[DM-35068] New ip_diffim algorithm breaks on macOS - Jira

XML

Word

Printable

Details

Type: Story
Status: Done
Resolution: Done
Fix Version/s: None
Component/s: ip_diffim
Labels:
- SciencePipelines

Story Points:
1
Epic Link:
f22a-ap-emergent
Sprint:
AP F22-1 (June)
Team:
Alert Production
Urgent?:
No

Description

The new image differencing algorithm from ~~DM-33001~~ breaks on macOS ARM.

$ pytest --log-level=DEBUG -x tests/test_subtractTask.py

==================================================================================== test session starts ====================================================================================

platform darwin -- Python 3.10.4, pytest-7.1.2, pluggy-1.0.0

rootdir: /Users/timj/work/lsstsw/build/ip_diffim, configfile: setup.cfg

plugins: xdist-2.5.0, remotedata-0.3.3, forked-1.4.0, arraydiff-0.3, asdf-2.10.1, mock-3.7.0, hypothesis-6.46.7, flake8-1.1.1, astropy-0.10.0, session2file-0.1.11, filter-subpackage-0.1.1, openfiles-0.5.0, doctestplus-0.12.0, subtests-0.7.0, astropy-header-0.1.2, cov-3.0.0

collected 16 items

tests/test_subtractTask.py s.F

========================================================================================= FAILURES ==========================================================================================

_____________________________________________________________________ AlardLuptonSubtractTest.test_auto_convolveScience _____________________________________________________________________

self = <test_subtractTask.AlardLuptonSubtractTest testMethod=test_auto_convolveScience>

    def test_auto_convolveScience(self):

        """Test that auto mode gives the same result as convolveScience when

        the science psf is the smaller.

"""

        noiseLevel = 1.

        science, sources = makeTestImage(psfSize=2.0, noiseLevel=noiseLevel, noiseSeed=6)

        scienceOrig = science.clone()

        template, _ = makeTestImage(psfSize=3.0, noiseLevel=noiseLevel, noiseSeed=7,

                                    templateBorderSize=20, doApplyCalibration=True)

        config = subtractImages.AlardLuptonSubtractTask.ConfigClass()

        config.doSubtractBackground = False

        config.mode = "convolveScience"

        task = subtractImages.AlardLuptonSubtractTask(config=config)

>       output = task.run(template, science, sources)

tests/test_subtractTask.py:219:

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

python/lsst/ip/diffim/subtractImages.py:234: in run

    subtractResults = self.runConvolveScience(template, science, kernelSources)

python/lsst/ip/diffim/subtractImages.py:324: in runConvolveScience

    kernelResult = self.makeKernel.run(science, template, sources, preconvolved=False)

python/lsst/ip/diffim/makeKernel.py:128: in run

    spatialSolution, psfMatchingKernel, backgroundModel = self._solve(kernelCellSet, basisList)

../../stack/lsst-scipipe-4.0.0/Darwin/utils/g894b03d5c6+82f6db4df0/python/lsst/utils/timer.py:345: in timeMethod_wrapper

    res = func(self, *args, **keyArgs)

python/lsst/ip/diffim/psfMatch.py:1040: in _solve

    raise e

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

self = <lsst.ip.diffim.makeKernel.MakeKernelTask object at 0x1694392d0>, kernelCellSet = <lsst.afw.math.SpatialCellSet object at 0x16944a8f0>

basisList = [<lsst.afw.math.FixedKernel object at 0x16945dab0>, <lsst.afw.math.FixedKernel object at 0x16945daf0>, <lsst.afw.math....x16945dc30>, <lsst.afw.math.FixedKernel object at 0x16945dbb0>, <lsst.afw.math.FixedKernel object at 0x16945dbf0>, ...]

returnOnExcept = False

    @timeMethod

    def _solve(self, kernelCellSet, basisList, returnOnExcept=False):

        """Solve for the PSF matching kernel

        Parameters

        ----------

        kernelCellSet : `lsst.afw.math.SpatialCellSet`

            a SpatialCellSet to use in determining the matching kernel

             (typically as provided by _buildCellSet)

        basisList : `list` of `lsst.afw.math.kernel.FixedKernel`

            list of Kernels to be used in the decomposition of the spatially varying kernel

            (typically as provided by makeKernelBasisList)

        returnOnExcept : `bool`, optional

            if True then return (None, None) if an error occurs, else raise the exception

        Returns

        -------

        psfMatchingKernel : `lsst.afw.math.LinearCombinationKernel`

            Spatially varying Psf-matching kernel

        backgroundModel : `lsst.afw.math.Function2D`

            Spatially varying background-matching function

        Raises

        ------

        RuntimeError :

            If unable to determine PSF matching kernel and ``returnOnExcept==False``.

"""

        import lsstDebug

        display = lsstDebug.Info(__name__).display

        maxSpatialIterations = self.kConfig.maxSpatialIterations

        nStarPerCell = self.kConfig.nStarPerCell

        usePcaForSpatialKernel = self.kConfig.usePcaForSpatialKernel

        # Visitor for the single kernel fit

        ps = pexConfig.makePropertySet(self.kConfig)

        if self.useRegularization:

            singlekv = diffimLib.BuildSingleKernelVisitorF(basisList, ps, self.hMat)

        else:

            singlekv = diffimLib.BuildSingleKernelVisitorF(basisList, ps)

        # Visitor for the kernel sum rejection

        ksv = diffimLib.KernelSumVisitorF(ps)

        # Main loop

        trace_loggers = [getTraceLogger(self.log.getChild("_solve"), i) for i in range(5)]

        t0 = time.time()

        try:

            totalIterations = 0

            thisIteration = 0

            while (thisIteration < maxSpatialIterations):

                # Make sure there are no uninitialized candidates as active occupants of Cell

                nRejectedSkf = -1

                while (nRejectedSkf != 0):

                    trace_loggers[1].debug("Building single kernels...")

                    kernelCellSet.visitCandidates(singlekv, nStarPerCell)

                    nRejectedSkf = singlekv.getNRejected()

                    trace_loggers[1].debug(

                        "Iteration %d, rejected %d candidates due to initial kernel fit",

                        thisIteration, nRejectedSkf

                # Reject outliers in kernel sum

                ksv.resetKernelSum()

                ksv.setMode(diffimLib.KernelSumVisitorF.AGGREGATE)

                kernelCellSet.visitCandidates(ksv, nStarPerCell)

>               ksv.processKsumDistribution()

E               lsst.pex.exceptions.Exception:

E                 File "src/KernelSumVisitor.cc", line 130, in void lsst::ip::diffim::detail::KernelSumVisitor<float>::processKsumDistribution() [PixelT = float]

E                   Unable to determine kernel sum; 0 candidates {0}

E               lsst::pex::exceptions::Exception: 'Unable to determine kernel sum; 0 candidates'

python/lsst/ip/diffim/psfMatch.py:948: Exception

------------------------------------------------------------------------------------- Captured log call -------------------------------------------------------------------------------------

DEBUG    lsst.psfMatch.background:subtractBackground.py:255 Ignoring mask planes: BAD, EDGE, DETECTED, DETECTED_NEGATIVE, NO_DATA

WARNING  lsst.psfMatch.background:subtractBackground.py:291 Too few points in grid to constrain fit: min(nx, ny) < approxOrder) [min(2, 2) < 6]

WARNING  lsst.psfMatch.background:subtractBackground.py:301 Reducing approxOrder to 1

DEBUG    timer.lsst.psfMatch.selectDetection:imagePsfMatch.py:806 runStartCpuTime=2.384572; runStartUserTime=1.994697; runStartSystemTime=0.389868; runStartMaxResidentSetSize=298434560; runStartMinorPageFaults=29968; runStartMajorPageFaults=20; runStartBlockInputs=0; runStartBlockOutputs=0; runStartVoluntaryContextSwitches=12; runStartInvoluntaryContextSwitches=877

DEBUG    timer.lsst.psfMatch.selectDetection:detection.py:250 detectFootprintsStartCpuTime=2.384666; detectFootprintsStartUserTime=1.994788; detectFootprintsStartSystemTime=0.389877; detectFootprintsStartMaxResidentSetSize=298434560; detectFootprintsStartMinorPageFaults=29969; detectFootprintsStartMajorPageFaults=20; detectFootprintsStartBlockInputs=0; detectFootprintsStartBlockOutputs=0; detectFootprintsStartVoluntaryContextSwitches=12; detectFootprintsStartInvoluntaryContextSwitches=877

DEBUG    lsst.psfMatch.selectDetection:detection.py:808 Detection threshold: VALUE value=1.4620134 (positive)

DEBUG    lsst.psfMatch.selectDetection.tempLocalBackground:subtractBackground.py:255 Ignoring mask planes: BAD, EDGE, DETECTED, DETECTED_NEGATIVE, NO_DATA

DEBUG    lsst.psfMatch.selectDetection:detection.py:808 Detection threshold: VALUE value=1.4879407 (positive)

INFO     lsst.psfMatch.selectDetection:detection.py:586 Detected 20 positive peaks in 14 footprints to 10 sigma

DEBUG    timer.lsst.psfMatch.selectDetection:detection.py:250 detectFootprintsEndCpuTime=2.392042; detectFootprintsEndUserTime=2.00193; detectFootprintsEndSystemTime=0.390109; detectFootprintsEndMaxResidentSetSize=298909696; detectFootprintsEndMinorPageFaults=30018; detectFootprintsEndMajorPageFaults=20; detectFootprintsEndBlockInputs=0; detectFootprintsEndBlockOutputs=0; detectFootprintsEndVoluntaryContextSwitches=12; detectFootprintsEndInvoluntaryContextSwitches=880

DEBUG    timer.lsst.psfMatch.selectDetection:imagePsfMatch.py:806 runEndCpuTime=2.392137; runEndUserTime=2.002021; runEndSystemTime=0.39011399999999996; runEndMaxResidentSetSize=298909696; runEndMinorPageFaults=30018; runEndMajorPageFaults=20; runEndBlockInputs=0; runEndBlockOutputs=0; runEndVoluntaryContextSwitches=12; runEndInvoluntaryContextSwitches=880

DEBUG    timer.lsst.psfMatch.selectMeasurement:imagePsfMatch.py:813 runStartCpuTime=2.392198; runStartUserTime=2.002082; runStartSystemTime=0.390115; runStartMaxResidentSetSize=298909696; runStartMinorPageFaults=30018; runStartMajorPageFaults=20; runStartBlockInputs=0; runStartBlockOutputs=0; runStartVoluntaryContextSwitches=12; runStartInvoluntaryContextSwitches=880

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:169 Mask plane "THISDET": plane 9, bitmask 512 = 0x200

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:169 Mask plane "OTHERDET": plane 10, bitmask 1024 = 0x400

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:370 Measured from image: clipped mean = -5.0545e-05, stdev = 0.999856

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:200 Using noise generator: FixedGaussianNoiseGenerator: mean=-5.0545e-05, std=0.999856

INFO     lsst.psfMatch.selectMeasurement:sfm.py:298 Measuring 14 sources (14 parents, 0 children)

DEBUG    lsst.psfMatch.selectMeasurement.base_SdssCentroid:baseMeasurement.py:371 MeasurementError in base_SdssCentroid.measure on record 1:

  File "src/SdssCentroid.cc", line 385, in std::pair<MaskedImageT, double> lsst::meas::base::(anonymous namespace)::smoothAndBinImage(std::shared_ptr<const afw::detection::Psf>, const int, const int, const MaskedImageT &, int, int, lsst::meas::base::FlagHandler) [MaskedImageT = lsst::afw::image::MaskedImage<float>]

    Object too close to edge {0}

lsst::meas::base::MeasurementError: 'Object too close to edge'

DEBUG    timer.lsst.psfMatch.selectMeasurement:imagePsfMatch.py:813 runEndCpuTime=2.427935; runEndUserTime=2.037612; runEndSystemTime=0.39032; runEndMaxResidentSetSize=299384832; runEndMinorPageFaults=30052; runEndMajorPageFaults=20; runEndBlockInputs=0; runEndBlockOutputs=0; runEndVoluntaryContextSwitches=12; runEndInvoluntaryContextSwitches=880

DEBUG    lsst.psfMatch.background:subtractBackground.py:255 Ignoring mask planes: BAD, EDGE, DETECTED, DETECTED_NEGATIVE, NO_DATA

WARNING  lsst.psfMatch.background:subtractBackground.py:291 Too few points in grid to constrain fit: min(nx, ny) < approxOrder) [min(2, 2) < 6]

WARNING  lsst.psfMatch.background:subtractBackground.py:301 Reducing approxOrder to 1

DEBUG    timer.lsst.psfMatch.selectDetection:imagePsfMatch.py:806 runStartCpuTime=2.454331; runStartUserTime=2.06373; runStartSystemTime=0.390596; runStartMaxResidentSetSize=300924928; runStartMinorPageFaults=30182; runStartMajorPageFaults=20; runStartBlockInputs=0; runStartBlockOutputs=0; runStartVoluntaryContextSwitches=12; runStartInvoluntaryContextSwitches=880

DEBUG    timer.lsst.psfMatch.selectDetection:detection.py:250 detectFootprintsStartCpuTime=2.454405; detectFootprintsStartUserTime=2.0638009999999998; detectFootprintsStartSystemTime=0.390602; detectFootprintsStartMaxResidentSetSize=300941312; detectFootprintsStartMinorPageFaults=30183; detectFootprintsStartMajorPageFaults=20; detectFootprintsStartBlockInputs=0; detectFootprintsStartBlockOutputs=0; detectFootprintsStartVoluntaryContextSwitches=12; detectFootprintsStartInvoluntaryContextSwitches=880

DEBUG    lsst.psfMatch.selectDetection:detection.py:808 Detection threshold: VALUE value=1.0181538 (positive)

DEBUG    lsst.psfMatch.selectDetection.tempLocalBackground:subtractBackground.py:255 Ignoring mask planes: BAD, EDGE, DETECTED, DETECTED_NEGATIVE, NO_DATA

DEBUG    lsst.psfMatch.selectDetection:detection.py:808 Detection threshold: VALUE value=1.1213353 (positive)

INFO     lsst.psfMatch.selectDetection:detection.py:586 Detected 18 positive peaks in 12 footprints to 10 sigma

DEBUG    timer.lsst.psfMatch.selectDetection:detection.py:250 detectFootprintsEndCpuTime=2.462878; detectFootprintsEndUserTime=2.072146; detectFootprintsEndSystemTime=0.39072999999999997; detectFootprintsEndMaxResidentSetSize=301694976; detectFootprintsEndMinorPageFaults=30238; detectFootprintsEndMajorPageFaults=20; detectFootprintsEndBlockInputs=0; detectFootprintsEndBlockOutputs=0; detectFootprintsEndVoluntaryContextSwitches=12; detectFootprintsEndInvoluntaryContextSwitches=880

DEBUG    timer.lsst.psfMatch.selectDetection:imagePsfMatch.py:806 runEndCpuTime=2.462965; runEndUserTime=2.072231; runEndSystemTime=0.39073199999999997; runEndMaxResidentSetSize=301694976; runEndMinorPageFaults=30238; runEndMajorPageFaults=20; runEndBlockInputs=0; runEndBlockOutputs=0; runEndVoluntaryContextSwitches=12; runEndInvoluntaryContextSwitches=880

DEBUG    timer.lsst.psfMatch.selectMeasurement:imagePsfMatch.py:813 runStartCpuTime=2.463024; runStartUserTime=2.07229; runStartSystemTime=0.390733; runStartMaxResidentSetSize=301694976; runStartMinorPageFaults=30238; runStartMajorPageFaults=20; runStartBlockInputs=0; runStartBlockOutputs=0; runStartVoluntaryContextSwitches=12; runStartInvoluntaryContextSwitches=880

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:169 Mask plane "THISDET": plane 9, bitmask 512 = 0x200

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:169 Mask plane "OTHERDET": plane 10, bitmask 1024 = 0x400

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:370 Measured from image: clipped mean = -0.000105243, stdev = 1.00975

DEBUG    lsst.psfMatch.selectMeasurement:noiseReplacer.py:200 Using noise generator: FixedGaussianNoiseGenerator: mean=-0.000105243, std=1.00975

INFO     lsst.psfMatch.selectMeasurement:sfm.py:298 Measuring 12 sources (12 parents, 0 children)

DEBUG    timer.lsst.psfMatch.selectMeasurement:imagePsfMatch.py:813 runEndCpuTime=2.475278; runEndUserTime=2.084471; runEndSystemTime=0.390804; runEndMaxResidentSetSize=301711360; runEndMinorPageFaults=30239; runEndMajorPageFaults=20; runEndBlockInputs=0; runEndBlockOutputs=0; runEndVoluntaryContextSwitches=12; runEndInvoluntaryContextSwitches=880

INFO     lsst.alardLuptonSubtract:subtractImages.py:486 template has 57600 good pixels (100.0%)

DEBUG    lsst.ip.diffim.generateAlardLuptonBasisList:makeKernelBasisList.py:205 Generating matching bases for sigma 3.00 pix -> 2.00 pix

DEBUG    lsst.ip.diffim.generateAlardLuptonBasisList:makeKernelBasisList.py:264 Target psf fwhm is the greater, deconvolution mode

DEBUG    lsst.ip.diffim.generateAlardLuptonBasisList:makeKernelBasisList.py:313 basisSigmaGauss: 1.6,2.5,4.6 basisDegGauss: 3,3,3

INFO     lsst.alardLuptonSubtract.makeKernel:makeKernel.py:329 Target psf fwhm is the greater, deconvolution mode

INFO     lsst.alardLuptonSubtract.makeKernel:diffimTools.py:426 Growing 14 kernel candidate stars by 27 pixels

INFO     lsst.alardLuptonSubtract.makeKernel:diffimTools.py:472 Selected 8 / 14 sources for KernelCandidacy

INFO     lsst.alardLuptonSubtract:subtractImages.py:213 Science PSF size: 4.709595

INFO     lsst.alardLuptonSubtract:subtractImages.py:214 Template PSF size: 7.064388

INFO     lsst.alardLuptonSubtract:subtractImages.py:226 `convolveScience` is set: convolving science image.

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 1 at 12463.579102, 67930.898438

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 2 at 12489.073242, 67934.562500

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 3 at 12396.318359, 67954.250000

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 4 at 12443.049805, 67961.398438

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 5 at 12411.178711, 67992.562500

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 6 at 12466.421875, 68013.226562

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 7 at 12435.289062, 68023.171875

DEBUG    lsst.alardLuptonSubtract.makeKernel:makeKernel.py:378 Candidate 8 at 12388.328125, 68036.453125

DEBUG    lsst.ip.diffim.generateAlardLuptonBasisList:makeKernelBasisList.py:205 Generating matching bases for sigma 2.00 pix -> 3.00 pix

DEBUG    lsst.ip.diffim.generateAlardLuptonBasisList:makeKernelBasisList.py:227 Reference psf fwhm is the greater, normal convolution mode

DEBUG    lsst.ip.diffim.generateAlardLuptonBasisList:makeKernelBasisList.py:313 basisSigmaGauss: 1.1,2.2,4.5 basisDegGauss: 4,2,2

INFO     lsst.alardLuptonSubtract.makeKernel:makeKernel.py:327 Reference psf fwhm is the greater, normal convolution mode

DEBUG    timer.lsst.alardLuptonSubtract.makeKernel:makeKernel.py:128 _solveStartCpuTime=2.530971; _solveStartUserTime=2.139767; _solveStartSystemTime=0.3912; _solveStartMaxResidentSetSize=303185920; _solveStartMinorPageFaults=30335; _solveStartMajorPageFaults=20; _solveStartBlockInputs=0; _solveStartBlockOutputs=0; _solveStartVoluntaryContextSwitches=12; _solveStartInvoluntaryContextSwitches=880

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:936 Building single kernels...

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:939 Iteration 0, rejected 0 candidates due to initial kernel fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:953 Iteration 0, rejected 2 candidates due to kernel sum

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:936 Building single kernels...

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:939 Iteration 0, rejected 0 candidates due to initial kernel fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:953 Iteration 0, rejected 0 candidates due to kernel sum

DEBUG    TRACE2.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:997 Spatial kernel built with 4 candidates

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:1005 Iteration 0, rejected 4 candidates due to spatial kernel fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:1009 0 candidates used in fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:936 Building single kernels...

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:939 Iteration 1, rejected 0 candidates due to initial kernel fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:953 Iteration 1, rejected 0 candidates due to kernel sum

DEBUG    TRACE2.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:997 Spatial kernel built with 2 candidates

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:1005 Iteration 1, rejected 2 candidates due to spatial kernel fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:1009 0 candidates used in fit

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:936 Building single kernels...

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:939 Iteration 2, rejected 0 candidates due to initial kernel fit

ERROR    lsst.alardLuptonSubtract.makeKernel:psfMatch.py:1037 ERROR: Unable to calculate psf matching kernel

DEBUG    TRACE1.lsst.alardLuptonSubtract.makeKernel._solve:psfMatch.py:1039

  File "src/KernelSumVisitor.cc", line 130, in void lsst::ip::diffim::detail::KernelSumVisitor<float>::processKsumDistribution() [PixelT = float]

    Unable to determine kernel sum; 0 candidates {0}

lsst::pex::exceptions::Exception: 'Unable to determine kernel sum; 0 candidates'

DEBUG    timer.lsst.alardLuptonSubtract.makeKernel:makeKernel.py:128 _solveEndCpuTime=2.639458; _solveEndUserTime=2.247661; _solveEndSystemTime=0.391794; _solveEndMaxResidentSetSize=307216384; _solveEndMinorPageFaults=30588; _solveEndMajorPageFaults=20; _solveEndBlockInputs=0; _solveEndBlockOutputs=0; _solveEndVoluntaryContextSwitches=12; _solveEndInvoluntaryContextSwitches=880

Eli Rykoff has determined that this is failing relatively early in determining the background values from https://github.com/lsst/ip_diffim/blob/ece46ecc3387e67bbf2a93f7fd24be2204ac7eb3/src/AssessSpatialKernelVisitor.cc#L103

Attachments

Issue Links

is triggered by

DM-33001 Design refactoring of ImageDifferenceTask

Done

relates to

DM-35081 Background modeling fails for some images on ARM processors

To Do

Activity

Ascending order - Click to sort in descending order

Hide

Permalink

Ian Sullivan added a comment - 02/Jun/22 6:19 PM

Further testing indicated that this was not coming from NaNs in the images, but was instead from the image size used in the tests. With 400x400 pixel test images all the tests pass, but with 200x200 pixel test images the tests pass on Linux and Intel Mac, but fail as reported above on a new ARM Mac.

Show

Ian Sullivan added a comment - 02/Jun/22 6:19 PM Further testing indicated that this was not coming from NaNs in the images, but was instead from the image size used in the tests. With 400x400 pixel test images all the tests pass, but with 200x200 pixel test images the tests pass on Linux and Intel Mac, but fail as reported above on a new ARM Mac.

Hide

Permalink

Ian Sullivan added a comment - 03/Jun/22 12:47 AM

This one-line change should fix the tests that are failing on ARM machines, but does not address the underlying bug. I have filed DM-35081 with details from the Slack thread for investigating and difference in KernelSolution.cc

Jenkins with this change:
https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/36747/pipeline/

Show

Ian Sullivan added a comment - 03/Jun/22 12:47 AM This one-line change should fix the tests that are failing on ARM machines, but does not address the underlying bug. I have filed DM-35081 with details from the Slack thread for investigating and difference in KernelSolution.cc Jenkins with this change: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/36747/pipeline/

Hide

Permalink

Tim Jenness added a comment - 03/Jun/22 3:55 AM

This fix works for me. Thanks.

Show

Tim Jenness added a comment - 03/Jun/22 3:55 AM This fix works for me. Thanks.

People

Assignee:: Ian Sullivan

Reporter:: Tim Jenness

Reviewers:: Tim Jenness

Watchers:: Eli Rykoff, Ian Sullivan, Nate Lust, Tim Jenness

Votes:: 0 Vote for this issue

Watchers:: 4 Start watching this issue

Dates

Created:: 02/Jun/22 5:17 PM

Updated:: 04/Jun/22 12:15 AM

Resolved:: 03/Jun/22 5:44 AM

Jenkins

No builds found.

Data Management

Details

Description

Attachments

Issue Links

Activity

People

Dates

Jenkins