Problem with 171.67.108.54?

[zodac]

Many people with AMD 5/6 series GPUs on my team have been doing 511pt WUs for the past few days, after consistently going 1,835pt WUs for weeks.

The server status page shows the server as fine, but the "WUs Avail/WUs To Go/WUs Wait" columns are all blank. Is it just a shortage, or something else?

[7im]

Projects come and go, and the next project may not bench out to the same total points, especially when the WUs are different sizes. Unless the server is down, there is nothing wrong, just natural progression of the project, or simply a hiatus while they crunch some results, and then release more WUs in the project, if needed.

[zodac]

Yeah, but all the WUs they're getting now are 511pt WUs, which are Core11. There are only two Core16 servers on the psummary page; one is only accepting, and the other is the one I'm talking about.

[7im]

Just like bigadv work units for SMP clients, there are not always work units available for all types of cores.

As always, fold what you get.

[zodac]

You're not really answering the question. I'm not complaining about the WUs and saying there should be other ones.

I'm asking if there's a reason there are very few Core16 WUs compared to the past few weeks.

[7im]

And you're not really getting the answer.

Yes, there is always a reason. WU volumes fluctuate up and down naturally as one project ends, and another one starts. Or while they are doing maintenance on a server. Etc. Asking for a specific reason is like asking why the wind blows harder today than it did yesterday.

If you must have the exact reason why the wind is blowing less today, then I suggest sending a PM a researcher doing fahcore_16 projects. Please share if you get an answer.

[zodac]

Aye, but usually when there are WU fluctuations, there are other similar WUs available (not including -bigadv shortages). Since this is the only server which is listed as active, I figured this was the place to actually ask about it.

[7im]

WU availability fluctuations aren't really considered a server problem, but this is still the best place to ask. Although a researcher may not respond to a normal occurance, hence the PM suggestion.