Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[RF] Improving the interface to the batch mode in RooAbsPdf::fitTo and RooAbsPdf::createNLL #9420

Merged
merged 3 commits into from
Dec 12, 2021

Commits on Dec 12, 2021

  1. [RF] Refactor likelihood creation code for new batch mode

    The code that was previously in RooAbsPdf.cxx got moved to a new file
    BatchModeHelpers.cxx. Furthermore, code duplication in retrieving the
    configuration parameters is removed.
    guitargeek committed Dec 12, 2021
    Configuration menu
    Copy the full SHA
    730a10c View commit details
    Browse the repository at this point in the history
  2. [RF] Make old batched evaluation interface available again

    This makes the old batch mode interface available again in
    `RooAbsPdf::fitTo` and `RooAbsPdf::createNLL`. It is not for the user,
    but the developer to easily make performance comparisons. One just has
    to pass the command argument `RooFit::BatchMode("old")`.
    
    The enum class `RooBatchCompute::BatchMode` was also moved to
    `RooFit::BatchModeOption` because it was never used int the batch
    computation library. The `Option` was added to the name to avoid a
    naming collision with the `RooFit::BatchMode` command argument.
    guitargeek committed Dec 12, 2021
    Configuration menu
    Copy the full SHA
    a00377a View commit details
    Browse the repository at this point in the history
  3. [RF] Use EXPECT_NEAR for batch mode comparisons in testLikelihoodSerial

    Before, and exact foating point match was required in the
    LikelihoodSerialTest.BatchedUnbinnedGaussianND unit test. However, the
    results of the new batch mode can be slightly different numerically, so
    we introduce a relative tolerance of 1e-14 when comparing the NLL from
    the new likelihood classes to the batch mode NLL.
    guitargeek committed Dec 12, 2021
    Configuration menu
    Copy the full SHA
    6515820 View commit details
    Browse the repository at this point in the history