c (2 quad-core Xeon 3GHz)
32GB: HP rx6600 (4 dual-core Itanium2 1.6GHz)

Disclosure Statement (all data as of Sept 27, 2007)

SPEC, SPECjbb, SPECjAppServer, SPECweb, reg tm of Standard Performance Evaluation Corporation, Info on www.spec.org. NotesBench Domino[R] R6iNotes, more info www.notesbench.org. Two-tier SAP Standard Sales and Distribution (SD) standard SAP ERP 2004/2005 application benchmark: SAP, R/3, mySAP reg TM of SAP AG in Germany and other countries. More info www.sap.com/benchmark.

SPECpower_ssj2008: HP Proliant DL160 G5 (2-chip QC Xeon E5450 3GHz), 698 overall ssj_ops/watt. SPECpower_ssj2008:HP Proliant DL580 G5 (4-chip QC Xeon L7345 1.86GHz), 546 overall ssj_ops/watt, 359523 ssj_ops and 387 watt at 100% target load, 255512 ssj_ops and 359 watt at 70% target load, and 71409 ssj_ops and 294 watt at 20% target load. Results from www.spec.org as of 12/11/07.

Posted by BM Seer on January 03, 2008 at 04:59 PM PST #

Perhaps one day you will get your wish as power is added to other SPEC benchmarks. I don't agree that it is as simple as you seem to think, but it is doable. In the meantime, the lack of a SPECpower_ssj2008 result from Sun does seem rather, well, odd. Run one benchmark, or rerun several. I have a different idea than you as to which is simpler :)

Posted by rick jones on January 03, 2008 at 06:18 PM PST #

You forgot to quote the ending part of the same article: "And Sun, with all of its talk about how the Sparc T1 and T2 chips are the most efficient processors in their performance class should have been first out with tests"

Posted by T1T2 on January 03, 2008 at 08:21 PM PST #

Reminder I've already covered the substance of your attacks, see...
http://blogs.sun.com/bmseer/entry/specpower_ssj2008_power_benchmark_needs

Sun has been showing power for years on all of the US T1 & US T2 benchmarks and nothing from HP, they wait till a new benchmark and use chopped configs and tweaked BIOS..

Here is Sun's measured data on all benchmarks:

2007: http://www.sun.com/servers/coolthreads/t5220/benchmarks.jsp

2005: http://www.sun.com/servers/coolthreads/t1000/benchmarks.jsp

2005: http://blogs.sun.com/bmseer/tags/t2000

By the say the reason I keep focusing on memory is I know how much
power Xeon-based memory uses -- see all kinds of past postings in the wattage category.

Posted by BM Seer on January 08, 2008 at 05:53 PM PST #

Post a Comment:
Comments are closed for this entry.