Xeon PC's Performace has Slowed Over Time?

[Chasteen]

I have 2 identical dual Xeon 5160 workstations I built a few months ago. Recently I noticed in bb that one was rendering slower. I re-tested that ws by rendering a small MR Max scene called MR Glacier and it performed 26% slower than when I first built it. I decided to re-install a clean version of XP64, but afterwards it performed identical. Anyone have any idea what direction I should take to identify the problem?

 

Parts list:

Supermicro X7DAE mobo

(2) Xeon 5160

(2) 1GB Kingston RAM @ 667MHz

XP64 SP1

Max 8 SP3

[AJLynn]

Check the speed settings in the BIOS, the CPU and RAM utilization while rendering in Task Manager, and try some other speed tests and benchmarks.

[Chasteen]

Thanks for the reply.

 

The CPU's show as 3GHz, the RAM at 667MHz and the FSB as 1333MHz in the BIOS. During rendering the tmgr shows all 4 CPU's at 100% and RAM at only a little over 500MB. I suspect the mobo may be responsible for the lower performance, but I don't know how to check that.

 

I don't need any additional proof of the slower performance. I have a few weeks of bb projects of various kinds documenting the ws's consistently slower performance than the it's twin. Are there any speed tests that might isolate the performance of the CPU, RAM & FSB?

[AJLynn]

http://en.wikipedia.org/wiki/Benchmark_(computing)

[Chasteen]

The problem is solved and I wouldn't believe it if I didn't experience it first hand. Heat was the cause. The only difference between the 2 ws's were the heatsinks. One set was the passive(no built-in fan) and the other was active(built-in fan). The passive set was running a little (not a lot) hotter so I swapped them out for an active set of heatsinks. Afterwards the ws consistently posted render times equal to its twin.

 

I have experience OC'ing and know the importance of cooling to create a stable PC under load, but I've never heard of improved performance from cooling a non-OC'ed pc.