Yesterday I got 4 tasks from the 10fe09ab.22806.* splitter run with short estimates, today 2 more from 10fe09ab.1847.* . All have

<autocorr_thresh>1000</autocorr_thresh>
<autocorr_per_spectrum>0</autocorr_per_spectrum>
<autocorr_fftlen>8</autocorr_fftlen>

rather than the usual settings. I suspect that's an accident.

The splitter estimate has a term which depends on the autocorr_fftlen, and doing DCTs at those short lengths will actually be a lot faster, so the reduced estimates may be OK.

In theory, all app versions ought to be able to handle the unusual processing, but probably none have actually been previously tested that way so there may be surprises. Beta testing is often interesting.

I'll ask Eric to comment if he has time.

                                                                  Joe

Hey Joe,

I think I ran across a bunch of those MB with the unusual auto correct. My 2 machines here both just hit the wall.

Ok,

found this. Now to decipher it slot 0


setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0


slot 6

setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 2
setiathome_CUDA: CUDA Device 2 specified, checking...
Device 2: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1123 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 2
setiathome_CUDA: CUDA Device 2 specified, checking...
Device 2: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1123 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0

slot 9

setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
setiathome_CUDA: Found 3 CUDA device(s):
Device 1: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 5, pciSlotID = 0
Device 2: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 1, pciSlotID = 0
Device 3: GeForce GTX 780, 3072 MiB, regsPerBlock 65536
computeCap 3.5, multiProcs 12
pciBusID = 2, pciSlotID = 0
In cudaAcc_initializeDevice(): Boinc passed DevPref 1
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0


I'm going to install the new version of BOINC, as Richard states it keeps a record for you all to look at. Right now both machines have 50 of these MB in a wait exit. I've not aborted them yet but not sure what to do with them

http://setiathome.berkeley.edu/forum_thread.php?id=76350&postid=1617018

Edit..

Here is link to new work unit

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6782596

I'm benching this Wu on my i5-3210M/GT650M/HD_Graphics_4000 host, just with the Stock 7.00 app, the AVX_2549 app, and the Intel_r2489 app,
I expect it to take 6 to 7 hours.

Claggy

Hey Joe,

I think I ran across a bunch of those MB with the unusual auto correct. My 2 machines here both just hit the wall.

Ok,

found this. Now to decipher it slot 0
...
Device 1: GeForce GTX 780 is okay
SETI@home using CUDA accelerated device GeForce GTX 780
pulsefind: blocks per SM 4 (Fermi or newer default)
pulsefind: periods per launch 100 (default)
Priority of process set to BELOW_NORMAL (default) successfully
Priority of worker thread set successfully

setiathome enhanced x41zc, Cuda 4.20

Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 0k elements.
Work Unit Info:
...............
WU true angle range is : 0.545577

Kepler GPU current clockRate = 1071 MHz

re-using dev_GaussFitResults array for dev_AutoCorrIn, 256 bytes
re-using dev_GaussFitResults+32x8 array for dev_AutoCorrOut, 256 bytes
Thread call stack limit is: 1k
Error on launch (ac_reducePartial<<<grid3, block3,blksize*sizeof(float3)>>>( (float *)dev_AutoCorrIn, dev_ac_partials )), file c:/[Projects]/__Sources/sah_v7_opt/Xbranch/client/cuda/cudaAcc_autocorr.cu, line 200: invalid configuration argument
Exiting
cudaAcc_free() called...
cudaAcc_free() running...
cudaAcc_free() PulseFind freed...
cudaAcc_free() Gaussfit freed...
cudaAcc_free() AutoCorrelation freed...
cudaAcc_free() DONE.
Cuda sync'd & freed.
Preemptively acknowledging a safe temporary exit->
Exit Status: 0
boinc_exit(): requesting safe worker shutdown ->
boinc_exit(): received safe worker shutdown acknowledge ->
Cuda threadsafe ExitProcess() initiated, rval 0
...
...

I'm going to install the new version of BOINC, as Richard states it keeps a record for you all to look at. Right now both machines have 50 of these MB in a wait exit. I've not aborted them yet but not sure what to do with them

I think just aborting them is sensible at this point. All attempts at processing them with CUDA app versions seem to get that same "...cudaAcc_autocorr.cu, line 200: invalid configuration argument" error which leads to 100 temporary exits, then BOINC kills the task for too many.

OpenCL app versions seem to be processing at least some of them correctly, but not all. The stock CPU app versions seem to be the only ones which are really unaffected. There will probably be many resends, and quite likely many of these WUs will end up with too many error tasks.

=================

I did get an email reply from Eric, possibly a recent cleanup in the pfb_splitter code introduced a bug.

                                                                  Joe

I did get an email reply from Eric, possibly a recent cleanup in the pfb_splitter code introduced a bug.

                                                                  Joe

My initial concern was the various straws in the wind we've had lately about Green Bank telescope data - staff working hard on GBT splitters, and the GPUUG buying new hardware for the recorder. If the changed parameters had been needed to match the GB telescope and recorder characteristics, then we might need to start work on an app refresh smartish. But 10fe09ab looks like a standard Arecibo tape from here, though at an unusual AR (0.545576...)

My Bench completed quicker than expected, Stock 7.00 and AKv8c_r2549 are Strongly similar:

------------
setiathome_7.00_windows_intelx86.exe -verb / 10fe09ab.4848.25021.140733193388047.16.119.wu :
AppName: setiathome_7.00_windows_intelx86.exe
AppArgs: -verb
TaskName: 10fe09ab.4848.25021.140733193388047.16.119.wu
Started at : 19:27:15.726
Ended at : 20:36:09.674
Result : stored as ref for validations.
4133.168 secs Elapsed
4104.745 secs CPU time

[ stderr ]
19:27:16 (18608): Can't set up shared mem: -1. Will run in standalone mode.
setiathome_v7 7.00 DevC++/MinGW/g++ 4.5.2
libboinc: 7.1.0

Work Unit Info:
...............
WU true angle range is : 0.545577

Optimal function choices:
--------------------------------------------------------
                            name   timing   error
--------------------------------------------------------
                v_BaseLineSmooth (no other)

              v_GetPowerSpectrum 0.000170 0.00000  test
             v_vGetPowerSpectrum 0.000057 0.00000  test
            v_vGetPowerSpectrum2 0.000100 0.00000  test
     v_vGetPowerSpectrumUnrolled 0.000058 0.00000  test
    v_vGetPowerSpectrumUnrolled2 0.000078 0.00000  test
           v_avxGetPowerSpectrum 0.000057 0.00000  test
             v_vGetPowerSpectrum 0.000057 0.00000  choice

                     v_ChirpData 0.005020 0.00000  test
                   fpu_ChirpData 0.013160 0.00000  test
               fpu_opt_ChirpData 0.005012 0.00000  test
             v_vChirpData_x86_64 0.065172 0.00000  test
               sse1_ChirpData_ak 0.009057 0.00000  test
             sse1_ChirpData_ak8e 0.007418 0.00000  test
             sse1_ChirpData_ak8h 0.006602 0.00000  test
               sse2_ChirpData_ak 0.006690 0.00000  test
              sse2_ChirpData_ak8 0.004618 0.00000  test
               sse3_ChirpData_ak 0.006475 0.00000  test
              sse3_ChirpData_ak8 0.004523 0.00000  test
                 avx_ChirpData_a 0.002521 0.00000  test
                 avx_ChirpData_b 0.003021 0.00000  test
                 avx_ChirpData_c 0.002960 0.00000  test
                 avx_ChirpData_d 0.002856 0.00000  test
                 avx_ChirpData_a 0.002521 0.00000  choice

                     v_Transpose 0.006639 0.00000  test
                    v_Transpose2 0.005101 0.00000  test
                    v_Transpose4 0.004064 0.00000  test
                    v_Transpose8 0.006608 0.00000  test
                  v_pfTranspose2 0.004339 0.00000  test
                  v_pfTranspose4 0.003793 0.00000  test
                  v_pfTranspose8 0.006166 0.00000  test
                   v_vTranspose4 0.002270 0.00000  test
                 v_vTranspose4np 0.002518 0.00000  test
                v_vTranspose4ntw 0.007721 0.00000  test
              v_vTranspose4x8ntw 0.003637 0.00000  test
             v_vTranspose4x16ntw 0.001134 0.00000  test
            v_vpfTranspose8x4ntw 0.007763 0.00000  test
            v_avxTranspose4x8ntw 0.003865 0.00000  test
           v_avxTranspose4x16ntw 0.001003 0.00000  test
            v_avxTranspose8x4ntw 0.007391 0.00000  test
          v_avxTranspose8x8ntw_a 0.004100 0.00000  test
          v_avxTranspose8x8ntw_b 0.004151 0.00000  test
           v_avxTranspose4x16ntw 0.001003 0.00000  choice

                 FPU opt folding 0.002286 0.00000  test
                 ben SSE folding 0.000664 0.00000  test
                  AK SSE folding 0.000540 0.00000  test
                  BH SSE folding 0.000476 0.00000  test
                JS AVX_a folding 0.000429 0.00000  test
                JS AVX_c folding 0.000503 0.00000  test
                JS AVX_a folding 0.000429 0.00000  choice

                   Test duration     4.99 seconds

Flopcounter: 13601301381159.039000

Spike count: 5
Autocorr count: 0
Pulse count: 0
Triplet count: 0
Gaussian count: 0
20:36:06 (18608): called boinc_finish
[ /stderr ]
------------
AKv8c_r2549_winx86-64_AVXxjfs.exe / 10fe09ab.4848.25021.140733193388047.16.119.wu :
AppName: AKv8c_r2549_winx86-64_AVXxjfs.exe
AppArgs:
TaskName: 10fe09ab.4848.25021.140733193388047.16.119.wu
Started at : 20:36:13.024
Ended at : 21:08:17.143
1923.779 secs Elapsed
1910.076 secs CPU time
Speedup : 53.47%
Ratio : 2.15x

R2: .\ref\ref-setiathome_7.00_windows_intelx86.exe-10fe09ab.4848.25021.140733193388047.16.119.wu.res
Result : Strongly similar, Q= 99.98%

[ stderr ]
20:36:13 (66084): Can't set up shared mem: -1. Will run in standalone mode.
Not using mb_cmdline.txt-file, using commandline options.

Build features: SETI7 Non-graphics FFTW USE_AVX x64
CPUID: Intel(R) Core(TM) i5-3210M CPU @ 2.50GHz

Cache: L1=64K L2=256K

CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSSE3 SSE4.1 SSE4.2 AVX
ar=0.545577 NumCfft=86985 NumGauss=433167160 NumPulse=56419559943 NumTriplet=112908674051
In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768

Windows optimized S@H v7 application
Based on Intel, Core 2-optimized v8-nographics V5.13 by Alex Kan
AVXxjf Win64 Build 2549 , Ported by : Raistmer, JDWhale

SETI7 update by Raistmer
Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.545577
Spike: peak=24.46589, time=33.55, d_freq=1421157876.12, chirp=5.3053, fft_len=128k
Spike: peak=24.58183, time=10.07, d_freq=1421162188.71, chirp=8.3701, fft_len=64k
Spike: peak=25.84634, time=10.07, d_freq=1421162188.71, chirp=8.3849, fft_len=64k
Spike: peak=25.60888, time=10.07, d_freq=1421162188.71, chirp=8.3997, fft_len=64k
Spike: peak=24.25809, time=46.98, d_freq=1421160606.28, chirp=-9.5513, fft_len=128k

Best spike: peak=25.84634, time=10.07, d_freq=1421162188.71, chirp=8.3849, fft_len=64k
Best autocorr: peak=0, time=-2.121e+011, delay=0, d_freq=0, chirp=0, fft_len=0
Best gaussian: peak=0, mean=0, ChiSq=0, time=-2.121e+011, d_freq=0,
score=-12, null_hyp=0, chirp=0, fft_len=0
Best pulse: peak=7.396154, time=102.5, period=2.307, d_freq=1421157416.34, score=0.894, chirp=99.399, fft_len=256
Best triplet: peak=0, time=-2.121e+011, period=0, d_freq=0, chirp=0, fft_len=0


Flopcounter: 11494065712493.554688

Spike count: 5
Autocorr count: 0
Pulse count: 0
Triplet count: 0
Gaussian count: 0
Wallclock time elapsed since last restart: 1919.9 seconds

21:08:14 (66084): called boinc_finish

The Intel r2489 app didn't run properly:

------------
MB7_win_x86_SSE_OpenCL_Intel_r2489.exe / 10fe09ab.4848.25021.140733193388047.16.119.wu :
AppName: MB7_win_x86_SSE_OpenCL_Intel_r2489.exe
AppArgs:
TaskName: 10fe09ab.4848.25021.140733193388047.16.119.wu
Started at : 21:08:22.363
Ended at : 21:08:36.243
13.760 secs Elapsed
7.909 secs CPU time
Speedup : 99.81%
Ratio : 519.00x

R2: .\ref\ref-setiathome_7.00_windows_intelx86.exe-10fe09ab.4848.25021.140733193388047.16.119.wu.res 
                ------------- R1:R2 ------------     ------------- R2:R1 ------------
                Exact  Super  Tight  Good    Bad     Exact  Super  Tight  Good    Bad
        Spike      0      0      0      0      0        0      0      0      0      5
     Gaussian      0      0      0      0      0        0      0      0      0      0
        Pulse      0      0      0      0      0        0      0      0      0      0
      Triplet      0      0      0      0      0        0      0      0      0      0
   Best Spike      0      0      0      0      0        0      0      0      0      1
Best Gaussian      0      0      0      0      0        0      0      0      0      1
   Best Pulse      0      0      0      0      0        0      0      0      0      1
 Best Triplet      0      0      0      0      0        0      0      0      0      0
                ----   ----   ----   ----   ----     ----   ----   ----   ----   ----
                   0      0      0      0      0        0      0      0      0      8

Unmatched signal(s) in R2 at line(s) 675 691 707 723 739 755 771 792
Result      : Different.

[ stderr ]
21:08:23 (66432): Can't set up shared mem: -1. Will run in standalone mode.
Not using mb_cmdline.txt-file, using commandline options.
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
GPU device # not found in init_data.xml
WARNING: BOINC was unable to find GPU device, using own enumeration
OpenCL platform detected: Intel(R) Corporation
WARNING: BOINC supplied wrong platform!
BOINC assigns device 0
WARNING: BOINC failed to provide OpenCL device, using own enumeration abilities

Build features: SETI7 Non-graphics OpenCL USE_OPENCL_INTEL OCL_SYNCHED OCL_CHIRP3 FFTW AMD specific USE_SSE x86
CPUID: Intel(R) Core(TM) i5-3210M CPU @ 2.50GHz

Cache: L1=64K L2=256K

CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSSE3 SSE4.1 SSE4.2 AVX
OpenCL-kernels filename : MultiBeam_Kernels_r2489.cl
ar=0.545577 NumCfft=86985 NumGauss=433167160 NumPulse=56419559943 NumTriplet=112908674051

Currently allocated 146 MB for GPU buffers
In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768

Windows optimized S@H v7 application
Based on Intel, Core 2-optimized v8-nographics V5.13 by Alex Kan
SSEx Win32 Build 2489 , Ported by : Raistmer, JDWhale

SETI7 update by Raistmer

OpenCL version by Raistmer, r2489

Number of OpenCL platforms: 2


OpenCL Platform Name: Intel(R) OpenCL
Number of devices: 1
Max compute units: 16
Max work group size: 512
Max clock frequency: 1100Mhz
Max memory allocation: 340158054
Cache type: Read/Write
Cache line size: 64
Cache size: 2097152
Global memory size: 1360632218
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 65536
Queue properties:
Out-of-Order: No
Name: Intel(R) HD Graphics 4000
Vendor: Intel(R) Corporation
Driver version: 10.18.10.3621
Version: OpenCL 1.2
Extensions: cl_intel_d3d11_nv12_media_sharing cl_intel_dx9_media_sharing cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_depth_images cl_khr_dx9_media_sharing cl_khr_gl_depth_images cl_khr_gl_event cl_khr_gl_msaa_sharing cl_khr_gl_sharing cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_image2d_from_buffer cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_intel_accelerator cl_intel_motion_estimation


OpenCL Platform Name: NVIDIA CUDA
Number of devices: 1
Max compute units: 2
Max work group size: 1024
Max clock frequency: 950Mhz
Max memory allocation: 536870912
Cache type: Read/Write
Cache line size: 128
Cache size: 32768
Global memory size: 2147483648
Constant buffer size: 65536
Max number of constant args: 9
Local memory type: Scratchpad
Local memory size: 49151
Queue properties:
Out-of-Order: Yes
Name: GeForce GT 650M
Vendor: NVIDIA Corporation
Driver version: 344.75
Version: OpenCL 1.1 CUDA
Extensions: cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_d3d9_sharing cl_nv_d3d10_sharing cl_khr_d3d10_sharing cl_nv_d3d11_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64


Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.545577
Used GPU device parameters are:
Number of compute units: 16
Single buffer allocation size: 64MB
max WG size: 512
period_iterations_num=20
ERROR: OpenCL kernel/call 'RepackInput_kernel' call failed (-54) in file ..\autocorr.cpp near line 480.
Waiting 30 sec before restart...

Claggy

The server is sending out those units again.
10fe09ab.4848.25021.140733193388047
All created today.

I get driver restarts so had to abort them.

---

With each crime and every kindness we birth our future.

Can someone please shoot me a PM over on main when this stops? Had to abort a few tasks too and set Beta to NNT.

Thanx and Happy New Year!!:D

Yup they are still here. 150 of them today.

None of them would start, Checked the stderr report, all had the same problems.

Gave up and had to abort all of them, now the server won't sent anymore of them.

Looks like going to have to be a daily event till they are all gone.

I`ve been told Jeff is working on a fix.

---

With each crime and every kindness we birth our future.

I`ve been told Jeff is working on a fix.

It might have already been fixed and deployed, My PIII has a fresh task that has been completed and validated by my wingmen:

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6799436

Can someone cancel all that broken work please.

Claggy

Had several of the 10fe09ab.*.vlar_8 task yesterday show up in my list.
Looking at the list they were marked as Invalid for EVERYONE who crunched them.
Wonder if they all have been sent out?

---

Had several of the 10fe09ab.*.vlar_8 task yesterday show up in my list.
Looking at the list they were marked as Invalid for EVERYONE who crunched them.
Wonder if they all have been sent out?

No, recent splitter runs have also shown the problem. Tasks I've been sent which were split yesterday, such as WU 6806063 are examples.

It seems maybe whenever a new splitter run starts, the first group of 256 WUs have a correct analysis_cfg section but subsequent groups have the bad parameters.

                                                                   Joe

Sorry, Jeff was supposed to have fixed those analysis configs last week. Either it never got applied or it didn't work. I'll try it again.

Turns out it didn't work when he tried it. I fixed the problem, so newly generated results should be OK. I'll try to cancel existing results.

---

Sorry, Jeff was supposed to have fixed those analysis configs last week. Either it never got applied or it didn't work. I'll try it again.

Turns out it didn't work when he tried it. I fixed the problem, so newly generated results should be OK. I'll try to cancel existing results.

It might be a good idea to check the Main project splitters too - I've just reported a couple of very strange WUs.

WU 1669833986
WU 1669840794

I noticed an unusual deadline, and very short processing time. CUDA applications are processing in v6 mode, without autocorr processing. OpenCL doesn't report any autocorr signal count.

Analysis_cfg:

  <analysis_cfg>
    <spike_thresh>24</spike_thresh>
    <spikes_per_spectrum>1</spikes_per_spectrum>
    <autocorr_thresh>0</autocorr_thresh>
    <autocorr_per_spectrum>0</autocorr_per_spectrum>
    <autocorr_fftlen>0</autocorr_fftlen>
    <gauss_null_chi_sq_thresh>2.47139668</gauss_null_chi_sq_thresh>
    <gauss_chi_sq_thresh>1.41999996</gauss_chi_sq_thresh>
    <gauss_power_thresh>3</gauss_power_thresh>
    <gauss_peak_power_thresh>3.20000005</gauss_peak_power_thresh>
    <gauss_pot_length>64</gauss_pot_length>
    <pulse_thresh>24.1975803</pulse_thresh>
    <pulse_display_thresh>0.5</pulse_display_thresh>
    <pulse_max>40960</pulse_max>
    <pulse_min>16</pulse_min>
    <pulse_fft_max>8192</pulse_fft_max>
    <pulse_pot_length>256</pulse_pot_length>
    <triplet_thresh>13.1977901</triplet_thresh>
    <triplet_max>131072</triplet_max>
    <triplet_min>16</triplet_min>
    <triplet_pot_length>256</triplet_pot_length>
    <pot_overlap_factor>0.5</pot_overlap_factor>
    <pot_t_offset>1</pot_t_offset>
    <pot_min_slew>0.00209999993</pot_min_slew>
    <pot_max_slew>0.0104999999</pot_max_slew>
    <chirp_resolution>0.333</chirp_resolution>
    <analysis_fft_lengths>262136</analysis_fft_lengths>
    <bsmooth_boxcar_length>8192</bsmooth_boxcar_length>
    <bsmooth_chunk_size>32768</bsmooth_chunk_size>
    <chirps>
    <chirp_parameter_t>
      <chirp_limit>20</chirp_limit>
      <fft_len_flags>262136</fft_len_flags>
    </chirp_parameter_t>
    <chirp_parameter_t>
      <chirp_limit>50</chirp_limit>
      <fft_len_flags>65528</fft_len_flags>
    </chirp_parameter_t>
  </chirps>
  <pulse_beams>1</pulse_beams>
  <max_signals>30</max_signals>
  <max_spikes>8</max_spikes>
  <max_autocorr>0</max_autocorr>
  <max_gaussians>0</max_gaussians>
  <max_pulses>0</max_pulses>
  <max_triplets>0</max_triplets>
  <keyuniq>-5170219</keyuniq>
  <credit_rate>2.8499999</credit_rate>
</analysis_cfg>

LoL, that's delightfuly retro! Chirp limits which date back before setiathome_enhanced, etc.

The reason the results are being marked invalid is because the SaH v7 validator insists that a v7 result which did not overflow must have a best_autocorr signal. That feature was implemented to keep anonymous platform users from processing v7 tasks with v6 applications.

If there were a good reason to run tasks without autocorrelation, it might make sense to remove that validator feature. The protection is probably no longer needed.

                                                                   Joe

The original reports - see Strange, utterly strange MB v7 run as MB v6 with MB v7 app.... all seemed to come from a single tape/channel:

19ap11ad.19368.xxxxx.140733193388035.12.xxx

but today more channels (still from the same tape) are reaching the front of the queue.

19ap11ad.18067.xxxxx.140733193388039.12.xxx
19ap11ad.22063.xxxxx.140733193388041.12.xxx
19ap11ad.23016.xxxxx.140733193388045.12.xxx
19ap11ad.23606.xxxxx.140733193388046.12.xxx
19ap11ad.24546.xxxxx.140733193388042.12.xxx
19ap11ad.25864.xxxxx.140733193388043.12.xxx

Not all have been processed and returned, but many (not all) have the same oddly short deadlines - just slightly over the current 'shorty' timespan. I've also had a couple of overflows validate, as we might expect from the known validator checks.

Eric,

Any progress in getting the splitters fixed for this issue?

Still getting freshly split work that errors on the GPU apps, and goes inconclusive, then invalid on the CPU apps:

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6887134

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6887134

although some Wu's are O.K:

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6887091

Claggy

I checked recent splitter runs to see how many were bad.

Group                AR          keyuniq  autocorr  triplet_thresh

27my12ac.28863.2120  0.0127258   -117761  bad       10.97682

27my12ac.11211.2529  0.0115601  -1325057  good       9.73841
27my12ac.11211.2938  0.0109193  -1328129  good      10.97682
27my12ac.11211.3347  0.0134823  -1747969  good      12.21523
27my12ac.11211.3756  0.0124419  -1758209  good      13.45364
27my12ac.11211.4165  0.0109803  -1925121  good      14.692049928753
27my12ac.11211.4574  0.0101923        -1  bad        9.73841

27my12ac.5321.4983   0.0133141  -1325057  good       9.73841
27my12ac.5321.5392   0.0143072  -1328129  good      10.97682
27my12ac.5321.5801   0.0139497  -1747969  good      12.21523                     
27my12ac.5321.6619   0.0143122  -1925121  good      14.69205
27my12ac.5321.7028   0.0142447  -1926145  good      15.930459928753
                     
27my12ac.4033.7437   0.0125166  -1325057  good       9.73841
27my12ac.4033.7846   0.0110970  -1328129  good      10.97682
27my12ac.4033.8255   0.0132942  -1747969  good      12.21523
27my12ac.4033.8664   0.0157456  -1758210  good      13.453639928753
27my12ac.4033.9073   0.1029475       -10  bad       13.3834276
                     
27my12ac.2218.9482   0.0143417  -1325057  good       9.73841
27my12ac.2218.9891   0.0150630  -1328130  good      10.976819928753
27my12ac.2218.10300  0.0128078        -1  bad        9.73841
27my12ac.2218.10709  0.0109577   -117761  bad       10.97682
27my12ac.2218.11118  0.0129995   -118785  bad       12.21523

Looks like something approaching 1/3 have the bad autocorr parameters. But they're all VLAR WUS, and if we include those with triplet_thresh differing from the expected 9.73841, only 4 of those 22 groups had proper analysis_cfg.

At main, all recent SaH v7 WUs I've checked have had keyuniq values between -9417729 and -9418728. They're all magnitude 9197*1024 + 100*AR.

                                                                  Joe

Some of these units are still floating around. I just aborted 7 of them (fortunately most of them were cancelled because they had been sent out too many times, but some of them are still out there).

Its the first 7 tasks from this list:

http://setiweb.ssl.berkeley.edu/beta/results.php?userid=17981&offset=0&show_names=0&state=6&appid=

Tom