I need to know likely data output volume of a task

[kennywho]

G'day

I'm on satellite broadband as I live in an Australian regional area. As such I have a finite volume of data available per month. If I go over that I get throttled down to 128kbps / 128kbps until the reset date. There is no option to add data volume.

2 days ago, a task generated 6GB of data. That represents 25% of my monthly daytime quota. Thus I need to be able to accept or reject daytime (0700-0100) tasks based on likely data volume. Is there any way I can do that please?

If not, is there any way I can schedule a task to run between the hours of 0100 and 0700 as I have 160GB of data available during that time.

Hope you can help. btw what do the letters stand for in the task display 'Work Unit (PRCG)'. I'm guessing C might be cpus and G graphics but if so, how is the single figure interpreted?).
Oops, re the last I should have said I am running WIN10 Pro, Intel i7-4790, 8xcpu, AMD Radeon RS-235X.

Thanks, kenny

[davidcoton]

Not sure how you got through 6GB, AFAIK units are just not that big.
There is no scheduler built in to F@H, particularly not to allow folding but hold the upload. F@H as a whole depends on the quickest possible return of WUs because each return generates the next WU in the run.
Each Work Unit is identified using four numbers, in the format: Project(Run, Clone, Generation).

[Joe_H]

A single WU download is typically under 50 MB, most of the CPU ones I have been processing have been under 10 MB. The upload size is a few times larger. If you want some idea of the sizes of the work assignments you have been getting, that is recorded for both download and upload in the log files. This topic includes directions for finding the log files and posting if needed - viewtopic.php?f=24&t=26036.

In the same directory as the current log, the client will keep up to 16 previous logs in a folder named logs.

There will be some additional network traffic in the form of requests for work, not sure how much that will add up to. Web Control also queries the stats database periodically, so you probably will want to have that closed when you do not need it, folding continues without it being open.

Scheduling to run at a certain time, that is not built into the client. It can be scripted, the client has an interface usable for such purposes.

As for the PRCG, that translates to Project, Run, Clone and Generation. The first is obvious, the Run and Clone indicate variations in initial starting conditions, ad the last is how far along the path the WU is. Different runs and clones can run at the same time, but Gen 0 is the very first WU in a sequence. Once someone finishes and uploads the WU to the server, it is processed and becomes the initial state for the next WU in the sequence. That is Gen n needs to come back so Gen n+1 can be created and sent out.

[kiore]

I used to fold full time in South Sudan with a satellite monthly limit of 6Gb until throttling, they were much smaller units then but I had 4 GPUs running 24/7 so I understand your predicament. The usage you report seems quite large but maybe you are better off just running this part time you can download a unit and set to finish so you know you have completed one without slowing down the project and when ready start another and set to finish. I understand this not exactly pumping out work but the work you do will be on time and you can view the up/download amounts on your log to see how much it actually used.

[kennywho]

Thanks also to Kiore for the info. Are log times GMT or local to me or other?

[Joe_H]

Log times are GMT.

[kennywho]

Thanks Joe.