Yes, I am sure that the M1 Mac mini with 8GB of memory is running 8 threads consistently. I think the potential difference (which I agree doesn't make a lot of sense and seems counter intuitive to me) could be what we discussed back in this thread between December 2nd and December 4th: https://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,42987_offset,33
Since I sometimes use this computer for other things during the week, I have a number of other programs on it running in the background. As we had discussed back in December, it counterintuitively seemed like opening several memory heavy applications actually helped with the BOINC agent being able to run more threads consistently. If you wanted to give it a try, you could try opening up a web browser and then open up a bunch of tabs (or otherwise open up a number of other applications) just to increase the memory pressure on your system and see if that causes the BOINC agent to start up some new threads.

In terms of MCM WUs that I'm averaging when crunching 24/7, over the past two days I'm seeing ~60-62 results per day. For the past 24 hours, the breakdown of the 60 completed results was 22 of the fast ones (37% of total) that took 1.68 hours on average and 38 of the slower ones (63% of total) that took 3.92 hours on average.

After I collect data for a few more days of running 8 threads simultaneously, I plan to limit the M1 Mac mini to 4 threads (via setting BOINC to only use 50% of CPUs) and see if I can reproduce the study I did previously in this post: https://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,42987_offset,45
From that past experience, I expect to see similar performance running only 4 threads vs 8 threads since as was previously discussed in this thread, the M1 chip has only 4 performance cores.

Since Apple has infinitely more resources than the MCM team, maybe you should ask Apple to fund this.

I don't have any inside information, but I doubt any current WCG projects will support the Apple M1 chips natively.

So, the question should be when will new WCG projects support the Apple M1 chips natively?

BTW, if the MCM team didn't visit the forum, we wouldn't get updates from them.

For several years now, few, if any researchers come to the forums and post. Well, all I can remember from recently are a couple of posts by Diogo from the OPN team about a year ago. Waiting for anyone from the research teams to come here and answer questions is, in my opinion, a waste of time. Best one can hope is a reply from the WCG team.

Support for the M1 architecture is likely very dependent on how many M1s are currently active at WCG.

I wonder how many M1s there are right now at WCG - a few hundred? I'm all for a Native M1 application for MCM, which is probably the only project where the investment in the application is worth since MIP is ending and OPN has OpenCL support, but right now, it's probably not something very high up on the to-do list, which as we all know, is a very long list for a small team.

You could try emailing WCG regarding this subject.