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

Use RTLD_GLOBAL for libgomp #4353

Merged
merged 3 commits into from
May 15, 2023
Merged

Use RTLD_GLOBAL for libgomp #4353

merged 3 commits into from
May 15, 2023

Conversation

spxiwh
Copy link
Contributor

@spxiwh spxiwh commented May 4, 2023

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

@josh-willis josh-willis self-requested a review May 4, 2023 17:14
Copy link
Contributor

@josh-willis josh-willis left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can see where the fftw_openmp checks succeeded, so LGTM

pycbc/scheme.py Outdated
@@ -33,7 +33,7 @@
from .libutils import get_ctypes_library

try:
_libgomp = get_ctypes_library("gomp", ['gomp'])
_libgomp = get_ctypes_library("gomp", ['gomp'], mode=ctypes.RTLD_GLOBAL)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@spxiwh do you think it is worth also trying to move the entire try/except block somewhere inside the CPUScheme class, as I suggested on Slack? In this way we reduce the amount of stuff that needs to happen at import time.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@titodalcanton I've moved this into the enter method now. I think that ctypes will never load a library twice, and so we don't duplicate if we make the call to ctypes multiple times (but its rare that we enter and exit schemes more than once).

spxiwh added 2 commits May 9, 2023 06:18
In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).
@spxiwh spxiwh force-pushed the spxiwh-patch-1 branch from 5d82bfd to 8f5bc5f Compare May 9, 2023 13:19
@titodalcanton
Copy link
Contributor

This looks good to me as well… Merge?

@spxiwh spxiwh merged commit 2aff594 into master May 15, 2023
@spxiwh spxiwh deleted the spxiwh-patch-1 branch May 15, 2023 13:11
GarethCabournDavies pushed a commit to xangma/pycbc that referenced this pull request May 25, 2023
* Use RTLD_GLOBAL for libgomp

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

* Move import of ctypes/gomp into __enter__

* Try this
yi-fan-wang pushed a commit to yi-fan-wang/pycbc that referenced this pull request May 30, 2023
* Use RTLD_GLOBAL for libgomp

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

* Move import of ctypes/gomp into __enter__

* Try this
yi-fan-wang pushed a commit to yi-fan-wang/pycbc that referenced this pull request Jun 1, 2023
* Use RTLD_GLOBAL for libgomp

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

* Move import of ctypes/gomp into __enter__

* Try this
titodalcanton pushed a commit to titodalcanton/pycbc that referenced this pull request Jan 19, 2024
* Use RTLD_GLOBAL for libgomp

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

* Move import of ctypes/gomp into __enter__

* Try this
titodalcanton added a commit that referenced this pull request Feb 6, 2024
* Remove reference to relative_example in docs

* Use RTLD_GLOBAL for libgomp (#4353)

* Use RTLD_GLOBAL for libgomp

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

* Move import of ctypes/gomp into __enter__

* Try this

* Revert "Revert "Allow SNR optimizer to use candidate point in initial array (#4393)""

This reverts commit 7be12f1.

We are now catching up with master, where the bug originally introduced
by #4393 is fixed properly, so here I am undoing the temporary fix.

* SNR optimisation options for pycbc_live (#4432)

* Moving the live optimizer option changes to my own branch

* Completing the snr optimization argument group

* updating pycbc_live

* re-adding bug fix

* removing TODO message

* Bug with d_e options

* Adding optimizer-seed

* fixing the d_e optimizer

* replacing run.sh code

* resolve merge conflict

* fixing run.sh

* cleaning up args_to_string func

* changing comment

* codeclimate fixes

* module docstring

* Update module docstring copyright

Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>

* Add gareth

* removing argv

* argument changing

* removing duplicated arguments

* minor CC points

* remove bug introduced when making CC happier

---------

Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>
Co-authored-by: Thomas Dent <thomas.dent@usc.es>

* Improvements to single-detector trigger fitting code for PyCBC Live (#4486)

* Cleanup

* Cleanup

* Refactor duration bin parsing code and add support for reading from bank

* Minor fix/cleanup to logging

* Update CLI checks for duration bins

* Cleanup

* Ignore inconsistent config when combining

* Fix bug

* Fix typo

Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>

* Comment from Gareth

---------

Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>

* [pycbc live] Don't add snr options to command if they don't exist (#4518)

* Don't run snr optimizer setup if not optimizing snr

* moving the check to a more appropraite place

* setting snr_opt_options to None if not optimizing

* [pycbc live] Allowing the use of psd variation in the ranking statistic for pycbc live (#4533)

* Modifying files to include psd variation in single detector statistic calculation

* ending variation.py with a blank line

* Changing to an increment agnostic solution

* removing change already fixed

* Updating function names and docstrings

* removing ToDos and adding more helpful comments

* Removing unused import

* Codeclimate fixes

* Removing excess logging and whitespace mistakes

* Removing unused objects + codeclimate fixes

* Updating comments and docstrings, removing matchedfilter changes

* Revert "Updating comments and docstrings, removing matchedfilter changes"

This reverts commit 0e6473a.

* Removing matchedfilter changes, updating comments and docstrings

* Move --verbose to the end of the commands

* more comment updates

* Repositioning filter recreation

* Changes to comments and removing whitespace

Co-authored-by: Thomas Dent <thomas.dent@usc.es>

* removing refchecks

* Adding option veification for psd variation

* Apply suggestions from code review

Co-authored-by: Thomas Dent <thomas.dent@usc.es>

* fixing EOL error

* Refactoring the filter creation function

* codeclimate fixes

* undo

* full_filt func

* removing indentation

* code climate

* code climate

* try to quiet codeclimate

* codeclimate doesn't know PEP8

* brackets obviate line continuation

---------

Co-authored-by: Thomas Dent <thomas.dent@usc.es>

* added scaling of initial pop in snr_optimizer (#4561)

* added scaling of initial pop

* init popn in optimize_di & pso func

* added changes in optimize_pso

* usig logging.debug for snr

* Do not set matplotlib's backend in internal modules (#4592)

* Set version to 2.1.4

* Remove reference to single_template_examples in docs

* Remove reference to hierarchical_model in docs

* Live: produce empty trigger fit plot for detectors with no triggers (#4600)

* Live: produce empty trigger fit plot for detectors with no triggers

* allow for below-threshold triggers

* fix thinko in option parsing for defaults (#4615)

* fix thinko in option parsing for defaults

When an option is not given at all getattr on the args object gives None, but we don't want to translate that into "--option-name None" on the command line.

* bugfix 

obviously we needed to define 'key_name' first ..

* Improvements to single fit plots (#4509)

* Improvements to single fit plots

* Apply suggestions from Gareth

Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>

---------

Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>

---------

Co-authored-by: Ian Harry <ian.harry@ligo.org>
Co-authored-by: Arthur Tolley <32394213+ArthurTolley@users.noreply.github.com>
Co-authored-by: Gareth S Cabourn Davies <gareth.cabourndavies@ligo.org>
Co-authored-by: Thomas Dent <thomas.dent@usc.es>
Co-authored-by: Praveen Kumar <86048588+PRAVEEN-mnl@users.noreply.github.com>
acorreia61201 pushed a commit to acorreia61201/pycbc that referenced this pull request Apr 4, 2024
* Use RTLD_GLOBAL for libgomp

In conda-forge/pycbc-feedstock#74 it was suggested to use RTLD_GLOBAL for libgomp. Let's see if this works fine with the test suite (which should answer @josh-willis 's concerns).

* Move import of ctypes/gomp into __enter__

* Try this
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants