Nehalem: Xeon Gets Core i7 Upgrade
Posted on: 03/29/2009 05:00 AM

Test Systems

Most of the reviews you will see hit the 'net today will be based on a swanky 1U server sporting a pair of Xeon W5570 processors. The X5570 is the top speed, 80W TDP Nehalem that clocks-in at 3.0ghz. You'll be seeing something a bit different here today, though, as my test system is a Supermicro workstation containing a pair of Xeon W5580 processors. The W5580 is the absolute top speed Nehalem, and it weighs-in at 3.2ghz with a TDP of 130W (there is a typo in the table below, I'll get it fixed ASAP).

Since the is the top-of-the-line, power-be-damned Nehalem, it seemed only fitting that I compare it to the current performance king (with a 150W TDP), the 3.4ghz Harpertown-based Xeon X5492.

Both systems are as identical as possible save a few components necessitated by the fundemental differences of the platforms.

The one glaring difference between the two platforms is the memory. While FB-DIMMs offered-up some benefits from a stability and flexibility standpoint, they got a pretty bad rap as being power hungry, hot and having high latencies. It will be interesting to see how the switch to Registered DDR3 will effect performance, especially coupled with QPI.

Notably absent from this comparison is a competitive product from AMD. I tried hard to procure a similar test system with a pair of "Shanghai" Opterons, but my calls went unanswered.


I used the following versions of our test applications: <ul><li>SiSoftware SANDRA 2009 SP2
<li>Black & Scholes Kernel (Optimized x64, custom VS2005 compile)
<li>Autodesk's 3ds Max 2009 64-bit[/url]
<li>SunGard Adaptiv Credit Risk Analysis 4.0
<li>CINEBENCH R10 x64
<li>POV-Ray 3.70 beta 30 x64
<li>SPECjbb2005 v1.07
<li>Euler3d CFD Benchmark v2.2
<li>Matrix Multiply v3.5
<li>FlamMap (FSPRO)
<li>x264 Benchmark HD v2.0</ul>All benchmark numbers presented were an average of three runs. All test systems' desktops were set at 1280x1024, 32bit color with a 60Hz refresh rate. "Optimized Defaults" were used in the BIOS on all machines.

Hardware Prefetching and advanced power management settings were enabled on both systems. The Nehalem machine was tested with SMT and Turbo set to ON for all tests.

All power measurments throughout this review were taken with an Extech 380803 power meter. Readings were for the computer itself, and all other devices (monitor, speakers, etc) were plugged-in seperately and not included in the measurments.

Printed from (,2.html)