Yeah, cost benefit is the bottom line. I only have one system, but have streamlined it, so for a single core machine, it is great. But, I'm hobbled running BOINC, which we all have to do now on WCG. It simply is not very efficient on a single core platform. Running on a different partition has had great results, though!

I'm just going to have buy myself an exceptional " core-duo" machine for Christmas, and make it aBoinc only machine. But withj xp, not vista! Maybe, linux!
And I'll keep my single core for gaming and give it to my son for his 11th B'Day!

I understand how to analyze lifecycle costs, but have to justify the cost of upgrading. If I can keep the same systems an additional year or 2 by switching them to Linux, and having several 'on the shelf' in case of hardware failure, how much more powerful machines can I get for the same price a year from now?
Just look how the price of that Q6600 dropped from $1100 initially to less than $300 in 6 months.
As far as electric cost, even if the new machines have a somewhat more energy efficient CPU, the periphiels will not be much different.
The consultant the boss has been talking to is trying to convince him to get Celeron Solo CPUs. While they may be cheaper, from a crunching standpoint,. that would suck! LOL
My best guess is, his company has a better markup in those systems (or they have a bunch they need to move) or he wouldn't be pushing them.
So far I have convinced the boss it's better to wait than to get machines like that.

Ah, then its one of those "best dont fix it if its not broke" situations.

Dont worry, I was just womdering if the concept project here could help you there. What im not trying to do is rock the boat.

Just curious

LOL, actually, the way we use computers; we could get away with thinclient machines. But those don't crunch very well...LOL
About all we use (officially) is the GM Website, email and Terminal Emulation software to connect remotely to an AS800 server.
It's nice to have real computers, but when the boss is tight, it's best to use what you have rather than 'upgrade' to something less than what you have...
I did convince them to install a Cisco VOIP phone system rather than spend the big $ for a digital phone system.
So far, I can demonstrate that it has saved $500,000 over what we would have had to do with the Toshiba system the consultant was pushing. Adding another node is as simple as adding another $1200 blade. With the other system, adding a node was over $50K and required days of programming and tweaking.
We have 6 locations around town and each would have required it's own node. By utilizing PTP T1s between locations, we can run all 6 locations off 1 PBX type switchboard....with 1 receptionist instead of 6.
And anyone can transfer a call to anyone without any complications.
With the VOIP system, if someone moves to a different desk (even in a different building), they take the phone with them. The phone number is assigned to the phone's IP, not the jack like a digital phone.
The phones use the same network as our computers and I use Power over Ethernet rather than having a power supply on each desk for the phone.