Re: [OMPI devel] Opal atomics question

Yes, Fujitsu MPI is running on sparcv9-compatible CPU. Though we currently use only stable-series (v1.6, v1.8), they work fine. Takahiro Kawashima, MPI development team, Fujitsu > Nathan, > > Fujitsu MPI is openmpi based and is running on their sparcv9 like proc. > > Cheers, > > Gilles > > On

Re: [OMPI devel] Opal atomics question

And obviously, ia64 is being used > On Mar 26, 2015, at 5:45 PM, Gilles Gouaillardet > wrote: > > Nathan, > > Fujitsu MPI is openmpi based and is running on their sparcv9 like proc. > > Cheers, > > Gilles > > On Friday, March 27, 2015, Nathan Hjelm > wrote: > > As

Re: [OMPI devel] Opal atomics question

Nathan, Fujitsu MPI is openmpi based and is running on their sparcv9 like proc. Cheers, Gilles On Friday, March 27, 2015, Nathan Hjelm wrote: > > As a follow-on. How many of our supported architectures should we > continue to support. The current supported list is: > > alpha > amd64* > arm* >

Re: [OMPI devel] Opal atomics question

Nathan, I run "make check" on those platforms, which I configure w/o any atomics-specific flags. Whatever coverage that yields is what I test. I also test with builds on x86-64 configured for osx-atomics and gcc-sync-atomics. -Paul (who would also test on Alpha if he had access) On Thu, Mar 26

Re: [OMPI devel] Opal atomics question

:+1: for getting rid of the atomic sub. George. On Thu, Mar 26, 2015 at 3:03 PM, Nathan Hjelm wrote: > > Good to know. That leaves alpha as the only one that isn't being > tested. > > BTW, do you know if you are testing the just the inline assembly or are > you also testing the stuff in opal

Re: [OMPI devel] Opal atomics question

Good to know. That leaves alpha as the only one that isn't being tested. BTW, do you know if you are testing the just the inline assembly or are you also testing the stuff in opal/asm? -Nathan On Thu, Mar 26, 2015 at 09:40:06AM -0700, Paul Hargrove wrote: >Nathan, >I test sparcv8+, spar

Re: [OMPI devel] Opal atomics question

Nathan, I test sparcv8+, sparcv9, ia64 and mips in release candidates. That isn't the same as *using* any of those platforms in production. I just mean to say that the implementations are known to pass "make check". -Paul On Thu, Mar 26, 2015 at 8:48 AM, Nathan Hjelm wrote: > > As a follow-on.

Re: [OMPI devel] Opal atomics question

As a follow-on. How many of our supported architectures should we continue to support. The current supported list is: alpha amd64* arm* ia32* ia64 mips osx* powerpc* sparcv9 sync_builtin* * - known to be in-use. Additionally, should we continue to support the atomics in opal/asm? Some of those

Re: [OMPI devel] Opal atomics question

In case anybody cares: In GASNet we have atomics for "add", "sub", "incr", "decr", and "dec-and-test". On some platforms all five are implemented in terms of a single inline-atomic for "add". There are platforms on which one or more of "incr", "decr" and "decr-and-test" have custom implementatio

[OMPI devel] Opal atomics question

I am working on cleaning up the atomics in opal and I noticed something odd. We define opal_atomic_sub_32 and opal_atomic_sub_64 yet only use opal_atomic_sub_32 once: ./opal/runtime/opal_progress.c:val = opal_atomic_sub_32(&num_event_users, 1); This could easily be changed to: val = opal_a