BM Seer Unofficial thoughts from an anonymous Sun employee

Dell games with little info, tiny mem, and low GHz

Tuesday Jan 22, 2008

Dell's comparison of power-performance done by Principled Technologies simply isn't very good. They leave out the critical configuration details to make the results look good.

Games Dell played:

  • They used tiny 4GB of memory configurations, Read here to see huge power implications of memory size games.
  • only filling half DIMM slots, and used small 1GB DIMMS
  • They used non-standard BIOS hacks to improve performance (ex: Disabled hardware Prefetcher and Adjacent Cache Line Prefetch!) This helps some Java benchmarks but really HURTS real performance on workloads. We talked about this trick before: Hacking non-standard BIOS
  • They only used 2.3GHz processors, hmmmm... Dell always talks about high-GHz on every other benchmark?
  • Dell doesn't talk about ways to drive up the system utilisation, which saves more power than any games they played on the benchmark. They talk about small percentage differences in watts/performance - utilisations save factors in terms of watts/performance.
  • older 10K RPM drives.
  • No 10G Gigabit Ethernet - only promised for the future!. In the power tests they only use one 1GbE.
  • showing perf/watt with 5 significant digits, to make this only appear accurate. I'm 99.97462452342534 sure about this.
  • Dell System is limited to 350W per blade, given their current supplies, fans, etc - that will continue to mean older CPUs and small memory configurations.
(P.S. IBM bloggers couldn't even find the memory configuration details of the test, all they had to do was look beyond the press release It was in the system's packaging report: http://www.principledtechnologies.com/Clients/Reports/Dell/DellHPIBMbladeserverOOB1207.pdf). I agree with those bloggers, GHz & memory details should be in the press release.

Blog info about other power-performance benchmarks with same of the same issues.

Register weighs in on this info-free announcement: http://www.theregister.co.uk/2008/01/21/new_blades_dell_m1000e/

[2] Comments
Like this post? del.icio.us | furl | slashdot | technorati | digg
Posted at 11:39AM Jan 22, 2008 in Competitor Follies
Tags:
Comments:

Your office must be really crowded with all the dead horses you are beating.

Your points 1, 2, 3, 4 and 6 are irrelevant to a comparison using the same processors, chipset and memory. Your extrapolation of the limits of 350W per blade is completely wrong - 350W is plenty to include the top processors that chipset supports, and a full load of 4GB DIMMs.

And, to jump on the most popular bandwagon in the comments area here, where are Sun's SPECpower_ssj2008 results? We all know the answer - the Sun "nominal" 90+ watts for the T2 is probably only a fraction of its actual power under load - hence the enormous 460+ watts shown for the T2-based systems. Subtract out the watts for the DIMMs, hard drives, etc and there are still over 300 watts left for the CPU. Hot as the Sun!!!!

Posted by Silver on January 22, 2008 at 06:22 PM PST #

It is clear that power benchmarking can not be decoupled from performance
benchmarks. On power-benchmarks like this vendors use Low-GHZ CPUs with tiny memory, hack the bios and make benchmark specials. Then they use
High GHz realistic configs on other benchmarks.

Also it is about the system not the low GHz chip with the same name to confuse everyone.

If you want to see SYSTEM comparisons with full memory I'd suggest that other vendors publish watts on all benchmarks, like Sun does:
http://www.sun.com/servers/coolthreads/t5220/benchmarks.jsp & http://www.sun.com/servers/coolthreads/t5120/benchmarks.jsp

Evidently vendors know they would lose so they avoid simply measuring on all benchmarks are resort to games like above.

Posted by BM Seer on January 23, 2008 at 10:55 AM PST #

Post a Comment:
Comments are closed for this entry.