p2501 wrote:... client generates new CPUID. Problem solved.
Judging by Beta v7 Client User Guide, solution seems to be well underway?
machine-id
If you run multiple clients on the same machine you can set this to a different value so that the work servers will give you different work units for each client. However, this option is mostly obsolete since the new client can run more than one core at a time and avoids the problems that required using machine-id in the past.
Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
Since I started folding 7+ years ago on v3.24, each new major client version has been an incrimental upgrade.
When a project like FAH gets this big, it's really hard to make big changes without causing a serious interruption. A cruise liner doesn't turn on a dime.
However, V7 is the first completely new code base for the client since, well, since fah started. So v6 is based on 10 year old programming methods, like running in a DOS-like command line window, and separate windows for each client, etc. Try out the V7 beta. It is a major step forward.
A Windows client can be copied from one machine to another without difficulty; the registry cannot. FAH only has a problem if you CLONE your copy of Windows, but that also violates Microsoft's licensing agreement: You should not run the same copy simultaneously on more than one piece of hardware. Stanford has a lot of problems that should be solved in their software but I don't expect that this one will have a high enough priority for them to do anything in the near future. After all, it's only a problem for those people who intentionally violate a licensing agreement (even if it's not their own EULA).
Bruce, I'm not sure if your post was directed at me, but to be clear I am running Linux on my cluster. I do have a windows machine or two that occasionally fold, but the bulk of my work is done on Linux nodes.
I was dismayed to find this thread months after the fact and see that, perhaps, I had wasted a lot of cycles performing duplicate work, and to add salt to the wound, some thought I was up to something fishy.
I didn't really direct the comment at anybody in particular. It's factual information related to the discussion.
You can copy a Linux directory to another machine as long as you do not copy the machinedependent.dat. Frankly, since you would then have two copies of the same WU, it doesn't really make sense to copy anything because you'd need to delete those files associated with one copy of the current WU. It's just easier to download a new copy. If the hardware is similar enough that you'd like to avoid configuring the client again, copy just client.cfg (V6) or config.xml (V7).