I reran the non-PRO install procedure with the option --openctl=legacy and it installed two more packages than it did before, and both were -PRO packages. Now, clinfo and clpeak work, again.
<core_client_version>7.14.2</core_client_version>
<![CDATA[
<stderr_txt>
putenv 'LAL_DEBUG_LEVEL=3'
2020-01-13 01:32:44.4563 (124933) [normal]: This program is published under the GNU General Public License, version 2
2020-01-13 01:32:44.4564 (124933) [normal]: For details see http://einstein.phys.uwm.edu/license.php
2020-01-13 01:32:44.4564 (124933) [normal]: This Einstein@home App was built at: Oct 24 2019 08:47:59
2020-01-13 01:32:44.4564 (124933) [normal]: Start of BOINC application '../../projects/einstein.phys.uwm.edu/einstein_O2MDF_2.02_x86_64-pc-linux-gnu__GW-opencl-ati'.
[DEBUG} GPU type: 2
[DEBUG} got GPU info from BOINC
[DEBUG} got VendorID 4098
2020-01-13 01:32:44.4972 (124933) [debug]: BSGL output files
2020-01-13 01:32:44.4976 (124933) [debug]: Flags: LAL_DEBUG, OPTIMIZE, HS_OPTIMIZATION, GC_SSE2_OPT, X64, SSE, SSE2, GNUC X86 GNUX86
2020-01-13 01:32:44.4976 (124933) [debug]: glibc version/release: 2.30/stable
2020-01-13 01:32:44.497633 - mytime()
2020-01-13 01:32:44.4978 (124933) [debug]: Set up communication with graphics process.
DEPRECATION WARNING: program has invoked obsolete function XLALGetVersionString(). Please see XLALVCSInfoString() for information about a replacement.
Code-version: %% LAL: 6.19.2.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALPulsar: 1.17.1.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALApps: 6.23.0.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
2020-01-13 01:32:44.7330 (124933) [normal]: Reading input data ... 2020-01-13 01:32:48.1974 (124933) [normal]: Search FstatMethod used: 'ResampOpenCL'
2020-01-13 01:32:48.1976 (124933) [normal]: Recalc FstatMethod used: 'DemodSSE'
2020-01-13 01:32:48.1976 (124933) [normal]: OpenCL Device used for Search/Recalc and/or semi coherent step: 'Ellesmere (Platform: AMD Accelerated Parallel Processing, global memory: 7789 MiB)'
2020-01-13 01:32:48.1976 (124933) [normal]: OpenCL version is used for the semi-coherent step!
2020-01-13 01:32:51.2029 (124933) [normal]: Number of segments: 6, total number of SFTs in segments: 9873
done.
% --- GPS reference time = 1177858472.0000 , GPS data mid time = 1177858472.0000
2020-01-13 01:32:51.2512 (124933) [normal]: dFreqStack = 1.902483e-07, df1dot = 4.494707e-13, df2dot = 1.589301e-19, df3dot = 0.000000e+00
% --- Setup, N = 6, T = 2592000 s, Tobs = 19750204 s, gammaRefine = 21, gamma2Refine = 11, gamma3Refine = 1
DEPRECATION WARNING: program has invoked obsolete function InitDopplerSkyScan(). Please see XLALInitDopplerSkyScan() for information about a replacement.
2020-01-13 01:32:51.4220 (124933) [normal]: INFO: No checkpoint checkpoint.cpt found - starting from scratch
% --- Cpt:0, total:176, sky:1/1, f1dot:1/176
DEPRECATION WARNING: program has invoked obsolete function FreeDopplerSkyScan(). Please see XLALDestroyDopplerSkyScan() for information about a replacement.
2020-01-13 01:45:50.6382 (124933) [debug]: resultfile '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_0' (len 94), current config file: 0
2020-01-13 01:45:50.6382 (124933) [debug]: renaming '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_0-BSGLtL' to '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_1'
2020-01-13 01:45:50.6383 (124933) [debug]: renaming '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_0-BtSGLtL' to '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_2'
Code-version: %% LAL: 6.19.2.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALPulsar: 1.17.1.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALApps: 6.23.0.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
FPU status flags: PRECISION
2020-01-13 01:45:50.6388 (124933) [debug]: worker done. return(0) to caller
2020-01-13 01:45:50.6388 (124933) [normal]: done. calling boinc_finish(0).
01:45:50 (124933): called boinc_finish
</stderr_txt>
Now, this is the worst possible result I can imagine from BOINC. The work unit ran for a few minutes and gave a successful exit code...but the output was not bad and I didn't get any credit! Yikes, the worst of both worlds!
I can imagine starting a new thread for this, but I figured I would try here as it is not clear how this relates to the original problem. Let me know if you think I should.
You'll notice several strange warnings about obsolete functions, including FreeDopplerSkyScan(). Now, that sounds like a function specific to E@H app, not BOINC or OpenCL related. What am I to make of that? Is this normal? I can't see Gary's computers, so I don't know. HOLMIS's tasks don't show anything similar in their logs.
... Now, this is the worst possible result I can imagine from BOINC. The work unit ran for a few minutes and gave a successful exit code...but the output was not bad and I didn't get any credit! Yikes, the worst of both worlds!
I presume you do realise that you need a result returned from a quorum partner before validation can even be attempted in the first place?
Currently, for recent GW tasks, you have 16 in total and all 16 are 'pending' - in other words, none of the 16 quorum partners have yet returned their copy. You have no errors or invalids so there seems to be every indication you have a working system again. The crunch times look pretty much as you would expect for an RX 480 running x1. You should be able to make a reasonable improvement by running x3.
You just need to be patient until someone else returns their copy of one of your tasks. That will trigger a validation attempt and then you will know the true outcome. You can actually improve your chances of finding out more quickly by allowing your host to download some more work. You might score some 'resends' (_2 or higher suffix) which you could promote to crunch immediately. There's a better chance that the other quorum partner may have already returned their copy since it's likely to have been first issued some time ago.
There is no need for a different thread. This one is fine.
Paul wrote:
You'll notice several strange warnings about obsolete functions, including FreeDopplerSkyScan(). Now, that sounds like a function specific to E@H app, not BOINC or OpenCL related. What am I to make of that? Is this normal?
Yes, it's entirely normal. I guess the Einstein Devs build their apps against older library functions that still work but are supposedly deprecated. That's their decision - nothing to do with BOINC or your computer.
The only thing unusual about the log you posted is that you show no long lines of 'dots' with the occasional 'c' for when checkpoints are written. That's the most important thing to see since each 'dot' represents a calculation loop and if you see lots of them then you know that calculation loops are completing, just as they should. I did check one of your results to make sure they were actually there and you had just left them out of what you posted. As far as I can see from a cursory glance, everything looks fine. You just need validations.
So put that machine back to work and start cracking the whip!! :-). It looks like congratulations to you for solving your own problem are in order!! :-).
Yeah, I remember that, but I also remember getting credit almost immediately in that past. Now, it's been like 5 years since BOINC has worked properly, so I'm probably not remembering the delays accurately at this point. You are right though, that work hasn't been validated, so, maybe it is is fixed. Fingers crossed. And thanks for all your help; if it were not for your hard work, showing that it was possible, I wouldn't have kept trying.
You'll notice several strange warnings about obsolete functions, including FreeDopplerSkyScan(). Now, that sounds like a function specific to E@H app, not BOINC or OpenCL related. What am I to make of that? Is this normal? I can't see Gary's computers, so I don't know. HOLMIS's tasks don't show anything similar in their logs.
I've just check one of my GPU tasks from today and I see the following warnings about obsolete functions:
EPRECATION WARNING: program has invoked obsolete function XLALGetVersionString(). Please see XLALVCSInfoString() for information about a replacement.
DEPRECATION WARNING: program has invoked obsolete function InitDopplerSkyScan(). Please see XLALInitDopplerSkyScan() for information about a replacement.
DEPRECATION WARNING: program has invoked obsolete function FreeDopplerSkyScan(). Please see XLALDestroyDopplerSkyScan() for information about a replacement.
So the windows version also uses some of these deprecated functions. But as long as the project is good with that and my results validate I'm not going to worry. But I think it would look better if the applications were upgraded to use the newer replacement functions, as long as they don't have a negative impact on the science.
I see you have started 'cracking the whip' :-). For GW tasks, 126 in progress, 52 pending, 36 valid, with no errors or invalids. You also are receiving some GRP tasks, 2 completed, 3 in progress, so you must have both searches enabled.
Just be aware that the scheduler seems to be prioritized towards sending GW which is why you have so few GRP. In a way, that's fortunate because there wont be enough GRP completions to seriously lower the estimate for the GW tasks. As the GW tasks get completed (more slowly than their initial estimates) they will tend to dominate and will cause the estimate for GRP tasks to be rather higher than what they will actually take.
If it were the other way around - lots of GRP tasks being crunched - the fast times for those would seriously lower the estimate for the GW tasks and could cause you to get far more of them, possibly even to the point of exceeding the relatively short 7 day deadline for GW. Just something to keep an eye on.
Okay! Thanks HOLMIS and Gary! Good to hear from you both.
I'm glad that it is even the apps on another platform that show these deprecated functions; that' builds my confidence.
Also, thanks for telling me about my own validations. I guess I picked the wrong WU to track; I was just sitting one that I found had a copy out to someone else and was refreshing that page, but they aren't done with it. I check all the tasks again and sure enough, there are a bunch that got validated. OMG! It's working, ha!
I'm especially excited to see with I get from SETI@Home, since that is my oldest (very very oldest) BOINC project. But, seems I'm still having a bit of bad luck; I haven't received any new WU in over 24 hours, just before I turned on GPU apps again for the first time ages. S@H seems to be functioning, but maybe I'm not reading the server stats right.
In any case, I'm mostly excited because I think what I did was repeatable and not terribly too much work. That's different from the last 5 years.
More good news: SETI@Home is also crunching after a long, long time not working. I'm really pleased. I hope I've learned something that will last me a while; fingers crossed.
I reran the non-PRO install
)
I reran the non-PRO install procedure with the option --openctl=legacy and it installed two more packages than it did before, and both were -PRO packages. Now, clinfo and clpeak work, again.
So, I tried E@H again and here is what I got:
x86_64-pc-linux-gnu
Stderr output
2020-01-13 01:32:44.4564 (124933) [normal]: Start of BOINC application '../../projects/einstein.phys.uwm.edu/einstein_O2MDF_2.02_x86_64-pc-linux-gnu__GW-opencl-ati'.
[DEBUG} GPU type: 2
[DEBUG} got GPU info from BOINC
[DEBUG} got VendorID 4098
2020-01-13 01:32:44.4972 (124933) [debug]: BSGL output files
2020-01-13 01:32:44.4976 (124933) [debug]: Flags: LAL_DEBUG, OPTIMIZE, HS_OPTIMIZATION, GC_SSE2_OPT, X64, SSE, SSE2, GNUC X86 GNUX86
2020-01-13 01:32:44.4976 (124933) [debug]: glibc version/release: 2.30/stable
2020-01-13 01:32:44.497633 - mytime()
2020-01-13 01:32:44.4978 (124933) [debug]: Set up communication with graphics process.
DEPRECATION WARNING: program has invoked obsolete function XLALGetVersionString(). Please see XLALVCSInfoString() for information about a replacement.
Code-version: %% LAL: 6.19.2.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALPulsar: 1.17.1.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALApps: 6.23.0.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
2020-01-13 01:32:44.7330 (124933) [normal]: Reading input data ... 2020-01-13 01:32:48.1974 (124933) [normal]: Search FstatMethod used: 'ResampOpenCL'
2020-01-13 01:32:48.1976 (124933) [normal]: Recalc FstatMethod used: 'DemodSSE'
2020-01-13 01:32:48.1976 (124933) [normal]: OpenCL Device used for Search/Recalc and/or semi coherent step: 'Ellesmere (Platform: AMD Accelerated Parallel Processing, global memory: 7789 MiB)'
2020-01-13 01:32:48.1976 (124933) [normal]: OpenCL version is used for the semi-coherent step!
2020-01-13 01:32:51.2029 (124933) [normal]: Number of segments: 6, total number of SFTs in segments: 9873
done.
% --- GPS reference time = 1177858472.0000 , GPS data mid time = 1177858472.0000
2020-01-13 01:32:51.2512 (124933) [normal]: dFreqStack = 1.902483e-07, df1dot = 4.494707e-13, df2dot = 1.589301e-19, df3dot = 0.000000e+00
% --- Setup, N = 6, T = 2592000 s, Tobs = 19750204 s, gammaRefine = 21, gamma2Refine = 11, gamma3Refine = 1
DEPRECATION WARNING: program has invoked obsolete function InitDopplerSkyScan(). Please see XLALInitDopplerSkyScan() for information about a replacement.
2020-01-13 01:32:51.4220 (124933) [normal]: INFO: No checkpoint checkpoint.cpt found - starting from scratch
% --- Cpt:0, total:176, sky:1/1, f1dot:1/176
0.% --- CG:1578018 FG:262815 f1dotmin_fg:-1.938525403367e-08 df1dot_fg:2.140336666667e-14 f2dotmin_fg:-7.224095454545e-20 df2dot_fg:1.444819090909e-20 f3dotmin_fg:0 df3dot_fg:1
...
2020-01-13 01:45:14.6850 (124933) [normal]: Finished main analysis.
2020-01-13 01:45:14.6858 (124933) [normal]: Recalculating statistics for the final toplist...
2020-01-13 01:45:50.2132 (124933) [normal]: Finished recalculating toplist statistics.
2020-01-13 01:45:50.2499 (124933) [debug]: Writing output ... toplist2 ... toplist3 ... done.
DEPRECATION WARNING: program has invoked obsolete function FreeDopplerSkyScan(). Please see XLALDestroyDopplerSkyScan() for information about a replacement.
2020-01-13 01:45:50.6382 (124933) [debug]: resultfile '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_0' (len 94), current config file: 0
2020-01-13 01:45:50.6382 (124933) [debug]: renaming '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_0-BSGLtL' to '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_1'
2020-01-13 01:45:50.6383 (124933) [debug]: renaming '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_0-BtSGLtL' to '../../projects/einstein.phys.uwm.edu/h1_1149.00_O2C02Cl4In0__O2MDFG2e_G34731_1149.40Hz_277_0_2'
Code-version: %% LAL: 6.19.2.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALPulsar: 1.17.1.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
%% LALApps: 6.23.0.1 (CLEAN da7b29d537da983b9ab60a7f2a2a29fd91238c6a)
FPU status flags: PRECISION
2020-01-13 01:45:50.6388 (124933) [debug]: worker done. return(0) to caller
2020-01-13 01:45:50.6388 (124933) [normal]: done. calling boinc_finish(0).
01:45:50 (124933): called boinc_finish
</stderr_txt>
Now, this is the worst possible result I can imagine from BOINC. The work unit ran for a few minutes and gave a successful exit code...but the output was not bad and I didn't get any credit! Yikes, the worst of both worlds!
I can imagine starting a new thread for this, but I figured I would try here as it is not clear how this relates to the original problem. Let me know if you think I should.
You'll notice several strange warnings about obsolete functions, including FreeDopplerSkyScan(). Now, that sounds like a function specific to E@H app, not BOINC or OpenCL related. What am I to make of that? Is this normal? I can't see Gary's computers, so I don't know. HOLMIS's tasks don't show anything similar in their logs.
Paul wrote:... Now, this is
)
I presume you do realise that you need a result returned from a quorum partner before validation can even be attempted in the first place?
Currently, for recent GW tasks, you have 16 in total and all 16 are 'pending' - in other words, none of the 16 quorum partners have yet returned their copy. You have no errors or invalids so there seems to be every indication you have a working system again. The crunch times look pretty much as you would expect for an RX 480 running x1. You should be able to make a reasonable improvement by running x3.
You just need to be patient until someone else returns their copy of one of your tasks. That will trigger a validation attempt and then you will know the true outcome. You can actually improve your chances of finding out more quickly by allowing your host to download some more work. You might score some 'resends' (_2 or higher suffix) which you could promote to crunch immediately. There's a better chance that the other quorum partner may have already returned their copy since it's likely to have been first issued some time ago.
There is no need for a different thread. This one is fine.
Yes, it's entirely normal. I guess the Einstein Devs build their apps against older library functions that still work but are supposedly deprecated. That's their decision - nothing to do with BOINC or your computer.
The only thing unusual about the log you posted is that you show no long lines of 'dots' with the occasional 'c' for when checkpoints are written. That's the most important thing to see since each 'dot' represents a calculation loop and if you see lots of them then you know that calculation loops are completing, just as they should. I did check one of your results to make sure they were actually there and you had just left them out of what you posted. As far as I can see from a cursory glance, everything looks fine. You just need validations.
So put that machine back to work and start cracking the whip!! :-). It looks like congratulations to you for solving your own problem are in order!! :-).
Cheers,
Gary.
Hi Gary, Yeah, I remember
)
Hi Gary,
Yeah, I remember that, but I also remember getting credit almost immediately in that past. Now, it's been like 5 years since BOINC has worked properly, so I'm probably not remembering the delays accurately at this point. You are right though, that work hasn't been validated, so, maybe it is is fixed. Fingers crossed. And thanks for all your help; if it were not for your hard work, showing that it was possible, I wouldn't have kept trying.
I just checked and when I
)
I just checked and when I post this 5 tasks have validated, so I concur with Gary, congratulations on fixing the problem!
Now you can begin the fun task of optimizing your system!
Paul wrote:You'll notice
)
I've just check one of my GPU tasks from today and I see the following warnings about obsolete functions:
EPRECATION WARNING: program has invoked obsolete function XLALGetVersionString(). Please see XLALVCSInfoString() for information about a replacement.
DEPRECATION WARNING: program has invoked obsolete function InitDopplerSkyScan(). Please see XLALInitDopplerSkyScan() for information about a replacement.
DEPRECATION WARNING: program has invoked obsolete function FreeDopplerSkyScan(). Please see XLALDestroyDopplerSkyScan() for information about a replacement.
So the windows version also uses some of these deprecated functions. But as long as the project is good with that and my results validate I'm not going to worry. But I think it would look better if the applications were upgraded to use the newer replacement functions, as long as they don't have a negative impact on the science.
And now you're up to 8 validations!
@Paul I see you have started
)
@Paul
I see you have started 'cracking the whip' :-). For GW tasks, 126 in progress, 52 pending, 36 valid, with no errors or invalids. You also are receiving some GRP tasks, 2 completed, 3 in progress, so you must have both searches enabled.
Just be aware that the scheduler seems to be prioritized towards sending GW which is why you have so few GRP. In a way, that's fortunate because there wont be enough GRP completions to seriously lower the estimate for the GW tasks. As the GW tasks get completed (more slowly than their initial estimates) they will tend to dominate and will cause the estimate for GRP tasks to be rather higher than what they will actually take.
If it were the other way around - lots of GRP tasks being crunched - the fast times for those would seriously lower the estimate for the GW tasks and could cause you to get far more of them, possibly even to the point of exceeding the relatively short 7 day deadline for GW. Just something to keep an eye on.
Cheers,
Gary.
Okay! Thanks HOLMIS and
)
Okay! Thanks HOLMIS and Gary! Good to hear from you both.
I'm glad that it is even the apps on another platform that show these deprecated functions; that' builds my confidence.
Also, thanks for telling me about my own validations. I guess I picked the wrong WU to track; I was just sitting one that I found had a copy out to someone else and was refreshing that page, but they aren't done with it. I check all the tasks again and sure enough, there are a bunch that got validated. OMG! It's working, ha!
I'm especially excited to see with I get from SETI@Home, since that is my oldest (very very oldest) BOINC project. But, seems I'm still having a bit of bad luck; I haven't received any new WU in over 24 hours, just before I turned on GPU apps again for the first time ages. S@H seems to be functioning, but maybe I'm not reading the server stats right.
In any case, I'm mostly excited because I think what I did was repeatable and not terribly too much work. That's different from the last 5 years.
More good news: SETI@Home is
)
More good news: SETI@Home is also crunching after a long, long time not working. I'm really pleased. I hope I've learned something that will last me a while; fingers crossed.