LHC@home

Yes I agree, the DB performance has been an issue in the past, though the issue was mainly on the submitting end (with too many tasks being submitted at the same moment - this was dealt with by using a traffic light to avoid flooding the DB) and rarely on the retrieving end.

Still, something to keep in mind, and investigate how it scales when using GPUs.

Thanks a lot!

Indeed, I work at CERN on the LHC. I'm Alessio's successor (in being the link between the beams and IT departments concerning BOINC).
Also, as until last year I was the main user of BOINC, you've probably been crunching my jobs :-)

Thanks for your recent posts. They have answered many questions I have had about Sixtrack.

---

Hello,
Currently our teams are developing a new tracking code, named XTrack, as an addition/alternative to SixTrack.

It will be open source like Sixtrack?

It's still bad since there is hard coded back off in BOINC client of 90 sec, for me still my GPU can do a WU in less than 90 sec, so it tries to get new WU every 90 sec which is then a DoS attack basically.

Kronos Group released updated documentation of C++ for OpenCl

Interesting article: CERN’s LHC Experiments Increase Use of GPUs to Improve Computing Infrastructure

Just curious for XTrack (a part of the new XSuite) i find this pdf and the project seems to be ambitious (see page 15).
And this is the source of the platform on github

Just curious for XTrack (a part of the new XSuite) i find this pdf and the project seems to be ambitious (see page 15)...

I'm intrigued as to how the run-time compilation of C by a python app (p. 54 on) would translate to running on an arbitrary Boinc client, especially a Windows one.

---

I'm intrigued as to how the run-time compilation of C by a python app (p. 54 on) would translate to running on an arbitrary Boinc client, especially a Windows one.

Mmmmm, virtual machine?

I'm intrigued as to how the run-time compilation of C by a python app (p. 54 on) would translate to running on an arbitrary BOINC client, especially a Windows one.

Mmmmm, virtual machine?

But the big selling point of Sixtrack as a BOINC project is precisely that it Just Works, without all the shenanigans of VirtualBox or CVMFS or whatever...

---

But the big selling point of Sixtrack as a BOINC project is precisely that it Just Works, without all the shenanigans of VirtualBox or CVMFS or whatever...

+1

... and now SQUID. That is a lot to ask of volunteer crunchers.

---

... and now SQUID. That is a lot to ask of volunteer crunchers.

Set up once, runs for months.
Less than 1 h of work for a halfway experienced computer user.
Doesn't require to be an freak but to be open minded.

BTW:
SixTrack is a typical application that does not require internet access while it's tasks are in progress.
Hence, it does not benefit from a proxy.

... and now SQUID. That is a lot to ask of volunteer crunchers.

Set up once, runs for months...

More to the point, you don't even need to start worrying about Squid until after you've been through the VirtualBox/CVMFS palaver at least once (to create useful traffic for a Squid to cache), and then only for the larger scale crunchers.
If Sixtrack becomes as painful to get running as, say, Atlas, then surely this will be a big turn-off when trying to encourage new, casual volunteers to join the project.

---

... and now SQUID. That is a lot to ask of volunteer crunchers.

Set up once, runs for months...

More to the point, you don't even need to start worrying about Squid until after you've been through the VirtualBox/CVMFS palaver at least once (to create useful traffic for a Squid to cache), and then only for the larger scale crunchers.
If Sixtrack becomes as painful to get running as, say, Atlas, then surely this will be a big turn-off when trying to encourage new, casual volunteers to join the project.

+1

If you run vbox apps you get correctly configured CVMFS and Frontier clients inside the VM.
If you simply don't want to run vbox the project team would still be happy to see a SixTrack-only cruncher.

Mmmmm, virtual machine?

But the big selling point of Sixtrack as a BOINC project is precisely that it Just Works, without all the shenanigans of VirtualBox or CVMFS or whatever...

This is my hope. No virtual machine, only native application.
They said they want to try new sixtrack/xtrack on gpu and i think it's difficult with VM

I think the challenge for the project would be that virtualization of GPUs is considered an enterprise feature so its blocked in consumer GPU drivers. I doubt that volunteers are willing to use enterprise GPUs to contribute.

I think the challenge for the project would be that virtualization of GPUs is considered an enterprise feature so its blocked in consumer GPU drivers. I doubt that volunteers are willing to use enterprise GPUs to contribute.

I think our posts it's only speculations.
No one of "new-generation" wus arrived in our pcs (and we don't even know if it will be feasible), so the only thing we can do is to wait.
And, if necessary, we can test new apps in lhc-dev

And this is the source of the platform on github

Seems that works are going on: new version of Xtrack is 0.9.5

I'm intrigued as to how the run-time compilation of C by a python app (p. 54 on) would translate to running on an arbitrary Boinc client, especially a Windows one.

In this pdf, the way seems to be hybrid