Skip to content

ComputeFstat: allow multi-detector Fa and Fb values to be retained in GPU memory for use in further computation

Liam Dunn requested to merge liam.dunn/lalsuite:fstat_expose_cuda_fafb into master

Description

This change allows the multi-detector Fa and Fb values computed by the ResampCUDA F-statistic implementation to be returned as pointers to GPU memory, allowing downstream CUDA-based code to make use of them without performing GPU<->CPU memory transfers.

API Changes and Justification

This change adds two fields to the FstatResults struct which contain GPU memory pointers for the Fa and Fb vectors. This matches the existing convention used for returning the 2F values, where twoF points to a result stored on the host, and twoF_CUDA points to a result stored on the GPU. It also adds FSTATQ_FAFB_CUDA to the FstatQuantities enum.

Backwards Compatible Changes

  • This change does not modify any class/function/struct/type definitions in a public C header file or any Python class/function definitions
  • This change adds new classes/functions/structs/types to a public C header file or Python module

Backwards Incompatible Changes

  • This change modifies an existing class/function/struct/type definition in a public C header file or Python module
  • This change removes an existing class/function/struct/type from a public C header file or Python module

Review Status

cc @karl-wette and @david-keitel

Merge request reports