It looks like CurieDock is slower than RxDock on Linux.

Ryzen Threadripper 2990WX: 567
Ryzen 9 3900X: 4
Xeon(R) CPU E5-2683 v4: 562
EPYC 7502P: 135
etc...

It looks like CurieDock is slower than RxDock on Linux.

That is my conclusion too, having completed over 100 total (about equal of each) on two different Ubuntu machines, one Intel and one Ryzen.

Out of work. The queue is empty.

---

Reno, NV
Team: SETI.USA

That is my conclusion too, having completed over 100 total (about equal of each) on two different Ubuntu machines, one Intel and one Ryzen.

Now average processing rates are changed and CurieDock is the faster one...

Now average processing rates are changed and CurieDock is the faster one...

Yes, I see that too. Good work.

Does CurieMarieDock have checkpoints? I saw no sign than RxDock does; it can run for half an hour without writing a checkpoint,

If it doesn't have checkpoints, are the future tasks expected to be significantly longer than the current ones? If so, expect users in areas with less reliable power to drop out of the project when they start getting the longer tasks. And probably also users who don't want to leave their computers running overnight.

If you're not familiar enough with the application source code to add checkpoints, you could try the virtual machine method of adding checkpoints. The RNA World BOINC project hasn't created any new workunits for months, but the last 30 are still running.

http://www.rnaworld.de/rnaworld/

The virtual machine has checkpoints, although the application program doesn't.

Does CurieMarieDock have checkpoints? I saw no sign than RxDock does; it can run for half an hour without writing a checkpoint,

If it doesn't have checkpoints, are the future tasks expected to be significantly longer than the current ones? If so, expect users in areas with less reliable power to drop out of the project when they start getting the longer tasks. And probably also users who don't want to leave their computers running overnight.

If you're not familiar enough with the application source code to add checkpoints, you could try the virtual machine method of adding checkpoints. The RNA World BOINC project hasn't created any new workunits for months, but the last 30 are still running.

http://www.rnaworld.de/rnaworld/

The virtual machine has checkpoints, although the application program doesn't.

Thank you for the feedback! The tasks will be of the same length. Also, CurieMarieDock runs faster, so they will finish faster. We hope this will more or less eliminate the need for checkpoints. We use BOINC Wrapper, and the main problem is, there are either checkpoints or correct suspending of tasks/project by the client. The latter feature is highly demanded.

[snip]

Thank you for the feedback! The tasks will be of the same length. Also, CurieMarieDock runs faster, so they will finish faster. We hope this will more or less eliminate the need for checkpoints. We use BOINC Wrapper, and the main problem is, there are either checkpoints or correct suspending of tasks/project by the client. The latter feature is highly demanded.

You may need to explain how to use the latter feature, then. I'd expect rather few users who haven't already used BOINC for years to already know,

I mean the suspend options with buttons in BOINC Manager or from command line,

boinccmd --project URL suspend

It seems many people are using them. BOINC Manager keeps suspended tasks in memory, so the work is not lost until BOINC/computer is not stopped. Yes, ideally there would be checkpoints too (they are automated and do not require user's actions). Maybe we find a solution one day.

Hi folks! Update: New application (CurieMarieDock) need a glibc version 2.27 or higher! As we see, most Linux computers meet these requirements.

Bad news for me. My machines are all running openSUSE Leap 15.2 and this currently only offers glibc 2.26. So I can only run RxDock. Are there any considerations to change the dependency from glibc 2.27? Is there an benefit of glibc 2.27?

Greetings
Ritschie

I had the same problem, all "Curie" are failing because missing glibc 2.27.
I'm running OpenSuse Leap 15.2

11:38:27 (30909): wrapper: running /home/cfjh/boinc/slots/27/build/cmdock (-r target.prm -p "/home/cfjh/boinc/slots/27/data/scripts/dock.prm" -f htvs.ptc -i ligands.sdf -o docking_out)
/home/cfjh/boinc/slots/27/build/cmdock: /lib64/libm.so.6: version `GLIBC_2.27' not found (required by /home/cfjh/boinc/slots/27/build/libcuriedock.so.0)

Any hints?

Any hints?

Exclude CurieDock and run RxDock only.

Ritschie wrote:

hoarfrost wrote:

Hi folks! Update: New application (CurieMarieDock) need a glibc version 2.27 or higher! As we see, most Linux computers meet these requirements.

Bad news for me. My machines are all running openSUSE Leap 15.2 and this currently only offers glibc 2.26. So I can only run RxDock. Are there any considerations to change the dependency from glibc 2.27?

I am using openSUSE Leap 15.2 too. But I can run CurieMarieDock just fine, due to a hack which I implemented earlier for a Universe@home speedup. I posted the hack in "Questions and Answers : Unix/Linux : GLIBC_2.27 Req".

Ritschie wrote:

Is there an benefit of glibc 2.27?

As far as I can tell from the 2.27 release plan, there is nothing interesting in it; no major feature, and no fix which a distributor wouldn't backport if strictly needed as a hotfix.

Does CurieMarieDock have checkpoints?

Thank you for the feedback! The tasks will be of the same length. Also, CurieMarieDock runs faster, so they will finish faster. We hope this will more or less eliminate the need for checkpoints. We use BOINC Wrapper, and the main problem is, there are either checkpoints OR correct suspending of tasks/project by the client. The latter feature is highly demanded.

QuChemPedIA had problems getting checkpointing to work at first. Keep trying we need it every ten minutes not every hour.
BTW, I will never run a BOINC project in VirtualCatBox. You have a very nice native Linux ap, please don't ruin it by putting in a virtual box.
I have no idea what you mean by your OR statement. If a WU is suspended it restarts from the last checkpoint. If you have no checkpoints then it restarts from the beginning.

---

I have no idea what you mean by your OR statement. If a WU is suspended it restarts from the last checkpoint. If you have no checkpoints then it restarts from the beginning.

You are probably using "leave application in memory"? In that case, you don't need checkpoints if you don't reboot. But BOINC wrapper apparently implements one or the other, but not both.

I have no idea what you mean by your OR statement. If a WU is suspended it restarts from the last checkpoint. If you have no checkpoints then it restarts from the beginning.

You are probably using "leave application in memory"? In that case, you don't need checkpoints if you don't reboot. But BOINC wrapper apparently implements one or the other, but not both.

Seriously?
"...there are either checkpoints OR correct suspending of tasks/project by the client."

---

Hello everyone!

We are glad to announce the new application CurieMarieDock (cmdock). Currently for Windows 64bit and MacOS 64bit, and more platforms will follow. If you observe errors or undesirable effects in it, please report.

With this application, there is no need for installing MS Visual C++ Redistributable for Visual Studio.

CurieMarieDock is a molecular docking application developed as a fork of RxDock by the authors of the basic COVID.SI project: Marko Jukić, Nejc Ilc, Davor Sluga, Črtomir Podlipnik, and Gašper Tomšič.

Thank you everyone for participating.

Without installing MS Visual C++ Redistributable for Visual Studio all my tasks failed on my notebook.

After installing it they seems to run.

Here is an example of the error:

https://www.sidock.si/sidock/result.php?resultid=10671101