The linux QM app runs fine but it requires some packages to be installed. People should expect infinite number of workunits...

We will run this for ever.

I still think, the main obstacle to gain a wider LINUX contributor base is to solve the start-up error when two WUs start at the same time. In my view, this hinders the implementation of the app on newer computers with more than 4 threads/cores, as the other available computer threads/cores are not loaded by other BOINC projects.

The linux QM app runs fine but it requires some packages to be installed. People should expect infinite number of workunits...

We will run this for ever.

I still think, the main obstacle to gain a wider LINUX contributor base is to solve the start-up error when two WUs start at the same time. In my view, this hinders the implementation of the app on newer computers with more than 4 threads/cores, as the other available computer threads/cores are not loaded by other BOINC projects.

True. I'm using only 2c/4t and 4c CPUs because of that. Ryzen 8c/16t is crunching WCG.

You can sort of learn to live with its idiosyncrasies after a while. They are more an annoyance at first. But I wonder if a non multi-core version would fix the start-up problems? It would be worth looking into.

May I touch upon this matter again after 2,5-3 weeks... at the risk of being a pain in the neck... is there any progress regarding concurrent QC tasks? I would really like to use a 8-16 core CPU (instead of my i3) and run several jobs at the same time. Hope that the admins can take some time to follow up, in view of the many jobs we still have to crunch.

Thanks.

---

I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

Hey,

What are the recommended PC specs you recommend for these Linux job units? (minimal CPU and ram?) Thanks

---

Cruncher/Learner in progress.

I have five QC running at the moment on two machines, and they are taking from 230 MB to 260 MB. I would plan on at least 300 MB to be safe.

I now have one running at 1016 MB for a few minutes, but now down to 932 MB. It looks like the upper limit is a bit elastic.

I now have one running at 1016 MB for a few minutes, but now down to 932 MB. It looks like the upper limit is a bit elastic.

Just to confirm, that's 1GB per work unit?

I have five QC running at the moment on two machines, and they are taking from 230 MB to 260 MB. I would plan on at least 300 MB to be safe.

pardon me for jumping in, does that mean it is possible now to run several QC tasks at the same time? For example, 4 Jobs on the Ryzen 1700?

---

I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

It is possible to run as many as you want concurrently. A bug unfortunately prevents simultaneous starts. I am investigating possible workarounds but no timeline yet, sorry. Requirements are in some other thread; they are rather mild (only thing you need to have the gcc package installed).

I have five QC running at the moment on two machines, and they are taking from 230 MB to 260 MB. I would plan on at least 300 MB to be safe.

pardon me for jumping in, does that mean it is possible now to run several QC tasks at the same time? For example, 4 Jobs on the Ryzen 1700?

Best to just run 1 per computer via app_config or you'll eventually get two starting at once. They'll crash at once, and end up in this loop.

I now have one running at 1016 MB for a few minutes, but now down to 932 MB. It looks like the upper limit is a bit elastic.

Just to confirm, that's 1GB per work unit?

Yes, but now I see 1466 MB for a single work unit, the most I have seen.

Note however that I use an app_config to limit them to only one CPU core per work unit, but I do not limit the number of work units that can run at a time. It would undoubtedly be more efficient use of memory to allow the default value of four cores to run on a single work unit. It probably would not use much more memory than a single core, but what the maximum is I don't really know. I have 32 GB, so I have never really paid much attention to it. I use BOINCTasks to measure the memory use by the way.

It is possible to run as many as you want concurrently. A bug unfortunately prevents simultaneous starts.

Okay, thank you :)

---

I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

Requirements are in some other thread; they are rather mild (only thing you need to have the gcc package installed).

I have looked around and cannot find any requirements for this specific project. The only CPU requirements I see are here
http://www.gpugrid.net/join.php

But that looks like general use to me.

I will make my question a little more simple... would a core i3 work fine with this project, and if so, would a gen 1 be fine, gen 2, or newer? My budget for another PC dedicated to cpu work is about 100 USD. I do not mind for a small desktop either without a big gpu slot. Thank you.

---

Cruncher/Learner in progress.

I will make my question a little more simple... would a core i3 work fine with this project, and if so, would a gen 1 be fine, gen 2, or newer? My budget for another PC dedicated to cpu work is about 100 USD. I do not mind for a small desktop either without a big gpu slot. Thank you.

For questions like this, wuprop is a fantastic resource. Check the link and you can see things like time to complete for various CPUs and other requirements like memory. In this case looks like you may need 776.5 MB of RAM and an i3 should do just fine.

I will make my question a little more simple... would a core i3 work fine with this project, and if so, would a gen 1 be fine, gen 2, or newer? My budget for another PC dedicated to cpu work is about 100 USD. I do not mind for a small desktop either without a big gpu slot. Thank you.

For questions like this, wuprop is a fantastic resource. Check the link and you can see things like time to complete for various CPUs and other requirements like memory. In this case looks like you may need 776.5 MB of RAM and an i3 should do just fine.

Thank you very much.

---

Cruncher/Learner in progress.

It is possible to run as many as you want concurrently. A bug unfortunately prevents simultaneous starts. I am investigating possible workarounds but no timeline yet, sorry.

One simple solution would be creating a TEMP file in the application directory. The first thing a job does is trying to create this file. In case the create command fails (because the file is already there or a multiple create/write conflict occurred) the job must back off for a while. If successful, this job is allowed to start and the others will try again in a couple of seconds. The now starting job shall delete the TEMP file in a timely manner, so that the others can also get started. So the TEMP file actually works like a (do-not-start) flag. You may also stop all tasks sometime and delete the TEMP at regular intervals, to make sure there is no failing or cancelled task leaving the file in place forever.

Frankly there are also other and more professional things like semaphores available in the OS, but the above is possibly the fastest solution.

---

I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

...You may also stop all tasks

Edit: I mean to pause the tasks, not to cancel them of course. By the way, does the error show up also when multiple tasks are paused and then re-started/continued at the same time?

---

I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

I am having a lot of CPU Quantum Chemistry WUs error out as soon as they start on my Ubuntu machine. Some work but most error out. Is this inherent with the CPU app or is it my machine?

Below are the machine's tasks:
http://www.gpugrid.net/results.php?hostid=424454

Me too:

http://www.gpugrid.net/result.php?resultid=17510462