I successfully ran a BRP5 v1.38 task on Albert just now. BRP5 v1.36 had failed consistently on my workstation earlier but v1.38 looks good on my setup.
If your worried about the "(Not checkpointed)" part I'd say that the part of the calculations/code that generate this message runs too quick for a checkpoint to be necessary.
We deployed BRP5 v1.38 on albert (test project) which contains another set of upstream patches that should (hopefully) finally fix the issue. In order to limit our deployment effort we just released the Win32-CUDA and Win64-OpenCL versions.
@Rechenkuenstler/Claggy: please test and report back.
Thanks,
Oliver
What were the upstream fixes? I've seen the same issue with Raistmer's latest Seti v7 AMD GPU app.
I successfully ran a BRP5
)
I successfully ran a BRP5 v1.38 task on Albert just now. BRP5 v1.36 had failed consistently on my workstation earlier but v1.38 looks good on my setup.
Successful run (interrupted in the middle by an automatic reboot for a Windows update):
http://albert.phys.uwm.edu/result.php?resultid=980776
Thanks,
Gustav
RE: "Statistics: count
)
If your worried about the "(Not checkpointed)" part I'd say that the part of the calculations/code that generate this message runs too quick for a checkpoint to be necessary.
That message has a level of
)
That message has a level of "[INFO]", and that's indeed right. It's just an information we wanted to get logged (and not new anyway)...
Einstein@Home Project
RE: We deployed BRP5 v1.38
)
What were the upstream fixes? I've seen the same issue with Raistmer's latest Seti v7 AMD GPU app.
Claggy
Hi Claggy, Have a look at
)
Hi Claggy,
Have a look at the last ~16 commits to BOINC's api/boinc_api.cpp, the oldest one being b98bc30.
HTH,
Oliver
Einstein@Home Project