vsoch commented on issue kubeflow/trainer#3312.
Thank you, @Raakshass. For a heads up, I will be out and back later this coming week. …
tzanio commented on issue mfem/mfem#5329.
Merged in next for testing……
gardner48 open issue llnl/sundials#935.
[BUG] CMake Error at cmake/tpl/FindPETSC.cmake:65 (find_package):
I am attempting to build SUNDIALS on top of PETSc, where PETSc has already installed Kokkos and Kokkos Kernels. …View Comment
BradWhitlock commented on issue llnl/axom#1866.
The failures on dane are the same as I was seeing in my PR. I’m testing whether PR #1868 fixes them….
markcmiller86 open issue llnl/H5Z-ZFP#165.
We don't document default behavior when filter is selected but no params are specified
We should probably document default behavior for the filter….View Comment
milroy commented on issue flux-framework/flux-sched#1467.
I added some commits to address the overbooking issue @jameshcorbett identified and have fixed it on all platforms but EL8. It appears that EL8 is exhibiting different match-allocate behavior than the other test platforms. …
markcmiller86 commented on issue visit-dav/visit#20936.
Hello @visit-dav/visit-core-team, does anyone have time to review? Feel free to ask questions….
tgamblin commented on issue spack/spack-packages#4829.
Isn’t 1.0 the version of the current compiler wrapper in Spack? I guess if this is the same as the current version it’s fine. Labeling this 1.1 or something would make a clear break and would make it easier to run, e.g., spack find compiler-wrapper@1.1: to find only the ones from github.
…
tgamblin commented on issue spack/packages.spack.io#37.
Resolved by #43…
alecbcs commented on issue llnl/hubcast#301.
Closing in favor of #312 …
JustinPrivitera closed issue llnl/conduit#1606.
`venn` example produces odd `mat_check` field
made using venn “full”…View Comment
cyrush commented on issue llnl/conduit#1607.
TODO: Test cases are WIP, need to add many more: …
markcmiller86 commented on issue llnl/Silo#552.
FYI…use 4.12.1-pre2 release assets. The release assets in 4.12.1-pre1 are bloated due to inclusion of ghpages subdir (for silo.llnl.gov website) which includes all prior .tar.gz releases….
jeremylt commented on issue libxsmm/libxsmm#855.
Is libxsmm moving away from making any more releases? While that would not be my preference, it would be handy to know if that is libxsmm’s plan so I can point our users at a specific commit hash as the dependency version instead of a version number…
jameshcorbett commented on issue flux-framework/flux-sched#1467.
If the overallocation issue I mentioned can be fixed I think this would be a huge step forward for the rabbits. However, there is a further step for the rabbits which I don’t think the PR addresses, which is fine, I just want to make sure that it’s known. …
trws commented on issue flux-framework/flux-sched#1467.
Ok, was that the source of the el8 failure or is that looking like a heisenbug?…
chu11 open issue flux-framework/flux-core#7624.
AddressSanitizer: heap-buffer-overflow on address 0x7f27cbf92030 at pc 0x7f27cf482676 bp 0x7fff0d0140c0 sp 0x7fff0d013880
I’m getting this a lot when running with ASAN on a fedora40 image. It notably occurs with the runpty command….View Comment
grondo commented on issue flux-framework/flux-core#7623.
Nothing obvious for rc3 from a login node in dry-run mode, just attr.get, config.get and module.list as expected (the of course the module.remove calls.)…
garlick commented on issue flux-framework/flux-core#7623.
Was it stuck after the last log entry there? (on “unload builtins”)?…
jameshcorbett closed issue flux-framework/flux-coral2#476.
revert handling of alloc_rabbit property
Problem: the alloc_rabbit property hack to allow rabbits to be used with rv1_nosched was perhaps a bad idea and should be reverted…View Comment
markcmiller86 commented on issue HDFGroup/hdf5#6153.
Just curious but where is it written or decided that for ZFP, rate=3.5, precision=16 is or should be considered an error? I mean, based on what I know about ZFP, it certainly seems like you can ask for one or the other but not both. …
jeremylt open issue CEED/libCEED#1965.
Revert Rust + CUDA CI Disabling
I temporarily disabled Rust + CUDA CI while troubleshooting the job. It should be restored…View Comment
cyrush commented on issue visit-dav/visit-website#21.
Resolves: https://github.com/visit-dav/visit/issues/20954…
cyrush closed issue visit-dav/visit#20954.
Web site wrong link to latest build visit script
From https://visit-dav.github.io/visit-website/index.html, link to latest build visit script…View Comment
jedbrown commented on issue openjournals/joss-reviews#9576.
Sorry to have missed this. I’m pretty underwater at the moment, but would be available late next month. …
hppritcha open issue open-mpi/ompi#11166.
Launch scaling issue: calling PMIx_Fence five times during MPI_Init/Finalize
Seen in HEAD of OMPI main branch …View Comment
v-dobrev commented on issue mfem/mfem#5316.
Merged in next for testing……
garlick commented on issue flux-framework/flux-sched#1481.
Please hold off on reviewing this - I’m redoing a few things after making a first pass on a pool class….
cmoussa1 commented on issue flux-framework/flux-accounting#869.
Thanks for reviewing @jameshcorbett! I’ve addressed your great suggestions above and force-pushed. I’ll set MWP here…
JustinPrivitera open issue visit-dav/visit#20955.
Silo Reader: handle null material names
see representative data here in this PR: https://github.com/llnl/conduit/pull/1595…View Comment
artv3 commented on issue llnl/RAJA#1857.
> @artv3 do we want to target this for our summer release? …
mplegendre commented on issue llnl/Spindle#177.
I addressed the identified issues. …
sam-maloney open issue flux-framework/flux-security#231.
test_device.t may fail if no loop devices in use
On my Rocky9.7 VM I was getting the following failure in test_device.t:…View Comment
trws commented on issue flux-framework/flux-docs#323.
The question really needs to go to legal. Technically all of our docs are licensed under LGPL3 because of the release process for flux as a whole. If they don’t object to re-licensing the docs as CC then I don’t think any of us object either, but it’s lab legal we have to ensure is alright with it. …
becker33 commented on issue spack/spack-packages#3968.
I believe it is #52029…
hppritcha commented on issue spack/spack-packages#4804.
okay in that event use the patch you selected. it should probably be applied to all builds out of the 5.0.x release stream….
dylan-copeland commented on issue mfem/mfem#5054.
@IdoAkkerman since I already approved this PR, should we move forward with merging it? If so, we can request review again from someone else….
vsoch commented on issue yoonho/flux-qrmi#2.
Understood, thank you @ohtanim….
vsoch commented on issue singularityhub/shpc-registry#480.
I do think if we used the docker api proper (instead of falling back to my lazy approach) we might get more fine tuned control of the requests. That direction sounds good to me!…
tzanio commented on issue mfem/mfem#5124.
Merged in next for testing……
BradWhitlock commented on issue llnl/axom#1860.
I’m pursuing an idea I had about reusing the MatsetView types to traverse the mixed field data since their representation changes depending on the structure of the matset….
vsoch commented on issue flux-framework/flux-core#7606.
Excellent - looks good! …
grondo commented on issue flux-framework/flux-core#7606.
Should be fixed now: flux-framework/flux-sched#1480…
chu11 commented on issue flux-framework/flux-core#7594.
just re-pushed, handling many of the corner cases that existed in the initial work. removing WIP …
JustinPrivitera commented on issue visit-dav/visit#20950.
venn should be robust to that case: https://github.com/llnl/conduit/issues/1600…
BradWhitlock commented on issue visit-dav/visit#20950.
I think the empty “background” material is the problem, which came from calling venn() with radius 5. It turns out a more suitable value is something like 0.25. Still, it exposes bug dealing with bad data….
alecbcs commented on issue spack/spack-packages#4550.
@spackbot run pipeline…
becker33 commented on issue spack/spack#52062.
```…
imguoguo closed issue sipeed/picoclaw#2784.
[BUG] README Baidu Search Free Tier update
https://github.com/sipeed/picoclaw#-web-search…View Comment
uvilla commented on issue pecos/tps#325.
A possibility cleaner way is warning suppression in valgrind. …
hppritcha commented on issue open-mpi/ompi#13847.
related to #13879 …
BradWhitlock open issue llnl/conduit#1602.
Flesh out Blueprint protocol documentation for mixed-vars / material-dependent fields
The documentation for making mixed-variables and material-dependent fields is very sparse. It would be helpful to have examples of these things, similar to what was done for the various matset flavors….View Comment
JustinPrivitera commented on issue llnl/conduit#1601.
We could also have a table of contents at the top like we did for the X Ray Image Query: https://visit-sphinx-github-user-manual.readthedocs.io/en/develop/using_visit/Quantitative/XRayImageQuery.html…
vsoch open issue flux-framework/flux-docs#323.
License: consider creative commons for docs
From the Linux Foundation:…View Comment
jedbrown commented on issue OSGeo/gdal#14500.
> As long as you put part of the book in the input prompt. …
jeremylt commented on issue CEED/libCEED#1963.
Checked CI locally - happy to merge if you want me to
…
cyrush commented on issue visit-dav/visit#20946.
I am not sure how robust this is b/c we dont test it….
becker33 commented on issue spack/spack-packages#4652.
@climbfuji can you try to create a more human-readable PR title in the future?…
tgamblin commented on issue spack/spack#52137.
Closing this until we can revisit it as an additional configuration option to control what failures are acceptable. We may also revisit making a 404 on URL fetch acceptable, but this is not critical for CI since GitLab’s reliability has been improved….
alecbcs commented on issue llnl/hubcast-test#5.
@lc-hubcast help…
cyrush open issue llnl/conduit#1597.
ci issues with umpire symbols
```…View Comment
imguoguo commented on issue imguoguo/Comments#38.
> @fktoday…
garlick closed issue flux-framework/flux-sched#1306.
json encoding inception with fluxion resource!
I noticed some unintentional turduckens here while playing with the tracing tools in flux-framework/flux-core#6345. …View Comment
jameshcorbett open issue flux-framework/flux-sched#1479.
Make bulk property-update RPCs send partial success messages
@trws noted in #1463 that if a bulk set-property or remove-property RPC fails,…View Comment
wihobbs open issue flux-framework/flux-sched#1477.
fluxion encodes some json as strings, resulting message is inconvenient to trace
Fluxion has a tendency in some callbacks to encode json as a string, which must then later be decoded (possibly taking up unnecessary cycles). As an example:…View Comment
trws commented on issue flux-framework/flux-sched#1477.
sigh…
trws commented on issue flux-framework/flux-core#7608.
Maybe rhel9 or 10 equivalent for the TOSS connection? Otherwise I’d probably default to whatever the current Ubuntu LTS is….
jameshcorbett commented on issue flux-framework/flux-core#7591.
> > FWIW, after messing around for a while, the best thing I could come up with to remove a hard dependency on ply was to move the ply imports from the top of modules into the bodies of every function or method that uses it. :( …
cmoussa1 commented on issue flux-framework/flux-accounting#858.
OK, so I’ve been messing around with a redesign for the better part of a week now and I think I’ve actually been able to get somewhere significant. The overhaul has been quite a lot of work thinking about a) how to make previously-hard-coded parameters configurable and extensible, and b) how to update an already-existing flux-accounting database to adopt the new design….
cyrush open issue Alpine-DAV/ascent#1719.
unexpected visualstudio update in github actions windows runner
We are using the windows-2025 gh-actions runner. …View Comment
gardner48 open issue llnl/sundials#926.
[BUG]
vsoch commented on issue llnl/cardioid#17.
Thanks! This is more than 3 years old and we finished and published the work. Closing….
samuelkgutierrez commented on issue hpc/quo-vadis#452.
qv_scope_split() is more expressive because it has an npieces parameter that split_at() does not….
garlick commented on issue flux-framework/flux-core#7605.
```…
vsoch commented on issue cosimameyer/awesome-pyladies-creations#112.
Yes of course! Thank you for the showcase!
v-dobrev open issue mfem/mfem#5275.
cudaErrorInvalidDevice abort in GPU builds due to missing device migration in HypreParMatrix::DropSmallEntries
When running a GPU-accelerated build (CUDA + Thrust), calling HypreParMatrix::DropSmallEntries results in a poisoned CUDA context that causes subsequent device operations to fail with a hardware abort. …View Comment
adrienbernede commented on issue llnl/radiuss-shared-ci#76.
In facts, I have mixed feelings: from the components perspective, I like how everything is explicitly stated in the inputs. But in the underlying setup, the perf processing job extends caliper to github, which extends convert to github, which itself has not script because the script should be specified by the user… But again, the components inputs makes the requirements clear….
wihobbs open issue flux-framework/flux-core#7601.
job-manager: plugins/validate-duration.c && update.c -- do they really need the full resource.R?
Some recent testing (see #7598 and flux-sched#1471) show that there are a few places where the resource.R is still fetched from the KVS. When that’s 202Mb, whatever’s doing it slows down a lot. …View Comment
wihobbs open issue flux-framework/flux-coral2#481.
callback error: `job-manager.err[0]: cray-pmi-bootstrap: Failed to post ports to job: Success`
On one of my many rounds of testing yesterday with --setattr=exec.test.run_duration, one round of 100 jobs got polluted with a ton of these messages. Apologies that I didn’t save great notes for what actual jobs caused this, though I can try to look in the persisted content-store. This didn’t show up consistently but I figured an issue would help us track and see if the codepath to cause this should be investigated….View Comment
BradWhitlock open issue mfem/mfem#5321.
MFEM appears to give incorrect quadrature-point coordinates for some custom tensor-product quadrature rules when the number of points differs by logical direction, like 5 x 3 in 2D or 5 x 3 x 2 in 3D.
- If I build a custom tensor-product IntegrationRule from unequal 1D rules and use it with Mesh::GetGeometricFactors(..., COORDINATES) on a curved quad/hex mesh, the physical quadrature-point locations can…View Comment
tomstitt commented on issue mfem/mfem#5309.
While I was trying to make a standalone reproducer I found that hipsparseSpMV_preprocess is not always needed, you can make a basic driver that works without it. I’m not sure yet if it’s the structure of our matrices, something about the function call ordering, or something else. I will update when I know more.
…
milroy commented on issue flux-framework/flux-sched#1466.
> We are maybe trying to get https://github.com/flux-framework/flux-sched/pull/1465 in as well @milroy ? …
tzanio commented on issue mfem/mfem#5230.
This PR is now under review (see the table in the PR description). To help with the review process, please do not force push to the branch….
cyrush commented on issue visit-dav/visit#20944.
Thanks for sharing, this pretty crazy that the compiler / system headers may be incompatible with the system standard libs. …
alecbcs commented on issue spack/spack-packages#4545.
@spackbot run pipeline…
tgamblin commented on issue spack/spack#52205.
I also like deprecated() over deprecated=True, for the cases above. How, though, would you handle the mappings in #52112? I suppose you could still have an optional replaecment:…
hppritcha commented on issue open-mpi/ompi#12458.
Should we keep this PR open?…
v-dobrev commented on issue mfem/mfem#5306.
Re-merged in next for testing……
cyrush commented on issue llnl/conduit#1594.
@Arlie-Capps let me know if these fixes resolve the fmt issues you encountered…
BradWhitlock open issue llnl/axom#1693.
Add support for uniform grid sampling in sampling shaper
The query points in our sampling-based workflow are currently defined by Gauss-Legendre quadrature node locations within each cell….View Comment
chu11 open issue flux-framework/rfc#517.
rfc23: add year suffix?
In https://github.com/flux-framework/flux-core/issues/6132, there is a configuration being developed for preserving history of drain/undrain events….View Comment
grondo closed issue flux-framework/flux-sched#1469.
loading `sched-fluxion-resource` with `load-format=hwloc` causes interner abort
I’m not even sure if this is supported, but I’m trying to use match-policy=locality on a single node and wanted to see if loading the hwloc XML topology would work, but ran into this issue immediately:…View Comment
trws commented on issue flux-framework/flux-sched#1449.
Same, that’s the normal way we do it in testing and everywhere else. All I can think is maybe we have a race between the hello protocol and the free event for the job. I’m trying to work out how to test it reliably, but it’s a tricky thing to pin down….
grondo commented on issue flux-framework/flux-core#7589.
Well, the obvious thing to do is to add a configuration flag to the job-list module that puts it in private mode, effectively adding and user:<requestor> to the constraints used to generate responses. Do you know if private-mode should also secure job-list.job-stats so guest users can’t see the general job stats? (shows counts of all jobs in all states) My guess is yes….
garlick commented on issue flux-framework/flux-core#7500.
As I’m working through a subclassing experiment I’m realizing this doesn’t quite solve the problem. I’ll close this PR for now and open some more once the design is solid….
trws commented on issue flux-framework/flux-coral2#473.
Ok, we could move it to a resolved state on the board or similar I guess. …
cmoussa1 commented on issue flux-framework/flux-accounting#863.
Looks good to me! Thanks bot…
adrienbernede open issue llnl/radiuss-shared-ci#75.
Manually triggering a flux job alone results in using the default queue instead of CI specific one.
Current workaround: manually trigger the shared alloc job first and on completion trigger the desired job….View Comment
JustinPrivitera commented on issue llnl/conduit#1111.
#1593…
markcmiller86 commented on issue llnl/Silo#535.
> Do you have a rough guess on when there will be a tagged release that incorporates this fix?…