mattekure

Members
  • Posts

    206
  • Joined

  • Last visited

Everything posted by mattekure

  1. They are burning through the work units as fast as they can generate them. One of the admins over on the Rosetta@Home forums posted this earlier today.
  2. Mine is the same. I have 6 queued up, but usually I have twice that many.
  3. If you notice that your Rosetta@Home BOINC client isnt getting any work units, its because of the MASSIVE number of new users who have joined in, including a bunch of us. They went from around 150,000 GigaFLOPS before the virus, to over 570,000 GigaFLOPS now. If you take a look at the server status, scroll down to the "Tasks by application" section, they have largely run out of tasks to issue. There are currently 1139105 tasks currently in process. https://boinc.bakerlab.org/rosetta/server_status.php
  4. Out of Memory errors are caused when the docker tries to use more memory than the system has. Usually its because limits were not set within the docker. The best way to prevent those is to determine for your system, how much memory the docker is safe to use and calculate that as a percentage of your total system memory. So if you have a total of 8Gb, and you feel it is safe for the BOINC docker to take up 3GB of that, you calculate 3GB/8GB = 37.5%. In boinc, go to the Options->Computing Preferences, and switch to the "Disk and Memory" tab. set that percentage in the memory section for both the in use and not in use fields. With those fields set, BOINC will limit itself to the amount of memory you have said it could use. BOINC isnt super strict though, it can fluctuate a little above the set percentage, so you may want to round down like to 35% or whatever fits your system specs.
  5. also remember that "not in use" for boinc means no keyboard/mouse activity. so with unraid, it would always register as not in use.
  6. Yes. See the docker FAQ https://forums.unraid.net/topic/57181-docker-faq/?do=findComment&comment=566088
  7. The best place is in the Options->Computing Preferences. Look for the tab Disk and memory. Calculate what % of the ram you want it to be able to use and input that number in the bottom section. For example, you have 8gb, and say you want the docker to use 3gb. you would put 3/8 = 37%. that will prevent it from using too much.
  8. very good point I forgot to mention. Thanks. Within BOINC I also limit the RAM usage so it never goes above 50% of my 32Gb.
  9. Some considerations on using the BOINC docker for Rosetta@Home. Performance and Memory concerns. BOINC defaults to using 100% of the CPUs. Also, by default, Rosetta will process 1 task per cpu core/thread. So if you have an 8 core machine (16 with HT) it will attempt to process 16 tasks at once. Even if you set docker pinning to specific cores, the docker image will see all available cores and begin 1 task per core/thread. If you want to limit the number of tasks being processed, change the setting for using % of CPUs. So using the 8 core machine example above, setting it to 50% would process 8 tasks at a time. Regardless of how you set up CPU pinning. RAM and out of memory errors. Some of the Rosetta jobs can consume a lot of RAM. I have noticed individual tasks consuming anywhere between 500Mb-1.5Gb of RAM. You can find the memory a task is using by selecting the task, and clicking the properties button. If the task runs out of memory, it may get killed, wasting the work done or delaying it. It is helpful to balance the number of tasks you have running to the amount of RAM you have available. In the example machine above, if I am processing 8 tasks, I might expect the RAM usage to be anywhere from 4Gb to 10Gb. The docker FAQ has instructions on limiting the amount of memory the docker container uses, but be aware that processing too many tasks and running out of memory will just kill them and delay processing. My real world example. CPU: 3900X 12-core (24 w/ Hyperthreading) RAM 32GB Usage limit set to 50%, so processing only 12 tasks at a time. RAM limited to 14G, I could go a little higher, but havent needed to. Most tasks stay under 1Gb CPU pinning to almost all available cores Actual CPU usage looks like Since putting those restrictions on, I have had very stable processing and no out of memory errors.
  10. Yeah, it totally makes sense that the work would ramp up now, not be going down. and I think our Unraid group is doing a great thing with donating the computing power.
  11. I imagine more covid related tasks are being submitted relatively frequently.
  12. I'm not so sure. I see a bunch of tasks with covid in the name, not starting with rb_.
  13. it seems right when it starts. it downloads and them immediately throws the error I'll let it run for a day or two, see if it was just those particular ones and if any future ones throw errors.
  14. I got it working, but I am noticing that in the jobs it downloads, the ones where the "Application" field is "Rosetta Mini 3.78" throw a Computation error. the ones marked Rosetta 4.08 seem to be processing fine.
  15. Are there any instructions on how to switch over from the official boinc container to this one?
  16. yes. Mine are running around 50-55C with no issues. As others have stated, thats nowhere close to the maximum for that processor. As you can see in my pic, I did use CPU pinning to leave 1 cpu (2 threads) open for Unraid and other processes.
  17. Thank you for posting this and announcing it through CA. Joined Team Unraid, doing my part.
  18. One suggestion for the Boinc Install instructions. In order for the File->Select Computer option to show up, you have to switch to Advanced View. Simple view does not show that option.
  19. If there is anything you can do to support using the larger font sizes, it would be greatly appreciated.
  20. I am using the White theme. I found that decreasing the font size resolved it, but it makes it harder to read. I had originally set it to Very Large (bad eyes), and by changing it to Normal, its no longer scrunched together.
  21. I'm getting the numbers scrunched up for Encoder and Decoder Utilization