Parallelize DiffractionFocussing - histogram branch

[Peter Peterson]

Basically copy the code for event processing with more than one group.

[Peter Peterson]

Pedantic hand optimization. Refs #4208.

Changeset: 72f68af7129188b68625f17cd3d319be3f30bd60

[Peter Peterson]

Checkpointing timing diagnostics. Refs #4208.

Changeset: 7b5c8d3cce6577242e5144e3cd4fb2ec6d9e6e05

[Peter Peterson]

Flipped the histogram loop. Refs #4208.

Now debugging the unit tests can begin for real.

Changeset: 5db0a46c55c3815e09e8f02f81d46b218752371a

[Peter Peterson]

Fixed the last unit test without editing the test. Refs #4208.

Changeset: 75891aca795fe74aeba428e8aac0a320ecdb2d31

[Peter Peterson]

Code cleanup. Refs #4208.

Removing old version of the histogram branch and removing lots of debug prints.

Changeset: 4dced69c9af0868e30a3c99323ad9c76a76ed602

[Peter Peterson]

Added openmp pragmas and removed debug print statements. Refs #4208.

Changeset: 4b854c3ff936a79d6df2be923f7b19019ab04e4a

[Peter Peterson]

Trying to fix mac builds. Refs #4208.

Changeset: beb05a9e4608478abde4f339965d1a50daa302fc

[Peter Peterson]

Stupid windows and its signed integers for openmp. Refs #4208.

Changeset: a2c11b31d6878234158d4572168f04ebd7d6bd07

[Peter Peterson]

Making constDataY/constDataE update both. Refs #4208.

Changeset: baca01cfc5bc34762fec44937715b9c355d5280b

[Peter Peterson]

Modified range check to improve code speed. Refs #4208.

Since the events and x-axis are ordered, don't check for the next event having a small x-value than the previous.

Changeset: 9790ab1f98b202d122ba33193ccbd0567ca3a971

[Peter Peterson]

Centralizing workspace index to group map. Refs #4208.

Since it is in both branches, make one version of the code.

Changeset: 7ad2d9d790db03f995f815402a7699815f57ab0b

[Peter Peterson]

Refactored to use centralized version of much of the code. Refs #4208.

The constructs introduced in the event branch that were copied into the histogram version are now centralized. The new central versions are now back into the event branch.

Changeset: c9d4c0dadfd4bad6c2343872f8e50b34a2d80f4c

[Russell Taylor]

DiffractionFocussing performance took a hit from either the commit in comment 9 ([baca01cfc5bc34762fec44937715b9c355d5280b]) or [5cc35f2c38be5ba65a1fea6691a7df9f2e8ee1f9] under #3866.

See ​http://scoop.sns.gov/jenkins/view/Tests/job/mantid_performance_tests/Performance_Tests_Report/AlgorithmsTest.DiffractionFocussing2TestPerformance.test_SNAP_event_six_groups_dontPreserveEvents.htm run 387.

[Martyn Gigg]

Refs #4208. Temporarily disable parallel loop on Mac.

Tests fail with parallelization but there is no immediately obvious cause.

Changeset: 8ee805e43d92ec988cd812d1a7ccbbdf2da32f79

[Janik Zikovsky]

I had similar issues with CompressEvents ( ticket #3976 ) failing in parallel on the mac (only). But I was not able to pinpoint the problem so I just disabled it instead...

[Peter Peterson]

Adding mask check to Instrument. Refs #4208.

Changeset: e7e8986f92856e75ff77363ddff20b9dba055cc4

[Peter Peterson]

Added another method for detector to group map. Refs #4208.

Since vectors are significantly faster than maps, added a function to generate one.

Changeset: 2ae1451a47e800272c287826a21c1d0d2ff07fbf

[Peter Peterson]

DiffractionFocussing serial speedups. Refs #4208.

Moved who calculates whether or not a pixel is masked and replaced a map with a vector.

Changeset: 05392d6910479e1660842ea8779b8f31589ae6a0

[Peter Peterson]

No longer throws exception when detector is not found. Refs #4208.

Changeset: 94e0564c30a14995ed42904b3e4f377c4fa2c59a

[Peter Peterson]

Fixing unit test. Refs #4208.

Changeset: afd75aa75737297974f374343c0d1f35daa2f273

[Peter Peterson]

The last commit finished this.

[Vickie Lynch]

NOMAD histogram data was diffraction focussed in 0.3 sec with my 8 cores being used.

[Stuart Campbell]

This ticket has been transferred to github issue ​5055