One unit is a command that tells a client to just do one unit and send it off. The it will wait to get a new work unit until you restart/tell it to.
I've got two of the standard large boxes going (both 4xCPUs, one 8GB one 16GB), but no idea at all how many PPD I'm getting from them. I thought a vague idea would be gathered by the bump in my average PPD, but since I had one physical box and one virtual machine decide to brick it over the last week, no joy there. But, after having had a play with some other work-related analysis programs, it would appear that it's not worth the cost for home folding by a long chalk, and only comparable to other co.s cloud computing svcs. Still, nice for the play with while we've had the chance.
I've had a fiddle and set up apache to get HFM to give me some stats, and one of the VMs appears to be giving me just under 5k ppd and the other nearly 11k ppd. No real idea why the discrepancy as they're both working on P6098 units and it's the 16GB one that's running the slower of the two, but maybe things will stabilise in a day or so.
On different servers? I find that setting up a bunch of 2 core servers has worked best, have a total of 20 each chucking out 5k PPD.
Yep. I only had one running for a week out of sheer laziness. If I get time I'll set up more but for now any PPD bump is more than I had previously. Do you know if there's a limit on the number we can create? I know I tried creating an 8-core and was rejected, but can we create dozens of duals?
Yeah the limit is 20GB of RAM per AZ- so 2x 10x 2vCPU+2GB RAM boxes. I was doing 12x 2vCPU's and 2x4vCPUs, and was getting about 38-42kPPD; now switched to all 2vCPU instances and clocking about 50kPPD instead. Shame its so expensive, but then, it's not a service that's really designed for us folders What it has acheived though is to get me hooked on folding!! New Gaming PC to be built with Christmas bonus me thinks...
So, free public beta ends on the 10th, think we can all get away with deleting our accounts on the 9th? Or would it be best doing it a few days before...
I was going to set my work units to finish on the evening of the 7th and then power them down on the 9th/once completed.
Same here; from the 7th onwards, once a WU is finished I'll be powering down the server and deleting the instance. Once they're all down, I'll be closing the account.
I'm gonna try skype as I got a few euros credit on it. Will have to wait until this afternoon probably for time difference.
To cancel the account (without having to call) all I did was email them using the details on their website stating my account number and that I had no instances running and could not phone them from the UK. I got an email back 24 hours later saying that they would process my cancelation.
Account closed. I dont think it took any of my skype credit either. I had to give them the tenent ID from the API key section of the dashboard.
Going to try closing mine tomorrow after last of these WUs finish. Will let you know how it goes, but don't envisage any problems.
Close request in, but one of my instances is still active(deleting). I have raised the issue with HP and they say its in hand, but that was 2 days ago now. Will see what happens by tomorrow.