Test Configuration

Software Configuration

Windows 2003 was configured with /3GB and /PAE switches in the boot.ini to support the 8GB of memory used for our tests. SQL Server Enterprise was set to use AWE extensions, and a maximum memory limit was set at 6144MB.

Intel Bensley 3.46 Pre-Production System
Dual 3.46GHz Dual-Core Dempsey Processors
Pre-production Blackford based Intel Motherboard
Windows 2003 Enterprise Server (32 Bit) SP1
8 x 36GB 15,000RPM Ultra320 SCSI drives in RAID-0
LSI Logic 320-2 SCSI Raid Controller

Opteron 280 System
Tyan S2882 K8S Motherboard
Dual Opteron 280 (Dual-Core) Processors
8GB Corsair PC3200 DDR
Windows 2003 Enterprise Server (32 Bit) SP1
8 x 36GB 15,000RPM Ultra320 SCSI drives in RAID-0
LSI Logic 320-2 SCSI Raid Controller

Measuring Power

To measure power consumption of each system, we used an EXTECH Instruments Power Analyzer Model 380803. This power analyzer allows us to view current power consumption, and log the consumption at various intervals during a test to a text file. For this test, we used the same Power Supply for both systems, although we recorded the difference between a 750W power supply and a 550W power supply, and it was less than 3 Watts. We should note that the Raid Array was powered by a separate power supply that was not plugged into our analyzer, so we were measuring strictly bare system power consumption. If you’re curious, the Raid Array used about 98 Watts spun up, and averaged 110 Watts during the database tests.

Idle – To measure a system at idle, we booted each system into Windows and let it stabilize by watching the task manager in Windows and the Wattage readings. Once we were at a stable reading, we began recording for 100 iterations of our data logger (which logs every 2 seconds). We then took those numbers and averaged them to get the idle power reading.

50% – We used our database benchmark to measure a loaded system, by adjusting the thread count for the test to a level that produced a half loaded system. Then, we would run our database test for its duration while recording to the data logger. Finally, we averaged those results.

100% – To produce a fully loaded system, we used the same technique as above, except increasing the number of threads until we achieved a fully loaded system.

The future is performance per Watt. Database Benchmark Results


View All Comments

  • OrphanBoy - Friday, December 16, 2005 - link

    Ok, I haven't got one of these new chips, but my work machine that I'm sitting at right now has 2x3.6GHz Noconas with a 7800 GTX and I often run it at full whack - the amount of power I must be drawing has to be huge!

    Nearly half a kilowatt per machine is a scary thought!
  • Cygni - Friday, December 16, 2005 - link

    Imagine a Pentium D EE on an Nvidia Intel SLI board with those quad-7800GT's from over at Toms. Maybe throw a few a nice RAID 5 array in too. Thats gotta pull down some SERIOUS wattage! Cant wait until the day that turning my gaming rig on for the first time trips my breaker, haha. Reply
  • Cygni - Friday, December 16, 2005 - link

    Of note, these arent 64bit results. I get the feeling that 64bit linux results would favor the Opty even more.

    The FB-DIMM controller and use of multiple FSB's (FINALLY!!!) really boosts the performance here to serious competition levels. This box would make a serious workstation powerhouse.

    But, as noted, Opty will likely have moved on quite a bit by the time this package is released. And it has better 64bit. And its platform is already available. But it is certainly interesting to see Intel really dominate a performance benchmark for the first time in a long time.
  • Peter - Friday, December 16, 2005 - link

    >use of multiple FSB's (FINALLY!!!)

    ... and only two years after AMD abandoned the dual-FSB approach and went HyperTransport.

    For those who forget quickly :) the Athlon-MP chipset (AMD762 north) was using dual independent FSB.
  • Viditor - Friday, December 16, 2005 - link


    Of note, these arent 64bit results

    A very good point...it doesn't make sense with the memory used to not use 64bit as well.

    Jason, could you let us know why you used 32 bit instead of 64 bit?
  • Jason Clark - Friday, December 16, 2005 - link

    We're going to look at 64bit in the new year. You have to realize most all of the general public running a windows 2003 server are running 32bit still. 64bit is not quite as adopted as you may think. That being said, we are going to look at 64 bit in the new year with sql 2005.

  • Viditor - Friday, December 16, 2005 - link


    we are going to look at 64 bit in the new year with sql 2005

    Fair enough, and thanks for the first peek!
    Enjoy your Holidays, then get out there and find us some MORE cool stuff to learn! ;)

  • Cygni - Friday, December 16, 2005 - link

    Its likely what Intel let them run at the time, as i doubt they shipped Anandtech a working system for their own use. :D

    Also, driver support is probably in its infancy. May not even have linux or x64 support today. Probably only Intel knows. But i think we can assume that 64bit will be similar to the current Xeons.
  • Griswold - Friday, December 16, 2005 - link

    Opteron 280 = available now
    "Bensley/Dempsey" = ???

    Btw, What about the Opteron 285SE at 2.6GHz that is exclusively used in the SUN Fire X4200? Should rectify the performance chart as well.
  • Heinz - Saturday, December 17, 2005 - link

    Yes that issue also catched my eye, actually it is a little bit a apples to oranges comparision. There are systems available now set up against systems available in Q2 or even H2 next year. Quote:

    Demspey is going to take us well into Q2 of next year, and Woodcrest will appear sometime in the second half of next year. Woodcrest will be a lower wattage part that is focused on performance per Watt.

    Well .. one should look at the AMD roadmap, what's in AMDs portfolio by then. Then you can declare a "winner" for the 2006 server market. Then the choice is really possible. Up to now there is not *any* choice. Everybody has to buy the AMD system, as the Intel is simply not available.

    I know that you cannot test Socket F now, thus you end up with the next best solution, which is an Opteron 280 system, but my point is, that you should have at least mentioned Socket F. These can/may (still speculation) deliver performance increases for AMD in 2006, too.

    Without that, the pure bottomline (performance) results of your article is, that a 2006 system is better than a 2005 one. Not really great news... even if it is about an Intel system being faster than AMD :)

    So for more objective articles, please try to cover all point of views of the industry. Without that, bad boys might question anandtechs independency, exspecially after you were invited to a nice(?) trip to the Intel headquarter ... no offense here, just trying to make a fair comment. After all it was a nice overview over the next-gen Intel platform with a lot of information.



Log in

Don't have an account? Sign up now