Folding Forum

When a WU is issued, two servers are chosen to receive the result: A Work Server and a Collection Server. (Later WUs often use a different Collection Server.) Suppose one of the two servers gets "full" and can no longer accept uploads. That can only be corrected by "unloading" so...

FAH doesn't use all the capabilities of GPUs. It outputs no video, so all of the components that genrate high-resolution / high frame rates take up their share of chip real-estate is essentially wasted space from FAH's perspective. The same goes for components that generate tensors (for AI) and do h...

I have confirmed that 1002:73af is in the latest version of GPUs.txt. Have you rebooted after deleting GPUs.txt from the FAH Data directory?

Several Work Servers that normally distribute WUs for CPUs are off-line. I'm not sure why. It puts more pressure on the servers that do have WUs so they run out quickly. It takes quite a bit of time and thought for the scientists to develop new projects. FAH does not distribute "busy work"...

Yes, processing by the GPU and the CPU are independent of each other. Most GPUs earn points faster than most CPUs but by default, you'll run both. The slider does have only 3 settings but there probably are a lot more choices if you open FAHControl. GPUs are either ON or OFF, with no intermediate se...

I had posted before noticing the thread belonged in the Intel iGP subforum (serves me right for just hitting the active topics button). Anyway, the two systems in my sig are just the systems setup for folding. One is CPU only folding and the other is GPU only folding. 04:27:02: ********************...

FAH only records points for WorkUnits that are completed and the results uploaded. Progress on the assigned WU can be seen with either WebControl or FAHControl together with the Timeout and Deadline. ETA is calculated based on the recent rate of completion for that project but it is decidely inaccur...

The GPU folding core is copying data and instructions to the GPU, and then copying back results of the calculations on that data continuously. The word "continuously" is subject to interpretation. Like all I/O tasks, it moves data whenever there's data to be moved ... and then it waits wh...

The actual GPU utilization varies depending somewhat on the number of atoms in the protein. A small WU may be assigned to a powerful GPU and there might not be enough atoms to keep all the shaders busy. On the other hand, the GT730 is one of the slowest GPUs that we support -- and it has a comparati...

Projects 134xx are not really single projects but rather groups of Moonshot projects. Typically there's one group of smaller projects which are only assigned to narrow GPUs and another group of larger projects which are only assigned to wider GPUs. It's an attempt to assign the right projects to the...

I'm seeing very fits and starts, server stats is showing that A7 units are going out in massive bursts based on what I'm observing. I saw it spike up to 16k/hr then immediately drop to zero. If there are N work servers with CPU projects to distribute and you happen to receive the last WU from that ...

FACore_a8 supposedly fixes this type of error but it's still present in _a7 and older cores. Projects started with older cores will continue to use that core (including follow-in clones of the same study) until the study is finished. Requesting "large prime' numers of cores (including multiples...

Fixed.

It should be distributable in a few minutes.

As I said elsewhere The smaller card will work up to its limits on any protein. The larger card won't run into it's internal limitations by a smaller protein but it'll still work better than the smaller card. If you happen to get the most powerful card available, you'll be spending money on capabili...

A GPU gains it's speed by performing a large number of operations in parallel. Any program contains many operations which MUST operate serially. In FAH's case, if a protein has a large number of atoms, the force calculations can be parallelized. If there is a relatively small number of atoms, the se...