uberenv update to spack 1.1.1 #1495

cyrush · 2026-01-09T18:42:05Z

No description provided.

cyrush · 2026-01-09T23:19:34Z

==> Added repo to config with name 'conduit'.
[adding spack package: 'uberenv-conduit@develop%gcc~mpi~hdf5~silo~parmetis~adios~python~doc']
[exe: /home/user/conduit/uberenv_libs/spack/bin/spack -D /home/user/conduit/uberenv_libs/spack_env add 'uberenv-conduit@develop%gcc~mpi~hdf5~silo~parmetis~adios~python~doc']
==> Adding uberenv-conduit@develop %gcc~adios~doc~hdf5~mpi~parmetis~python~silo to environment /home/user/conduit/uberenv_libs/spack_env
[concretizing spack env]
[exe: /home/user/conduit/uberenv_libs/spack/bin/spack -D /home/user/conduit/uberenv_libs/spack_env concretize --fresh ]
==> Compilers have been configured automatically from PATH inspection
==> Error: No such variant {'python', 'doc', 'silo', 'hdf5', 'adios', 'parmetis', 'mpi'} for spec: 'gcc~adios~doc~hdf5~mpi~parmetis~python~silo'
[spack version: 1.1.0 (0c2be44e4ece21eb091ad5de4c97716b7c6d4c87)
]
[exe: /home/user/conduit/uberenv_libs/spack/bin/spack -D /home/user/conduit/uberenv_libs/spack_env spec --fresh --install-status --very-long]
==> Error: No such variant {'mpi', 'python', 'silo', 'adios', 'parmetis', 'hdf5', 'doc'} for spec: 'gcc~adios~doc~hdf5~mpi~parmetis~python~silo'

[exe: /home/user/conduit/uberenv_libs/spack/bin/spack -D /home/user/conduit/uberenv_libs/spack_env -k install --fail-fast --fresh ]
==> Warning: You asked for --insecure. Will NOT check SSL certificates.
==> Error: No such variant {'mpi', 'silo', 'python', 'adios', 'hdf5', 'parmetis', 'doc'} for spec: 'gcc~adios~doc~hdf5~mpi~parmetis~python~silo'

It looks like the spec is getting thrashed some how. It should not be gcc as the package name.

Answer:

New spack needs compiler at the end of a spec.

cyrush · 2026-01-13T16:21:11Z

For hdf5 2.0, the source tarball on HDF5 Group website has a different shasum than the source tarball on github.

https://support.hdfgroup.org/releases/hdf5 tarball:

 curl -O https://support.hdfgroup.org/releases/hdf5/v2_0/v2_0_0/downloads/hdf5-2.0.0.tar.gz 
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 40.0M  100 40.0M    0     0  10.5M      0  0:00:03  0:00:03 --:--:-- 10.5M
$ shasum -a 256 hdf5-2.0.0.tar.gz 
6e45a4213cb11bb5860e1b0a7645688ab55562cc2d55c6ff9bcb0984ed12b22b  hdf5-2.0.0.tar.gz

github tarball:

$ curl -LO https://github.com/HDFGroup/hdf5/releases/download/2.0.0/hdf5-2.0.0.tar.gz
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
100 40.0M  100 40.0M    0     0  9321k      0  0:00:04  0:00:04 --:--:-- 9896k
$ shasum -a 256 hdf5-2.0.0.tar.gz 
f4c2edc5668fb846627182708dbe1e16c60c467e63177a75b0b9f12c19d7efed  hdf5-2.0.0.tar.gz

cyrush · 2026-01-13T17:02:40Z

scripts/build_conduit/build_conduit.sh

+  zfp_cmake_dir=${zfp_install_dir}/lib64/cmake/zfp/
+else
+  zfp_cmake_dir=${zfp_install_dir}/lib/cmake/zfp/
+fi


I think this will resolve #1499

cyrush · 2026-01-21T23:37:55Z

windows build_conduit -- umpire smoke test is failing:

    10/116 Test  #10: t_umpire_smoke .............................Exit code 0xc0000135
  ***Exception:   0.11 sec

I think it is failing to find a dll. Not sure which one is missing yet ...

cyrush · 2026-01-22T17:53:40Z

Updating to new TPLs with spack yields two errors:

In t_python_conduit_node:

64: ======================================================================
64: ERROR: test_set_external_with_buffer (t_python_conduit_node.Test_Conduit_Node.test_set_external_with_buffer)
64: ----------------------------------------------------------------------
64: SystemError: Objects/abstract.c:430: bad argument to internal function
64: 
64: The above exception was the direct cause of the following exception:
64: 
64: Traceback (most recent call last):
64:   File "/home/user/conduit/src/tests/conduit/python/t_python_conduit_node.py", line 699, in test_set_external_with_buffer
64:     n.set_external(s_compact,ra)
64:     ~~~~~~~~~~~~~~^^^^^^^^^^^^^^
64: SystemError: <method 'set_external' of 'Node' objects> returned a result with an exception set

This is using:

Python 3.13.11 (main, Jan 22 2026, 00:08:06) [GCC 11.4.0] on linux

In t_relay_io_silo:

96: [ RUN      ] conduit_relay_io_silo.round_trip_save_option_nameschemes_n_files_n_domains
96: [/home/user/conduit/src/libs/relay/conduit_relay_io_silo.cpp : 8344]
96:  Silo save: Overlink: topo name not provided or not found.
96: [/home/user/conduit/src/libs/relay/conduit_relay_io_silo.cpp : 8352]
96:  Silo save: Overlink: topo name defaulting to topo
96: unknown file: Failure
96: C++ exception with description "
96: file: /home/user/conduit/src/libs/conduit/conduit_node.cpp
96: line: 15028
96: message: 
96: Invalid child index: 0 (number of children: 0)
96: " thrown in the test body.
96: 
96: [  FAILED  ] conduit_relay_io_silo.round_trip_save_option_nameschemes_n_files_n_domains (61 ms)

96: [  FAILED  ] conduit_relay_io_silo.round_trip_save_option_nameschemes_n_files_n_domains
96: 
96:  1 FAILED TEST
1/1 Test #96: t_relay_io_silo ..................***Failed   42.69 sec

…duit into task/2026_01_ci_updates

cyrush · 2026-01-22T18:57:24Z

Silo issue, seems a load is failing (maybe the write is failing as well)

load_mesh number of children: 0
save_mesh number of children: 5
save mesh child idx: 0

Catchpoint 1 (exception thrown), 0x00007f12b13a94a1 in __cxa_throw () from /lib/x86_64-linux-gnu/libstdc++.so.6
(gdb) bt 
#0  0x00007f12b13a94a1 in __cxa_throw () from /lib/x86_64-linux-gnu/libstdc++.so.6
#1  0x0000564ff817a991 in conduit::utils::default_error_handler (msg="Invalid child index: 0 (number of children: 0)", 
    file="/home/user/conduit/src/libs/conduit/conduit_node.cpp", line=15028)
    at /home/user/conduit/src/libs/conduit/conduit_utils.cpp:433
#2  0x0000564ff817aa11 in conduit::utils::handle_error (msg="Invalid child index: 0 (number of children: 0)", 
    file="/home/user/conduit/src/libs/conduit/conduit_node.cpp", line=15028)
    at /home/user/conduit/src/libs/conduit/conduit_utils.cpp:466
#3  0x0000564ff813ff4f in conduit::Node::child (this=0x7ffefd9981b0, idx=0)
    at /home/user/conduit/src/libs/conduit/conduit_node.cpp:15028
#4  0x0000564ff8140395 in conduit::Node::operator[] (this=0x7ffefd9981b0, idx=0)
    at /home/user/conduit/src/libs/conduit/conduit_node.cpp:15102
#5  0x0000564ff8071864 in conduit_relay_io_silo_round_trip_save_option_nameschemes_n_files_n_domains_Test::TestBody (
    this=0x564ff9866040) at /home/user/conduit/src/tests/relay/t_relay_io_silo.cpp:1870

@JustinPrivitera can ideas on what could trigger this?

cyrush · 2026-01-22T19:14:25Z

mrgtree_name                  =    (null)
tv_connectivity               =    0
disjoint_mode                 =    0
topo_dim                      =    not specified
file_ns                       =    "|silo_save_option_nameschemes_n_files_n_domains_yes_spiral.cycle_000000/domain_%06d.silo|n"
block_ns                      =    "|mesh/topo"
block_type                    =    500
repr_block_idx                =    not specified
alt_nodenum_vars              =    NULL
alt_zonenum_vars              =    NULL
empty_cnt                     =    0
empty_list                    =    NULL
meshids                       =    NULL
meshnames                     =    NULL
meshtypes                     =    NULL
dirids                        =    NULL
extents                       =    NULL
zonecounts                    =    NULL
has_external_zones            =    NULL
lgroupings                    =    0
groupings                     =    NULL
groupnames                    =    NULL

cyrush · 2026-01-23T19:36:58Z

windows path style issue on some of the tests on windows:

2026-01-23T18:31:06.5590054Z D:\a\conduit\conduit\src\tests\relay\t_relay_io_silo.cpp(2288): error : Expected equality of these values: [D:\a\conduit\conduit\build\RUN_TESTS.vcxproj]
2026-01-23T18:31:06.5590148Z     meshnames[domid]
2026-01-23T18:31:06.5590512Z       Which is: "overlink_save_option_nameschemes_m_domains_n_files_no_spiral\\domfile1.silo:domain3/MESH"
2026-01-23T18:31:06.5590566Z     meshname
2026-01-23T18:31:06.5590816Z       Which is: "overlink_save_option_nameschemes_m_domains_n_files_no_spiral/domfile1.silo:domain3/MESH"

cyrush · 2026-01-28T00:37:21Z

Newer spack build of parmetis yields different result for the polyhedral test:

left : old (current baseline)
right: new

Input distribution for parmetis poly test for reference

@BradWhitlock -- I think a rebaseline is fine here, what do you think?

…o task/2026_01_ci_updates

cyrush · 2026-01-30T21:34:47Z

with new builds of metis and parmetis, we get a different result for the polyhedra test on macos than on linux.

cyrush · 2026-02-02T21:43:15Z

parmetis polyhedra results are not the same on all platforms.

(They used to be)

Linux metis and parmetis spack spec:

root@ed773f55ad04:/home/user/conduit/uberenv_libs/spack_env# spack spec

[+]      ^[email protected]~gdb~int64~ipo~no_warning~real64+shared build_system=cmake build_type=Release generator=make patches:=4991da9,93a7903,b1225da platform=linux os=ubuntu22.04 target=zen2 %c,[email protected]
[+]      ^[email protected]~gdb~int64~ipo+shared build_system=cmake build_type=Release generator=make patches:=4f89253,50ed208,704b84f platform=linux os=ubuntu22.04 target=zen2 %c,[email protected]

macOS metis and parmetis spack spec:

[harrison37@zeliak spack_env (task/2026_01_ci_updates)]$ spack spec
[+]      ^[email protected]~gdb~int64~ipo~no_warning~real64+shared build_system=cmake build_type=Release generator=make patches:=4991da9,93a7903 platform=darwin os=sequoia target=m1 %c,[email protected]
[+]      ^[email protected]~gdb~int64~ipo+shared build_system=cmake build_type=Release generator=make patches:=4f89253,50ed208,704b84f platform=darwin os=sequoia target=m1 %c,[email protected]

Spack specs are the same.

The result is subject to a seed, which is the same on all platforms.

cyrush · 2026-02-02T21:52:38Z

Old spec for linux build:

[spack version: 0.23.1 (2bfcc69fa870d3c6919be87593f22647981b648a)

2026-02-02T17:46:43.7131194Z  -   zqois2f      ^[email protected]%[email protected]~gdb~int64~ipo~real64+shared build_system=cmake build_type=Release generator=make patches=4991da9,93a7903,b1225da arch=linux-ubuntu20.04-zen2

2026-02-02T17:46:43.7132915Z  -   gyulqlm      ^[email protected]%[email protected]~gdb~int64~ipo+shared build_system=cmake build_type=Release generator=make patches=4f89253,50ed208,704b84f arch=linux-ubuntu20.04-zen2

cyrush · 2026-02-02T21:59:08Z

No leads on spack spec differences or patches being the cause. I think we are subject to a difference in rng.

Going to change the baseline to look for a better balance (desired outcome) instead of an exact baseline match.

…f specific baseline

cyrush · 2026-02-03T17:07:29Z

In the docker container test, mpi tests are failing with:

Invalid rank has value 1 but must be nonnegative and less than 1

I think this is an environment issue, i suspect be docker in the azure env is locked down in some way and we can't oversubscribe processes to run mpi.

blt mpi smoke test passes - they are testing a reduce and would pass if somehow the mpi run has 1 rank instead of the expected 4.

cyrush added 3 commits January 9, 2026 10:36

uberenv update to spack 1.1.0

31b86d5

spack repo with new silo and az ci updates

d92166b

use real container name

09d21e4

cyrush added 6 commits January 9, 2026 15:31

spec order change to reflect new reality

516b118

use versions that are in spack

0d2798a

add local pkg with hdf5 2.0 recipe

ffb6f5e

ver and path fixes

23d91bd

found the ghost - missing backtick

da9575e

add macos and win build_conduit testing

569f795

checksum fix, zfp install layout fix

a513966

cyrush commented Jan 13, 2026

View reviewed changes

cyrush mentioned this pull request Jan 13, 2026

h5zzfp build script fails on redhat systems using lib64 instead of lib paths #1499

Open

cyrush added 7 commits January 16, 2026 10:55

fix bc logic typo

1a2af59

updates for spack 1.1.0

f669cdb

new configs

ed89a81

reflect config cleanup

0442377

Merge branch 'develop' into task/2026_01_ci_updates

048a689

update docker recipe

65a787a

update

bca00cf

cyrush added 2 commits January 21, 2026 15:52

try older spack pkg repo

bcda16c

adj umpire dll search logic

18c41dc

cyrush added 3 commits January 22, 2026 18:35

adj python buffer view use

a36c538

Merge branch 'task/2026_01_ci_updates' of https://github.com/LLNL/con…

b802a05

…duit into task/2026_01_ci_updates

small change

6c89bcb

cyrush added 2 commits January 23, 2026 18:09

restore silo tests but ignore ns test issue

122dbf3

revert silo debugging changes

5ac4b0f

cyrush added 11 commits January 23, 2026 15:04

filter hdf5 linker settings for Threads::Threads cmake target

c064507

equal is not strequal ...

447b408

platform agnostic paths

b0acddd

path fixes for py fort ci test and docker example build

4c5ff53

more fs path fixes

44c7c72

silo ns patching

10b8788

fix for mpi tag limit discovery

5da88c3

update python helper to use python3 cmd

756e62e

typo in spack pkg

c04d890

new configs

540f425

silo changes

916fa9b

cyrush changed the title ~~uberenv update to spack 1.1.0~~ uberenv update to spack 1.1.1 Jan 28, 2026

cyrush added 5 commits January 27, 2026 16:37

Merge branch 'develop' into task/2026_01_ci_updates

281eb9c

baseline update for parmetis poly case

583d0a8

Merge branch 'task/2026_01_ci_updates' of github.com:LLNL/conduit int…

5210891

…o task/2026_01_ci_updates

take a journey back in packages time

269eec9

more travel backwards in time

fa40c47

ver fixes

ba8651f

refactor polyhedral test case to check for improved balance instead o…

eb81a9f

…f specific baseline

cyrush added 2 commits February 3, 2026 09:36

test better mpi smoke test

a3ae4e7

shed some light

96dd7b2

uberenv update to spack 1.1.1 #1495

Are you sure you want to change the base?

uberenv update to spack 1.1.1 #1495

Conversation

cyrush commented Jan 9, 2026

Uh oh!

cyrush commented Jan 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cyrush commented Jan 13, 2026

Uh oh!

cyrush Jan 13, 2026

Choose a reason for hiding this comment

Uh oh!

cyrush commented Jan 21, 2026

Uh oh!

cyrush commented Jan 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cyrush commented Jan 22, 2026

Uh oh!

cyrush commented Jan 22, 2026

Uh oh!

cyrush commented Jan 23, 2026

Uh oh!

cyrush commented Jan 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cyrush commented Jan 30, 2026

Uh oh!

cyrush commented Feb 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cyrush commented Feb 2, 2026

Uh oh!

cyrush commented Feb 2, 2026

Uh oh!

cyrush commented Feb 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

cyrush commented Jan 9, 2026 •

edited

Loading

cyrush commented Jan 22, 2026 •

edited

Loading

cyrush commented Jan 28, 2026 •

edited

Loading

cyrush commented Feb 2, 2026 •

edited

Loading