PRRTE v4.1.0rc1

This is the first release candidate for the second release in the v4 family.

---

IMPORTANT
This series is based on a forking of the PRRTE master branch. The v4.1.0 release contains quite a few bug fixes, plus the following significant changes:

• minimum PMIx required for build and execution is now v6.1.0
• minimum Python required to build from Git clone is v3.6
• MPMD jobs can now specify mapping directives for each app
• Group construct now returns all required info as per Standard
• MCA parameter support has been extended to include the display
  and runtime-options and cpus-per-rank cmd line directives
• fixed debugger daemon rendezvous on remote nodes and debugger
  connection to launcher spawned by the debugger itself

---

Commits since branch was forked:

• PR #2397: Update NEWS for rc1
• PR #2396: Remove the "compat" macro definitions
• PR #2394: Purge checks for earlier PMIx versions
• PR #2392: Require cherry-picks on this release branch
• PR #2391: Update NEWS and VERSION

---

Commits from master branch since v4.0.0 include:

• PR #2389: Initialize variable
• PR #2388: Correct cflags used for check_compiler_version.m4
• PR #2387: Initialize variable before use
• PR #2386: Cleanup prte_info support
• PR #2383: Pass group info in PMIx server callback
• PR #2382: Support backward compatibility with PMIx v6.0.0
• PR #2381: Stop PMIx progress thread during abnormal shutdown
• PR #2380: Treat a NULL name as indicating "stop all threads"
• PR #2379: Use session_dir_finalize to preserve output files
• PR #2378: Stop PMIx progress thread at beginning of finalize
• PR #2376: Update the directive list in check_multi
• PR #2374: Provide an MCA param to control hwloc shmem sharing
• PR #2372: update-my-copyright.py: properly support git workspaces
• PR #2371: Update handling of log requests
• PR #2370: Fix typo and add MCA param for default output options
• PR #2369: Reduce min PMIx version t v6.0.1
• PR #2368: Silence some Coverity warnings
• PR #2367: Add MCA param to set personality
• PR #2365: Initialize ESS before prterun fallback to prun_common()
• PR #2363: Create a custom signature if asymmetric topology found
• PR #2361: Cleanup bind output when partial allocations exist
• PR #2360: Revert back to explicitly setting hwloc support for xml import
• PR #2357: Update prun to new cmd line option spellings
• PR #2356: Enable MCA param support for display and runtime-options
• PR #2353: Don't force map display for donotlaunch if bind display requested
• PR #2351: Fix hetero node launch
• PR #2350: Potential use after free (alerts 10-12)
• PR #2349: Workflow does not contain permissions
• PR #2348: Correct computation of nprocs for ppr
• PR #2347: Support multi-app map-by specifications
• PR #2346: Fix debugger daemon rendezvous on remote nodes
• PR #2344: Multiple commits
  • Fix ppr mapper
  • Extend existing build check CI
• PR #2340: Slight mod to the indirect debugger example
• PR #2338: Correct the nprocs check to support moving to next node
• PR #2337: Perform sanity check on cmd line
• PR #2335: Allow setting default cpus/rank
• PR #2333: Retrieve PMIX_PARENT_ID with wildcard rank
• PR #2331: Remove stale references to LIKWID mapper
• PR #2329: Centralize quickmatch to ensure consistency
• PR #2327: Add the rmaps/lsf component
• PR #2326: Update ras to allow rank/seq files to define allocation
• PR #2325: Remove LSF references in rank_file mapper
• PR #2324: Some cleanup on hostfile parsing
• PR #2323: Do not override user-specified num procs
• PR #2322: Correctly set the number of procs for rankfile apps
• PR #2319: Fix prte_info to output correct version
• PR #2317: Fix printing of binding ranges
• PR #2316: Correct ordering of macro variables
• PR #2315: Provide skeleton for passing allocation requests through RAS
• PR #2313: Implement support for resource usage monitoring
• PR #2311: PLM: mark uncached daemons as reported
• PR #2308: Fix prun tool
• PR #2307: Cleanup and improve autohandling of hetero nodes
• PR #2306: Tweak the forwarding of signals
• PR #2305: Improve hetero node detection a bit
• PR #2304: Correct show-help message
• PR #2303: Allow PMIx group construct caller to specify the order of the final membership
• PR #2301: Fix map-by pe-list when using core CPUs
• PR #2300: Add launching-apps section to docs
• PR #2299: Extend timeout to child jobs
• PR #2298: Replace sprintf with snprintf
• PR #2297: Fix relative node processing
• PR #2296: Let seq and rankfile mappers compute their own num-procs
• PR #2295: Error out when asymmetric topologies cannot support ppr requests
• PR #2294: Fix printout of binding cpus
• PR #2293: Do not assign DVM's bookmark to the application job
• PR #2291: Add some minor verbose debug output
• PR #2290: Check only for existence of PMIx capability flag
• PR #2289: Fix the definition checks for tool_connected2 and log2 integration
• PR #2288: Update with new log2 and tool_connected2 upcalls
• PR #2287: Clarify help messages
• PR #2281: Bugfix: inconsistently setting PMIX_JOB_RECOVERABLE
• PR #2280: Fix display of physical vs logical CPUs
• PR #2279: Fix precedence ordering on envar operations
• PR #2278: Fix the colocation algorithm
• PR #2277: Silence Coverity warnings
• PR #2276: Silence more Coverity warnings
• PR #2275: Silence even more Coverity warnings
• PR #2274: Silence more Coverity warnings
• PR #2273: Extend testbuild launchers support
• PR #2272: Silence yet more Coverity warnings
• PR #2271: Silence more Coverity warnings
• PR #2270: Silence more Coverity warnings
• PR #2268: Extend inheritance to app level
• PR #2266: Inherit env directives if requested
• PR #2265: Move prte function into librar
• PR #2264: Silence more Coverity warnings
• PR #2262: Silence Coverity warnings
• PR #2261: Ensure we have HNP node aliases
• PR #2260: Extend control over client connections
• PR #2257: Resolve Coverity issues
• PR #2252: Extend support for specifying tool connection parameters
• PR #2251: Cleanup queries and completely register tools
• PR #2250: Correct handling of tool-based spawn requests
• PR #2249: Tool registration updates
• PR #2248: Include node object when registering tool
• PR #2246: Properly implement the "abort" operation
• PR #2243: Adjust top session dir name
• PR #2241: Add some finer-grained connection support
• PR #2240: Customize the OMPI "allow-run-as-root" doc snippet
• PR #2239: Runtime check both min/max PMIx versions
• PR #2238: Runtime check that PMIx meets min requirement
• PR #2237: Check for PMIx version too high
• PR #2234: Add support for client_connected2 server module upcall
• PR #2233: Declare the process set during registration
• PR #2231: Provide error message when ssh fails
• PR #2230: Add some missing command strings for debug output
• PR #2228: Minor cleanups in tool connection
• PR #2227: Process deprecated "stop" CLI
• PR #2226: Update CI
• PR #2225: Ensure to progress job launch for singleton
• PR #2224: Properly handle sigterm when started by singleton
• PR #2221: iof/hnp: correctly handle short write to stdin
• PR #2217: Update OAC submodule
• PR #2216: Update child job's fwd environment flag
• PR #2213: Remove debug output
• PR #2211: Check for pthread_np.h header
• PR #2209: src/docs/prrte-rst-content: Add missing file to Makefile.am
• PR #2208: Preserve formatting in show-help output
• PR #2207: Support fwd-environment directives for spawned child jobs
• PR #2206: Preserve source ID across API call
• PR #2205: Reduce min Python version to 3.6

SHASUMS

4786696cba20d0adfb8c54aea2e4f0f626e203e8  prrte-4.1.0rc1.tar.bz2
5caf504b8307ac13288bedee1b83a47350e2954d  prrte-4.1.0rc1.tar.gz
2a36bfe104bc909f051d482760a0dedc03f3ec09  prrte-4.1.0rc1-1.src.rpm

PRRTE v3.0.13

NOTE
The v3.0 release series may be going dormant after this release.

---

IMPORTANT
This release includes the following critical changes:

• repair of the indirect debugger connection procedure. The prterun launcher was using the incorrect proc name to retrieve its parent's processID
• repair of the LSF mapper when LSF affinity has been specified
• a number of fixes to various mapper components and printing of binding ranges
• an improved mechanism for automatically detecting the presence of heterogeneous nodes
• providing MCA param equivalents to the five major (map, rank, bind, display, runtime) cmd line options

---

Detailed changes include:

• PR #2385: Final update of NEWS
• PR #2384: Update NEWS and VERSION for release
• PR #2377: Multiple commits
  • Provide an MCA param to control hwloc shmem sharing
  • Update the directive list in check_multi
• PR #2362: Multiple commits
  • Revert back to explicitly setting hwloc support for xml import
  • Cleanup bind output when partial allocations exist
• PR #2359: Update NEWS
• PR #2358: Multiple commits
  • Enable MCA param support for display and runtime-options
  • Update prun to new cmd line option spellings
• PR #2355: Update NEWS and VERSION to rc3
• PR #2354: Don't force map display for donotlaunch if bind display requested
• PR #2352: Multiple commits
  • Extend existing build check CI
  • Fix debugger daemon rendezvous on remote nodes
  • Workflow does not contain permissions
  • Potential use after free (alerts 10-12)
  • Fix hetero launch
• PR #2345: Fix ppr mapper
• PR #2343: Update NEWS with important note
• PR #2342: Update VERSION and NEWS
• PR #2341: Slight mod to the indirect debugger example
• PR #2339: Correct the nprocs check to support moving to next node
• PR #2336: Allow setting default cpus/rank
• PR #2334: Retrieve PMIX_PARENT_ID with wildcard rank
• PR #2330: Centralize quickmatch to ensure consistency
• PR #2328: Multiple commits
  • Correctly set the number of procs for rankfile apps
  • Do not override user-specified num procs
  • Some cleanup on hostfile parsing
  • Remove LSF references in rank_file mapper
  • Update ras to allow rank/seq files to define allocation
  • Add the rmaps/lsf component
• PR #2321: Fix typo in examples Makefile
• PR #2320: Fix prte_info to output correct version
• PR #2318: Multiple commits
  • Correct ordering of macro variables
  • Fix printing of binding ranges
• PR #2314: Roll rc1
• PR #2312: Mark uncached daemons as reported
• PR #2309: Multiple commits
  • Inherit env directives if requested
  • Extend inheritance to app level
  • Extend testbuild launchers support
  • Fix the colocation algorithm
  • Fix precedence ordering on envar operations
  • Bugfix: inconsistently setting PMIX_JOB_RECOVERABLE
  • Clarify help messages
  • Do not assign DVM's bookmark to the application job
  • Error out when asymmetric topologies cannot support ppr requests
  • Let seq and rankfile mappers compute their own num-procs
  • Fix relative node processing
  • Replace sprintf with snprintf
  • Add launching-apps section to docs
  • Improve hetero node detection a bit
  • Tweak the forwarding of signals
  • Cleanup and improve autohandling of hetero nodes
  • Fix prun tool

SHASUMS

324fc50313dacd6f55372fb673d9e692c64c045c  prrte-3.0.13.tar.bz2
8bb1f1daf3cc798ca943ca48839f5f64e63e99b7  prrte-3.0.13.tar.gz
95b566d8ae209e14157d472bac3a37f415ef1a66  prrte-3.0.13-1.src.rpm

PRRTE v3.0.12

This is a bug-fix release in the PRRTE v3.0 series. Detailed changes include:

• PR #2269: Final update of NEWS and VERSION for release
• PR #2263: Multiple commits
  • Extend control over client connections
  • Ensure we have HNP node aliases
• PR #2255: Update NEWS
• PR #2254: Replace some incorrectly removed code
• PR #2253: Multiple commits
  • Update NEWS
  • Extend support for specifying tool connection parameters
  • Cleanup queries and completely register tools
  • Correct handling of tool-based spawn requests
  • Change no-foreign-tools default to "false"
  • Include node object when registering tool
  • Properly implement the "abort" operation
• PR #2245: Multiple commits
  • Adjust top session dir name
  • Update NEWS
• PR #2244: Multiple commits
  • Declare the process set during registration
  • Add support for client_connected2 server module upcall
  • Customize the OMPI "allow-run-as-root" doc snippet
  • Add some finer-grained connection support
• PR #2236: Protect against running with PMIx versions too high
• PR #2235: Check for PMIx version too high
• PR #2232: Multiple commits
  • Properly handle sigterm when started by singleton
  • Ensure to progress job launch for singleton
  • Update CI
  • Process deprecated "stop" CLI
  • Minor cleanups in tool connection
  • Add some missing command strings for debug output
  • Provide error message when ssh fails
• PR #2222: Multiple commits
  • iof/hnp: correctly handle short write to stdin
  • check for pthread_np.h header
• PR #2219: Update OAC submodule
• PR #2214: Multiple commits
  • Preserve source ID across API call
  • src/docs/prrte-rst-content: Add missing file to Makefile.am
  • Add external OMPI build CI
  • Remove debug output

SHASUMS

21d31cfe32d0d61364586948f766510410605228  prrte-3.0.12.tar.bz2
13d0b8a48a04fab48f0302c31839fc423949b770  prrte-3.0.12.tar.gz
823d81ed920c9a06c36530b8caf827c8ef248b5a  prrte-3.0.12-1.src.rpm

PRRTE v4.0.0

This is the first release in the v4 family. The intent for this series is to provide regular "reference tags", effectively serving as milestones for any development that might occur after the project achieved a stable landing zone at the conclusion of the v3 series. It is expected, therefore, that releases shall be infrequent and rare occurrences, primarily driven by the completion of some significant feature or some particularly critical bug fix.

For this initial release, that feature is completion of support for the Group family of PMIx APIs. This includes support for all three of the group construction modes, including the new "bootstrap" method.

A few notes:

1. Starting with this release, PRRTE requires Python >= v3.7 to build a Git clone (ie., not a tarball). Certain elements of the code base are constructed at build time, with the construction performed by Python script. The constructed elements are included in release tarballs.

2. PRRTE >= v4.0 is not compatible with PMIx < v6.0 due to internal changes (e.g., show-help messages are now contained in memory instead of on-disk files). Configure will therefore error out if the detected PMIx version does not meet this criterion.

A full list of individual changes will not be provided here, but will commence with the v4.0.1 release.

SHASUMS

bd2dc81f2d97d082f28fd50d42735489d7f964a5  prrte-4.0.0.tar.bz2
bf2d33a6f38aacff3fabbe53bf7e0e637113ef1f  prrte-4.0.0.tar.gz
7a2e7fe757e5b13c43a669a155b027ed483459cc  prrte-4.0.0-1.src.rpm

PRRTE v3.0.11

** NOTE **

This is a minor bug fix release to resolve a problem when running with schedulers that allocate at the individual CPU (vs the node) level. Users operating in such an environment need to add the --hetero-nodes option to their prte or prterun cmd lines, or set the prte_hetero_nodes MCA param to a value of 1, t, T, "true", or "True". Administrators of such systems can set the MCA param for their users by adding it to the default MCA param file for the 3.0.11 installation.

---

Changes include:

• PR #2203: Update VERSION and NEWS for release
• PR #2201: Add missing OMPI cmd line option
• PR #2199: Properly handle hetero node configurations

SHASUMS

f1bbdf3d8842bd418d4ddf987161d99d6cdeb514  prrte-3.0.11.tar.bz2
38d4abbff7f6fb916a5bc91c93dc0b409c4af110  prrte-3.0.11.tar.gz
d39edbf9d0ab59a7c11490cc0bc3301c779d1966  prrte-3.0.11-1.src.rpm

PRRTE v3.0.10

Detailed changes include:

• PR #2195: Update NEWS for release
• PR #2194: Multiple commits
  • Remove stale Solaris references
  • docs: Add a "Last updated on:" HTML footer
• PR #2190: Multiple commits
  • Enable shim-based testing of PLM launchers
  • Roll version to 3.0.10
• PR #2186: Fix one more typo
• PR #2184: Fix typo
• PR #2182: Properly terminate RST label
• PR #2181: Add new CI test - build OMPI internally
• PR #2178: Don't link to pmixcc if we don't find it
• PR #2176: Multiple commits
  • Extend support for envar operations
  • Do not require PMIx wrapper compiler

SHASUMS

b58108bf91dd7e3e61642332f38994ac519f042a  prrte-3.0.10.tar.bz2
7d04bc3b544efbecf1a02da8ac205185f2561e5e  prrte-3.0.10.tar.gz
a2b22caadbe63a22c669283eda1fabe3555ead1b  prrte-3.0.10-1.src.rpm

PRRTE v3.0.9

** IMPORTANT **
This release contains configure logic to protect it from PMIx versions that have been updated to support the LTO optimizer as that infrastructure conflicts with the PRRTE v3.0 series. PMIx versions v6.0 and above are not supported by the PRRTE v3.0 series.

---

This is the final anticipated bug-fix release in the v3.0 series. Changes include:

• PR #2172: Update NEWS for release
• PR #2171: Multiple commits
  • Remove ALPS support
  • Require pmixcc be available to build PRRTE
  • Fix oob include/exclude matching
  • Constrain PMIx versions
• PR #2163: Protect against stone age HWLOC
• PR #2162: Multiple commits
  • Break thread deadlock
  • data_server: changes to handle first_read better
  • Allow "not supported" to be returned by query upcall
  • Try adding timeout to CI
  • Add PMIx v4.2 build
  • Fix the data server
  • Just ignore any lto settings
  • Accommodate changes in Slurm version string
  • Correctly support the OMPI stream-buffering option
  • Correct the handling of "prefix"
  • Correct syntax errors in RST files
• PR #2157: Update OAC submodule pointer
• PR #2147: Cleanup the pending lookup request when resolved
• PR #2145: Multiple commits
  • Fix inheritance
  • Provide a user error message when srun fails
  • Update NEWS
• PR #2142: Multiple commits
  • gcc compilation failure fixes
  • Update CI workflow
  • Silence warnings exposed by revised CI
• PR #2141: Multiple commits
  • Silence warnings exposed by revised CI
  • Add build against older PMIx CI
  • Add mpi4py CI
• PR #2137: Multiple commits
  • Add spawn_multiple test
  • Minor cleanups
  • Implement query support for PMIx resolve functions
  • Add ability to direct app GPU support
• PR #2125: hurd: fix build
• PR #2123: Refactor ranking algorithms
• PR #2109: Multiple commits
  • Add a test for CI purposes
  • Try a DVM-based action
  • Update users of pmix_os_dirpath_create
  • Remove the group CI as this release branch doesn't support it

SHASUMS

7012853cb1ce3252c17fe0435585aaeaed809285  prrte-3.0.9.tar.bz2
c9a17aae07903e6960bc13655c4ea9bea6e50960  prrte-3.0.9.tar.gz
db8fa9784867ac48d209475733f5044838cb3217  prrte-3.0.9-1.src.rpm

PRRTE v3.0.8

This is the final anticipated bug-fix release in the v3.0 series. Changes include:

• PR #2103: Final update for release
• PR #2102: Remove stale config .m4 files
• PR #2100: Update NEWS and VERSION
• PR #2099: maintain correct ref count of child jobs
• PR #2097: Correctly parse map-by modifiers
• PR #2086: Propertly compute number of procs when not given
• PR #2082: Multiple commits
  • Enclose XML output in a high-level tag
  • Add support for show_help XML output
  • Properly build the nidmap
• PR #2076: Multiple commits
  • Register the singleton's nspace with the PMIx server
  • Fix data type declaration in data server

SHASUM

44e8d4b71284bfc8646e92c435c73003889ba3c9  prrte-3.0.8.tar.bz2
4f3ed67d57030e16ba8e99ad34de5d751dfa6f11  prrte-3.0.8.tar.gz
3500b5e5f200a5af267a73ed6f1cb121109dff5e  prrte-3.0.8-1.src.rpm
daa3e72ac77bc77346b9b177c5103f638294ff7c  prrte-3.0.8-2.src.rpm

PRRTE v3.0.7

This is a bug-fix release in the v3.0 series.

IMPORTANT: This release represents the expected end of the v3.0 series. Any follow-on bug fixes will be committed to the release branch, but are unlikely to generate an official release tarball.

Changes include:

• PR #2067: Update NEWS and VERSION for final release
• PR #2064: Multiple commits
  • Update help message to PRTE MCA parameter handling
  • Fix typo in Makefile
  • Fix rankfile use of relative node indexing
• PR #2060: Update NEWS and VERSION for release
• PR #2059: Protect against old PMIx versions
• PR #2057: Fix support for hetero LSF environments
• PR #2049: Cleanup show-help files
• PR #2047: Implement the PMIX_JOB_CHILD_SEP support
• PR #2044: Multiple commits
  • Add a threaded client example
  • Restore parsing of OMPI param files
• PR #2039: Unlock stdin target
• PR #2037: Pull a couple of fixes from master branch
• PR #2034: Fix deprecation warnings for ppr on socket objects
• PR #2032: Multiple commits
  • avoid warnings (as errors) spots (init vars, etc.)
  • remove unused Level var
  • init installdirs structure (avoids warn-as-error)
  • Pay attention to interface include/exclude params
  • Protect against the envar version of the Slurm custom args param
• PR #2018: fix support for MPIEXEC_TIMEOUT
• PR #2017: Update prte_portable_platform_real.h from upstream gasnet
• PR #2015: Always have show-help available
• PR #2013: Remove unused yaml
• PR #2011: Cleanup show help formatting
• PR #2009: docs: update for Python 3.12
• PR #2007: Fix a segfault when no arguments are provided
• PR #2005: Multiple commits
  • Remove unused function
  • Revise cmd line parsing to handle special case
• PR #2002: Update OAC to latest HEAD
• PR #2000: Use the PMIx functions to check params
• PR #1998: Protect against LTO optimizer
• PR #1996: Read the Docs updates
• PR #1994: Provide a warning of potentially unknown Slurm params

SHASUMS

75d19187a0c1357ad1bafb24d25c2be1db8d96ad  prrte-3.0.7.tar.bz2
855e0ca9cada12e532f531bac83e2aff775c6e52  prrte-3.0.7.tar.gz
77e86b85881cae8fac91226850ec22873533cc99  prrte-3.0.7-1.src.rpm

PRRTE v3.0.6

This is a bug-fix release in the PRRTE v3.0 series. Changes since v3.0.5 include:

• PR #1992: Update NEWS and VERSION for release
• PR #1991: Multiple commits
  • Protect against missing HWLOC object types
  • Minor fixes to allow compile with pre-stone age HWLOC
• PR #1988: Github action: bring back MacOS builds
• PR #1986: Tailored backport of "various fixes for singleton support"
• PR #1985: Multiple commits
  • Don't strip quotes from cmd line entries
  • Update tar format to tar-pax
• PR #1983: Check the runtime version of PMIx
• PR #1979: Multiple commits
  • Cleanup issues surfaced by devel-check
  • Cleanup unused var warning
  • Add CI build with clang
  • Enable devel-check by default in Git repos
  • Stop warning on LSF detection when not requested
  • Silence warning
  • MPI 4.1: add support for memory-alloc-kinds
  • Remove MacOS CI builds
  • Add support for PMIX_MEM_ALLOC_KIND
• PR #1971: Remove stale resilience document
• PR #1968: Try to improve an error message a bit
• PR #1966: Fix typo
• PR #1964: Repair the binding algorithm
• PR #1962: Correct name of help file for --hostfile
• PR #1959: Multiple commits
  • Update the allocation simulator
  • Use of pmix_getline from PMIx started in v4.2.5
• PR #1956: Multiple commits
  • Allow execution without an active plm component
  • Fix the binding algorithm to handle partial disablement

SHASUMS

1d2888e2b762f722721837b7e10dd5a7cbd35332  prrte-3.0.6.tar.bz2
f21b7cd5f0d0207cd20e8f38c61e68bf3f57e246  prrte-3.0.6.tar.gz
5766cfc4c0c7c91264e32568ab173b41780e5c25  prrte-3.0.6-1.src.rpm