Your right in your assumption that BRP4 is a CPU only search but there is a GPU variant called BRP4G.
The log you linked tells med that you asked for GPU work but then I only see messages about the scheduler checking for CPU work and it's not able to match any to your host.
I don't understand why it wouldn't find CPU matches. I have some CPU wu running now (FGRPSSE). Is it saying I don't need any *more* CPU work? (That is also not true, but just a thought.)
I haven't messed with those preferences. All apps are checked.
I still suspect an ATI thing because I did a system update today. I got a new kernel and, looking back, I see there was a mesa update, too. That makes me suspicious of the platform matching.
You know, one thing that is a real problem with BOINC, now, is that the old historical log is missing (stdout.txt). I don't know what sub-sub-version of MESA BOINC found 12 hours ago, so I can't tell if that changed or not. I'm going to start making my own log of this, but it's really frustrating.
Sun 02 Oct 2016 02:21:00 PM | | OpenCL: AMD/ATI GPU 0: AMD PITCAIRN (DRM 2.45.0 / 4.7.5-200.fc24.x86_64, LLVM 3.8.0) (driver version 12.0.3, device version OpenCL 1.1 Mesa 12.0.3, 1024MB, 1024MB available, 1680 GFLOPS peak) Sun 02 Oct 2016 02:21:00 PM | | OpenCL: AMD/ATI GPU 1: AMD PITCAIRN (DRM 2.45.0 / 4.7.5-200.fc24.x86_64, LLVM 3.8.0) (driver version 12.0.3, device version OpenCL 1.1 Mesa 12.0.3, 1024MB, 1024MB available, 1680 GFLOPS peak)
I don't understand why it wouldn't find CPU matches. I have some CPU wu running now (FGRPSSE). Is it saying I don't need any *more* CPU work? (That is also not true, but just a thought.)
I was a bit sloppy with my previous reply, Boinc didn't ask for CPU work so the scheduler could not give any.
Paul_134 wrote:
I haven't messed with those preferences. All apps are checked.
I still suspect an ATI thing because I did a system update today. I got a new kernel and, looking back, I see there was a mesa update, too. That makes me suspicious of the platform matching.
That might indeed be part of the problem, at the moment there aren't any GPU tasks to send out so we have to wait for the staff to fix that first.
Paul_134 wrote:
You know, one thing that is a real problem with BOINC, now, is that the old historical log is missing (stdout.txt). I don't know what sub-sub-version of MESA BOINC found 12 hours ago, so I can't tell if that changed or not. I'm going to start making my own log of this, but it's really frustrating.
You should have stdoutdae.txt and stdoutdae.old in your Boinc data directory with the logs you seek. Unless the system update wiped that folder.
stdoutdae.txt hasn't been updated since Jun. I don't know what changed at that time, but I think I upgraded to 7.6.33 around then. You haven't noticed this?
I'm running on Windows and both of those files are being updated and both show recent changes. Maybe your upgrade of Boinc got it to use another data directory? Otherwise Boinc works differently on Linux but I don't think so. A restart of Boinc and check in the startup messages in the Event log should show the path to the data directory.
@Paul: Your right in your
)
@Paul:
Your right in your assumption that BRP4 is a CPU only search but there is a GPU variant called BRP4G.
The log you linked tells med that you asked for GPU work but then I only see messages about the scheduler checking for CPU work and it's not able to match any to your host.
Have you selected Binary Radio Pulsar Search (Arecibo, GPU) in your project preferensces?
Thanks for the help! I don't
)
Thanks for the help!
I don't understand why it wouldn't find CPU matches. I have some CPU wu running now (FGRPSSE). Is it saying I don't need any *more* CPU work? (That is also not true, but just a thought.)
I haven't messed with those preferences. All apps are checked.
I still suspect an ATI thing because I did a system update today. I got a new kernel and, looking back, I see there was a mesa update, too. That makes me suspicious of the platform matching.
You know, one thing that is a real problem with BOINC, now, is that the old historical log is missing (stdout.txt). I don't know what sub-sub-version of MESA BOINC found 12 hours ago, so I can't tell if that changed or not. I'm going to start making my own log of this, but it's really frustrating.
Sun 02 Oct 2016 02:21:00 PM | | OpenCL: AMD/ATI GPU 0: AMD PITCAIRN (DRM 2.45.0 / 4.7.5-200.fc24.x86_64, LLVM 3.8.0) (driver version 12.0.3, device version OpenCL 1.1 Mesa 12.0.3, 1024MB, 1024MB available, 1680 GFLOPS peak)
Sun 02 Oct 2016 02:21:00 PM | | OpenCL: AMD/ATI GPU 1: AMD PITCAIRN (DRM 2.45.0 / 4.7.5-200.fc24.x86_64, LLVM 3.8.0) (driver version 12.0.3, device version OpenCL 1.1 Mesa 12.0.3, 1024MB, 1024MB available, 1680 GFLOPS peak)
Paul_134 wrote:Thanks for the
)
I was a bit sloppy with my previous reply, Boinc didn't ask for CPU work so the scheduler could not give any.
That might indeed be part of the problem, at the moment there aren't any GPU tasks to send out so we have to wait for the staff to fix that first.
You should have stdoutdae.txt and stdoutdae.old in your Boinc data directory with the logs you seek. Unless the system update wiped that folder.
stdoutdae.txt hasn't been
)
stdoutdae.txt hasn't been updated since Jun. I don't know what changed at that time, but I think I upgraded to 7.6.33 around then. You haven't noticed this?
I'm running on Windows and
)
I'm running on Windows and both of those files are being updated and both show recent changes. Maybe your upgrade of Boinc got it to use another data directory? Otherwise Boinc works differently on Linux but I don't think so. A restart of Boinc and check in the startup messages in the Event log should show the path to the data directory.
Thanks again, HOLMIS. The
)
Thanks again, HOLMIS. The data directory is where I think it is. Are you on 7.6.33?
I'm still on 7.6.22.
)
I'm still on 7.6.22.
Hi HOLMIS. I started a new
)
Hi HOLMIS. I started a new thread.
This thread is quite old, but
)
This thread is quite old, but i found it searching a solution for AMD+mesa+boinc
tried to compile libclc now from github to make my AMD-card running under lubuntu..its saying
"libclc requires LLVM >= 4.0"
newest version i found is 3.9- any ideas anyone ? Do i have to wait until 4.0 is released ?
in this
)
in this thread
https://einsteinathome.org/content/christening-rx-460-debianstable
there isn't a solution to the LLVM question either, but maybe the hints mean something to you ;-)