Are two processors better than one? Hardware system #4

As a consultant for many years of business systems, it always seemed better to recommend multiple processors when specifying server or high end workstation equipment.  Long before the Dual Core CPU came about were the multi-cpu boards and systems.  One of my favorite boards that hung on the wall for a long time was a dual 386 processor board.  It had two Intel 386, 33Mhz processors built onto it.  The board came out of an early server used in bank and financial transactions.
Intel and AMD both have a huge business market for multiple processor systems.  Recently it was shown that personal computer shipments were down last year, but Intel made a huge profit primarily due to the server market.
Even in todays multi-core CPU world, multiprocessor systems play a huge role, especially in virtualization.  I recently got to drool all over some servers we ordered.  The servers were only 1U rack space, 1.75" tall.  The amazing part were these servers housed two processors each.  Inside each of those processors were six cores, and they were hyperthreaded.  It is amazing to pull up a Windows 2008 Enterprise server and see 24 CPU threads in task manager.
So unfortunately many times did I have to explain to the companies ordering these multiprocessor servers that because it was two processors, it did not mean exactly double or more performance.
The other misconception was that if the server had two 2Ghz processors, that the server was running at or equivalent to a single 4Ghz processor system.

Into the mix of the benchmarking I decided to throw a nice video and photo editing workstation.  Now granted this system dates back before the dual cores and includes the hyperthreading technology of the Pentium 4 HT class chips.  I also tried to make this system as close to the benchmark systems being used here in the other posts.  So I outed the usual RAID 5 of 10,000 RPM SCSI drives and thankfully the board had an onboard SATA controller.  Also removed was the 8GB of memory typically used and the 1GB video card.
Following are the specifications I ended up with.
System Components:
Hardware system #4
CPU Dual (2) Intel Xeon 3.4Ghz HT, 1 Core / 2 Threads (Hyperthreading) each CPU, 1MB Cache /CPU, 800Mhz Rated FSB
Dual Socket 604, Nocona 90nm, Stepping 1, Revision E0, Multiplier 17x, FSB-200Mhz
MMX, SSE 1,2,3, EM64T

Board OEM Intel Dell 0u7565 (Precision Workstation 600 series, PowerEdge 1800/2800 Server Series)
Intel E7525 / 82801EB (ICH5), Dual Socket 604
2 * SATA, 2 * IDE, Ultra320 SCSI RAID, 10 * USB, 2 * IEEE 1394, 1 * PCI-e x16, 2 * PCI-x, 6 * PCI

Video -- Trust me here if the workstation had some kind of onboard VGA I would use it.  Many server and workstation boards usually come with a 16MB - 64MB PCI, this board did not.
So I picked the lowest powered PCI Express x16 card I could find as to not skew the benchmarks too bad towards the video
Add-In Board Nvidia Quadro FX540, NV43
Revision A4, 110nm, PCI-E x16, DDR 128-Bit, DirectX 9c,
128MB Dedicated Memory, 300Mhz GPU

RAM - 2GB (2 * 1GB) Infineon DDR2 400r (PC2-3200R) Registered ECC
Running at 200Mhz Dual Channel Mode DDR2-400, CAS 3-3-3-11, 1.8V
2046MB Total, 1722 Free after Clean install

As with servers and workstations, most of them will not use standard desktop memory.

I again picked modules that were close to the others benchmarked but compatible with the ECC needed.

Not in use for testing:
Sound ADI 198x SoundMax AC97 audio
LAN Intel Pro1000 MTW Gigabit Onboard

Common Components to be used in all tests unless noted otherwise:

Hard Drive Western Digital Caviar Blue 250GB, 7200RPM, 8MB Cache, SATA II 3Gbps

Optical AOpen DUW1608 Dual Layer DVD Writer ATAPI/EIDE

KVM USB KVM with Logitech Optical Mouse, 104 Keyboard, 17" LCD DVI/VGA

Power Supply -- Dell 650WT Workstation/Server unit

At least I was able to use the same hard drive to try and keep this test close to the normal deviation of the other hardware configurations.