Best Settings for maximum output

Moderators: Site Moderators, FAHC Science Team

Post Reply
MrScott
Posts: 9
Joined: Tue Dec 12, 2017 12:27 am

Best Settings for maximum output

Post by MrScott »

This somewhat of a multi-part question but I trust there are some helpful souls will to advise.

1) I would like to know if it is possible to change my Folding USER name without losing my accrued points? I want to start my own Group under my company name and would like my points to carryover.

2) I have multiple computers I can login and FOLD when not in use. I frequently find, like today, I cannot get more than on PC connected at a time. Is this something in my setup? I have a 250MB DL and 50MB UL
Internet connection so I doubt it is my connection to the web.

3) One of my machines is a 56 Thread Dual XEON WORKSTATION with a GTX 980 SC GPU. What is the best settings to get maximum use out of the cores. The best I have been able to achieve in AUTOMATIC (default mode) is 71 % use from my CPU's.

Thanks for your assistance,

MrScott
bruce
Posts: 20910
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Best Settings for maximum output

Post by bruce »

Welcome to foldingforum.org, MrScott
 

1) No. Once accrued, point stay with that user name and that team.

Since it's possible (on the official points site) to search by passkey, however, you can start an account under a new name and then always remember to consider the total under the passkey to be your authentic total. There are other ways that work, too, such as creating a new account under MrScott@myemail.com or MrScott@email2.com.

2) I'm not aware of any such restrictions. I fold with multiple computers using the same Name/.Team/Passkey as long as you perform a fresh install on each one. If you clone your OS, however, the server will see both computers as the same one. As part of the installation procedure, each client is assigned a separate (hidden) identifier.

3) When folding with multi-threaded CPUs (or multiple CPUs) you will have to create multiple CPU slots. The servers treat each slot independently, assigning a different WU to each slot and collecting the results under the same account. Certain numbers of CPUs are restricted, but there still will be many ways to divide up 56. One should always be left free to move data to/from each GPU. I believe 32 is the most that can be assigned to a slot but you may wish to use smaller numbers. Each project is configured to be assigned to a specific range of CPU counts so if there are no projects that can use all 32, then you'll want to reduce that number to get assignments. The beta client adds a feature to negotiate between the client and server for some lower number that you've assigned if that can get you an assignment. If you use the V7.4.4 released client, it's pretty rigid.
 
Post Reply