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

Threads squash rebase #6

Open
wants to merge 29 commits into
base: main
Choose a base branch
from

Conversation

hppritcha
Copy link
Owner

No description provided.

npe9 and others added 8 commits February 4, 2019 20:20
use the OPAL_CONFIG_POSIX_THREADS macro

Signed-off-by: Howard Pritchard <howardp@lanl.gov>
These can be changed later if need be.

Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
configure.ac Outdated
@@ -931,7 +931,7 @@ OPAL_CHECK_BROKEN_QSORT
#
# Check out what thread support we have
#
OPAL_CONFIG_THREADS
#OPAL_CONFIG_THREADS
Copy link
Owner Author

@hppritcha hppritcha Feb 5, 2019

Choose a reason for hiding this comment

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

we need to set these compiler flags now in the threads framework in the various components, we may need to do something special here. We think we need to rework OPAL_CONFIG_THREADS to be more agnostic about type of threads. TODO: @hppritcha will enhance OPAL_CONFIG_THREADS

return 0;
}

int opal_uses_threads = 0;
Copy link
Owner Author

@hppritcha hppritcha Feb 5, 2019

Choose a reason for hiding this comment

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

Do we really still use this file? Maybe not include this in the final PR?


static int mca_threads_base_register(mca_base_register_flag_t flags)
{
// Do I need to register anything here?
Copy link
Owner Author

Choose a reason for hiding this comment

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

this is where you'd define base MCA parameters for the threads component

fi

if test "$mutex_base_include" = "" ; then
# mutex_base_include="base/mutex_base_null.h"
Copy link
Owner Author

Choose a reason for hiding this comment

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

TODO: @npe9 to predefine mutex_base_ncude as NULL

@@ -131,7 +133,7 @@ static inline int opal_mutex_trylock(opal_mutex_t *m)
#endif
}

static inline void opal_mutex_lock(opal_mutex_t *m)
static inline void opal_mutex_lock(struct opal_mutex_t *m)
Copy link
Owner Author

Choose a reason for hiding this comment

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

check why the struct needed to be added
TODO: @npe9

# status: e.g. active, maintenance, unmaintained
#
owner: SNL
status: unmaintained
Copy link
Owner Author

Choose a reason for hiding this comment

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

active rather than unmaintained
TODO: @npe9

@@ -0,0 +1,62 @@
/*
Copy link
Owner Author

Choose a reason for hiding this comment

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

@npe9 says these null.h files can be blown off.

@@ -0,0 +1,193 @@
/* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil -*- */
Copy link
Owner Author

Choose a reason for hiding this comment

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

TODO: @npe9 need to actually move from opal/threads/thread_usage.h down to the framework

@@ -61,6 +61,10 @@ char *opal_timing_output = NULL;
bool opal_timing_overhead = true;
#endif

#if OPAL_ENABLE_DEBUG
Copy link
Owner Author

Choose a reason for hiding this comment

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

this variable wasn't conditionally included in master - may get some pushback here. leave it for now.

@@ -62,7 +62,8 @@ endif # OPAL_INSTALL_BINARIES
endif # OPAL_WANT_SCRIPT_WRAPPER_COMPILERS

opal_wrapper_SOURCES = opal_wrapper.c
Copy link
Owner Author

Choose a reason for hiding this comment

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

we're fixing this as part of OPAL_CONFIG_THREADS refactor so it will not be in final PR
TODO: @hppritcha will remove as part of this rework.

npe9 and others added 12 commits February 5, 2019 13:25
rework the threads framework configury to move a good chunk
of setup to the top config directory to make sure thread
package specific compile/link args are defined early in
the configure process.

Signed-off-by: Howard Pritchard <howardp@lanl.gov>
remove workaround introduced into the wrappers Makefile.
This can be done due to the configury refactor.

Signed-off-by: Howard Pritchard <howardp@lanl.gov>
shintaro-iwasaki and others added 9 commits February 13, 2019 17:15
- Add double quotations
- Change Argobot to Argobots
If the poll time is too long, MPI hangs.

This quick fix just sets it to 0, but it is not good for the
Pthreads version. Need to find a good way to abstract it.

Note that even 1 (= 1 millisecond) causes disastrous performance
degradation.
…atomic-fix

threads/argobots: cast pointers for atomic functions
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