Topic/enable atomics #288
Topic/enable atomics #288
Conversation
|
Test PASSed. |
|
Test PASSed. |
|
Test PASSed. |
|
Test PASSed. |
|
@bosilca George -- can you rebase this PR branch such that your unique commits are on the top of the current master? Right now, I can't read this patch to know what is new / part of this PR vs. what is already on the master. (ping me if you want some help doing this rebase) Thanks! |
|
Test PASSed. |
|
What a mess. I finally manage to rebase this is a decent way (I had to use --force-with-lease to do this). Hopefully this was the correct way. |
|
Test PASSed. |
|
Test PASSed. |
|
George, there are still 21 commits added by this PR just to add what looks like a single small change, and some of them look badly broken. Multiple merge commits appear to still be present on this branch. The right thing to do is to avoid ever merging in the first place, which will make your rebases work properly. Since you already have all these merges on this branch, you'll now need to do something a bit more complicated. Ping me on IM if you want help with this. The |
|
Test PASSed. |
as reported by Coverity with CID 1196722
as reported by Coverity with CIDs 1269848, 1269852 and 1269862
as reported by Coverity with CIDs 710628, 1196713 and 1269855
as reported by Coverity with CIDs 1269901 ans 1269902
as reported by Coverity with CID 710627
as reported by Coverity with CIDs 72125, 72126, 1269899 and 1269900
as reported by Coverity with CID 72202
as reported by Coverity with CIDs 1196737 and 1269850
as reported by Coverity with CIDs 72127, 72145, 72146, 72177, 72179, 72186, 731276, 731278, 1269888, 1269890
as reported by Coverity with CIDs 1196785, 1196787 and 1269896
as reported by Coverity with CIDs 71091, 71230, 71231, 72274, 72389, 1196718 and 1196719
This way, if the call to ibv_fork_init() fails, the job will still continue.
of datatypes. Thanks Bogdan Sataric for the bug report.
OMPI *requires* a C99 compiler, so we don't need the AC_C_INLINE or AC_C_RESTRICT tests anymore (because the C99 compiler guarantees to support them). Indeed, in some cases (see #491), AC_C_INLINE gets the wrong answer and defines `inline` to be empty, which screws up some systems (e.g., OS X with clang). As an added bonus, we get to get rid of a bunch of gcc 2.96-specific code (!) in configure.ac.
The oob/ud configure was not honoring the case if the ompi is configured with --with-verbs=no. This fixes that problems. Fixes #522 Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Instead, use a safe environment variable name (that is SCOPE_PUSHed and SCOPE_POPed).
…igure == Short version Do not export special variables into the environment (e.g., LIBS, LDFLAGS, etc.) when invoking subdir configure scripts. This prevents problems described in #471. == More detail Exporing special env variables before invoking a subdir configure script causes problems in some cases. E.g., in #471, when the user configures with `--with-hwloc=/path/to/hwloc`, and that directory is *not* in a default linker search location will cause the libevent subdir configuration to fail. This happens because: 1. We'll pass LIBS="-L/path/to/hwloc/lib -lhwloc" to the libevent configure script 1. Meaning: configure-generated executables will link successfully 1. But unless LD_LIBRARY_PATH (or some other tell-the-linker-where-to-find-things mechanism) includes /path/to/hwloc/lib, the executable can't run. Specifically, the libevent "hey, does the compiler generate proper executables?" check will fail, and configure will abort (because OMPI needs libevent). I checked the history: exporting these vars dates all the way back to LAM/MPI. I can't think of a reason why we need to export these variables -- AC_CONFIG_SUBDIRs doesn't do it; subdir configure scripts should be orthogonal from the upper-layer configure script (and its variables). So let's remove these export statements and see if anything breaks.
…mponent can be used. This doesn't matter on the master, but it does matter on the 1.8 branch as the MTL select logic is different over there.
This commit fixes a typo in mca_btl_vader_progress_endpoints where OPAL_THREAD_LOCK was used when OPAL_THREAD_UNLOCK was intended. Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
…sub-modules to be loaded.
… the user has asked LSF to assign bindings. Fix a couple of typos in lex parser definitions. Tell hostfile parser to ignore binding designations in hostfiles. Add an attribute to indicate that cpusets were provided as physical cpu ids. Once validated, a version of this will be backported to the v1.8.4 release.
… the semi-colon vs comma situation in hwloc slot_lists
…n we really should abort.
Also alphebetize the ignored files in test/class.
… the user has asked LSF to assign bindings. Fix a couple of typos in lex parser definitions. Tell hostfile parser to ignore binding designations in hostfiles. Add an attribute to indicate that cpusets were provided as physical cpu ids. Once validated, a version of this will be backported to the v1.8.4 release.
|
Refer to this link for build results (access rights to CI server needed): Build Log Test FAILed. |
|
@bosilca Can you rebase this branch instead of merging? |
|
This is a complete mess. I want to see this PR either rebased to make for a clean commit or else closed. |
|
@bosilca Want to catch me on IM to figure out how to rebase it properly? |
|
Can one of the admins verify this patch? |
|
I think we still want this but its hard to tell which commits are the PR. I see ffff47f which is simple enough. Are there others? |
|
This has been replaced by #633. |
…c-update v1.10 usnic libfabric update
Fix some typos and naming errors
As most compilers gain support for builtin atomics I propose to switch the default Open MPI behavior with regard to atomics. Instead of relying on our own atomics, we should leverage what the compiler, or the OS, is supporting. This patch change the default behavior for builds on OSX or with a compiler supporting gcc-type atomic builtins.
@goodell and @hjelmn please review.