241
38%
main: 42%

Ran 17 Jan 2018 10:09PM UTC

Jobs 1

Files 17

Run time 1s

Badge

Embed ▾

pending completion

Build # 241

Build Type

push

travis-ci

Committed by

tskisner

Commit Message

This is a substantial refactor with the goal of improving MPI performance while
still maintaining existing multiprocessing performance.  Changes / features
include:

    - Targets are distributed across MPI processes, but MPI is not required.

    - DESI targets are distributed on read, so they are never all kept in memory
      on a single process.

    - BOSS targets are currently read on rank zero, and then distributed to all
      processes.  This could be changed in the future if needed.

    - Single codepath for MPI and multiprocessing, except for two specific
      locations where multiprocessing wrappers are used in the case where MPI
      is disabled.

    - In the multiprocessing case, target data is packed once into shared memory
      as in the original redrock.  This target data is then unpacked to process-
      local memory in each of the 2 multiprocessing blocks.

Run Details

657 of 1758 relevant lines covered (37.37%)

0.37 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
6	241.6 (PYTHON_VERSION=3.5 SETUP_CMD='test --coverage' CONDA_DEPENDENCIES=$CONDA_ALL_DEPENDENCIES PIP_DEPENDENCIES=$PIP_ALL_DEPENDENCIES)	17 Jan 2018 10:09PM UTC	0	37.37	Travis Job 241.6

Source Files on build 241

Detailed source file information is not available for this build.

desihub / redrock / 241
38%
main: 42%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 241

desihub / redrock / 241 38% main: 42%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 241

desihub / redrock / 241
38%
main: 42%

README BADGES
x