For information -- the last time there was a beta test for MCM1 (October 2020) it included a 32-bit ARM version (which ran quite well, as far as i could tell, and used the system better than OPN1 does!). However, nothing further came of that at the time -- i suspect some other aspects of the beta test failed, so no version change (and no ARM version...)

So it is possible that once the Jurisica Lab folks aren't so busy solving current problems an ARM (non-Android) MCM1 program will be back on the radar -- after all, it is their project :-)

Cheers - Al.

Yeah, I am running a RPi4 myself, which is actually my currently unutilized personal web server.
Currently, it is only crunching OPN1, when available, but I am pretty sure that in the past, I had different projects running on it as well.

Providing a different platform is a bit more involved than most people might think. It requires at times a completely new tool chain to produce the target binary that is doing the actual crunching, it requires testing in house, before it can be released for a public beta and then finally for "general consumption". That all requires resources, which some research teams may or may not have. And that all applies not only to different CPUs as in the case of ARM vs x86, but also different OS. Of which there may or may not be up to 4 at this point (Windows, Linux, macOS, Android). And GPU usage commonly requires a complete redesign/rethinking of the computation process, due to the highly parallel nature of those GPUs.

If we will see any other WCG projects including ARM in future iterations might be seen. Yes, would be nice, regardless how "efficient" the effort might be seen by some, but then the whole "distributed computing" concept using volunteers over things like BOINC should be more inclusive than just considering "power crunchers and their GPU farms". It all got started, and it is still the major "selling point" of projects, that spare resources of pretty much any computer could be used for a common goal, a typical "strength through numbers" scheme.

Ralf

Well, it's actually not that simple (what issue in computer land is ). Most newer (certainly ARM64) ARM can actually switch endianess, some even on the fly and that might be a difference also between 32bit and 64bit ARM code. So using gcc might make things easier, but there is no guarantee. I do vividly remember issues "way back when" on the RPi 2 that certain application programs crashed because while using the gcc on that machine itself, which was big-endian, some of the libraries (sqlite?) had been compiled on a different system which was little-endian. On those ARM6 CPUs, endianess was either fixed on the die, one way or the other, or on some versions of the chip, it was selectable upon boot initialization but then fixed for the time the machine was powered on.
As mentioned, rather than "WCG" it's an issue of the efforts by the research team behind any of the sub projects. And as mentioned, it could result in having to do things even for 3 different platforms, x86, ARM and GPU (and to make things even more interesting, the code would have to be developed and tested for possibly more than 1 type of GPU, my programming machine, when crunching Einstein@Home, gets both NVidia & Intel GPU WUs, while OPNG runs only on the NVidia chip).
WCG became quickly my preferred DC project, for multiple reasons.
For one, I had started many, many moons ago on SETI@Home, well pre-BOINC. When they the switched to using BOINC, I had to drop out because the first versions of the BOINC client use made the host machine unusable for doing actual work on the machine. Didn't crunch for a while until I a) had a short bout with stomach cancer and b) was pointed into the direction of Rosetta@Home, which runs in a similar setup/situation like WCG, with Baker Lab being "housed" at WSU. And R@H had a lot of issues in those days, with a +8h outage at least once a week, though this has gotten better since. So on January 2nd, 2011 I signed up and attached to WCG on all of my clients. And for the longest time, WCG ran just smooth as a Rolls Royce, barely any hitch, and if there was any kind of issues, the techs responded and fixed things in no time. Things got a bit worse when they started to switch from "real iron" to cloud based clusters. But still WCG seemed to have been leaps and bounds over anything else out there. But unfortunately, this all started to change a year ago

In general, I consider the type of projects that have been processed by WCG more "worthy" than hunting for signals from space or prime numbers or the like, which all seem to be en vogue by a lot of folks because they "pay more"...

Let' see how this all works out, but so far the reaction time from Krembil to ANY issue has been rather underwhelming...

Ralf