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

  • Furen - Sunday, December 18, 2005 - link

    I think the comparison is fair enough considering that Benseley should be coming out within the next 3-5 months. The most I'd expect from AMD by then is maybe a bump in clock speed. Socket F is scheduled to come out in Q4 2006, if I remember correctly, and mentioning it would serve no purpose since we know absolutely nothing about it. Reply
  • Heinz - Monday, December 19, 2005 - link

    Ok, if your information would be correct I would agree, however Socket F is due to H1/2006, i.e. the same 3-5 month timeframe we have to wait for Bensley. Thus Bentley is not competing with the tested S940 Sytem but with a Socket F System.

    Look here:

    Apart from that, I do not see your point about not being interested in Socket F at all. Of course, there is little information about it, but what it is sure is that it will be presented in 2006. Because new product generations are normally faster/better than the old ones (old tradition in the computer, if not any market ;-). I would be interested to know that.
    It is like saying the new Volkswagen is next year best car on the market, because it is better than the current competition, without mentioning that there will also be a new Toyota model.

    If the information from pcstats is incorrect and anandtech got better information about a delayed/later launch of the Socket F platform, I apologize. But then again ... it should have been mentioned in the article :)


  • Furen - Thursday, December 22, 2005 - link

    Interesting, that's the first time I've seen that road map. I was kind of hoping that Socket F would come out at the end of the year with FB support, since I think tri/quad-core CPUs may be bandwidth limited with two DDR2 channels, then again the fact that they'll be DDR667 dimms may help enough. Reply
  • IntelUser2000 - Friday, December 16, 2005 - link

    The first Intel dual core for dual core servers is out and the CPU is called Paxville. It uses the aging Lindenhurst chipset with single 800MHz FSB while Bensley will use dual 1066FSB. The highest clock for Paxville is 3.0GHz.


    It says Bensley with Dempsey will be out Q1 of 2006.


    3.2 and 3.46GHz will be 1066FSB.
    2.5, 2.83 and 3GHz will be 667FSB.

    MV versions will be 3.2GHz with 1066FSB.
    LV versions will be 2GHz with 667FSB(ouch).
  • coldpower27 - Friday, December 16, 2005 - link

    Interesting, though with that projected clock frequency on the LV Dempsey, they might as well use the Sossaman processor. As the Roadmap in this article no longer points out LV Dempsey parts.
  • Griswold - Friday, December 16, 2005 - link

    Mostly information I didnt ask for / can be found in the AT article. Thank you for that. Reply
  • Frallan - Friday, December 16, 2005 - link

    I dont mind that AMD has held the lead for quite some time bc they needed it badly. But it wasn't good that Intel had nothing that could compare. Now at least Intel is on the map again as AMD which might acctually force AMD to go a bit faster again.

    Then on the topic of power, if it is 8k, 20k or 50k you save per year by buying product A instead of B isn't important if everything else is equal. 8k is more then enough to rule in favour of AMD (remember everything else equal).

    The future looks interestin and Im gonne pick a Denmark up soon :0)
  • menting - Friday, December 16, 2005 - link

    problem is that even if performance is equal..price parity on the cost of servers aren't... :)
    you get a huge discount if you go mostly intel.
  • phaxmohdem - Friday, December 16, 2005 - link

    Wow! A Dual core server/workstation chip. Way to go Intel! AMD better get its $hit together soon.... Oh wait...

    Do they make a special Silver stake or bullet that will ever kill fvcking Netburst?
  • coldpower27 - Friday, December 16, 2005 - link

    These are probably the last NetBurst based server processor you will see for this segement, the next slated update is Woodcrest core over the Dempsey core used now, which is part of Intel's NGMA. Reply

Log in

Don't have an account? Sign up now