tzanio commented on issue mfem/mfem#5190.
Re-merged in next for testing……
vsoch commented on issue flux-framework/flux-core#7341.
The beauty of try/except!…
JustinPrivitera open issue llnl/conduit#1514.
Generalized way to walk material sets
we need a more general way to walk matsets; no more bespoke solutions…View Comment
grondo commented on issue flux-framework/flux-core#7341.
You can also use modprobe to schedule these kinds of tasks before and after the execution of a batch or alloc instance….
chu11 commented on issue flux-framework/flux-core#7303.
re-pushed going with a slightly alternate implementation. …
tzanio closed issue GLVis/glvis#254.
Screenshot does not behaves reliably in glvs script
Following this tutorial, I tried to make a .gif file from a series of solutions. …View Comment
alecbcs commented on issue spack/spack-packages#3144.
@adamjstewart I’ve rebased this PR now that #3119 is merged. I’ll need you to approve before we can proceed since you requested changes….
cyrush commented on issue spack/spack-packages#3200.
I appreciate the proposed solution, but I don’t think we should take such measures in the spack package. …
tgamblin commented on issue spack/spack-packages#3200.
One other thought. We could: …
alecbcs commented on issue spack/spack#51521.
Gotcha. I’ll drop the tag on the PR. Thanks for checking….
becker33 commented on issue spack/spack#51632.
This PR does not apply cleanly to releases/v1.0 and will be removed from v1.0.3…
tgamblin commented on issue spack/spack#51881.
I’m ok with this (enthusiastic, even) if we can add an audit like the one @haampie mentions. I think we need that soon though – otherwise I think we’re going to start getting a lot of packages with invalid constraints….
hppritcha closed issue open-mpi/ompi#13377.
ompio `collective_buffering` MPI_Info hint does not work correctly
The following test program can be used to illustrate the problem:…View Comment
camierjs commented on issue mfem/mfem#5190.
For lower orders, the performances are fine - this issue appears specifically at higher ones….
adrienbernede commented on issue llnl/radiuss-spack-configs#157.
> Ah I see, that makes sense. And yes, I do have access to Corona, but just not the ability to relaunch radiuss-spack-config jobs from gitlab (I do have access for Umpire and maybe RAJA) …
cyrush commented on issue llnl/conduit#1495.
with new builds of metis and parmetis, we get a different result for the polyhedra test on macos than on linux. …
amagela commented on issue llnl/benchmarks#9.
@pearce8 , I haven’t finished this quite yet. …
trws commented on issue flux-framework/flux-sched#1425.
That sounds right. I think as it stands, the R returned by the scheduler is what the shell uses for this, since that provides the actual resources to use, but for correct output wherever the jobspec_parse is used instead we would need either a separate annotation or to consult the R to determine what the result is….
trws commented on issue flux-framework/flux-core#7340.
@grondo mentioned offering a flux python<num> for each of the installed bound pythons, these could have a “get path” option or similar, or maybe we could have arfuments on flux python itself to list or select the available version to use?…
garlick commented on issue flux-framework/flux-core#7333.
Thanks - I added that and will set MWP….
garlick commented on issue flux-framework/flux-coral2#446.
Thanks!…
trws commented on issue desul/desul#164.
That would be pretty painful yeah. Do any of the implementations actually trap on an atomic add to a non-writable location? I haven
camierjs commented on issue GLVis/glvis#353.
Closing for in-place solution in #336…
jeremylt open issue CEED/libCEED#1830.
CeedVectorFilter
@zatkins-dev would it be adequate to your purposes to add this in libCEED? (Though you could access it via PETSc also)…View Comment
trws commented on issue spack/spack-packages#3144.
Darn, thanks for trying it!…
jedbrown commented on issue openjournals/joss-reviews#9244.
Thank you @valsson!…
v-dobrev commented on issue mfem/mfem#5210.
I thought about this issue and I think a better solution is to = delete the move-assignment operator of DenseMatrix. This way assignments to DenseMatrix will always be copy-assignments. The reason I think this is a better solution is that re-defining the move-assignment (especially at the level of Array) to do a deep copy, in the case when the LHS is an alias, goes against the idea of move-assignment. We will lose the ability to move-assign DenseMatrix but that is not a big deal, I think – this was already the case for a long time before we added it in #4028.
…
cyrush open issue visit-dav/visit#20803.
import visit, when given site packges dir `linux-x86_64-gfx942` seems to still look for `linux-x86_64` w/o the gfx942
### Describe the bug…View Comment
gardner48 commented on issue llnl/sundials#836.
#837 deprecates the <TPL>_WORKS options and replaces them with SUNDIALS_ENABLE_<TPL>_CHECKS to make the purpose of the option clearer (and namespace the option). …
bgunnar5 open issue llnl/merlin#588.
Non-blocking runs
User shouldn’t need to wait for the workflow to complete before resuming terminal interaction…View Comment
wihobbs commented on issue flux-framework/flux-security#213.
@chu11 Can you add the MWP label? I don’t have that ability; this repo must not inherit maintainer privileges from the core team (IIRC pmix doesn’t either)….
cmoussa1 commented on issue flux-framework/flux-accounting#810.
Thanks @jameshcorbett! Setting MWP…
cbritopacheco commented on issue cbritopacheco/rodin#165.
@copilot Add manufactured tests for P1_H1 mixed problems similarly as you did with the P0_P1 tests. Put them in the folder called P1_H1 and go up to K = 2….
markcmiller86 open issue llnl/Silo#516.
Add abidiff CI
Add CI to check for ABI diffs in the committed code. Maybe allow them to main but not to any RC branch….View Comment
adrienbernede commented on issue llnl/RAJAPerf#577.
@rhornung67 This is ready for review….
wihobbs commented on issue flux-framework/flux-core#7325.
> Did you spot a way to easily “exempt” paths from a rule in dorny/paths-filter? This action seems to easily enable conditional execution based on files changed, not based on ‘only files changed’,…
emily-howell closed issue Alpine-DAV/ascent#1386.
RZ mesh support
Need to add RZ mesh support to Ascent. A user reported the following: …View Comment
BradWhitlock closed issue llnl/conduit#1485.
Failure with find_reference_node.
I ran across a strange case where find_reference_node fails when I move mesh domain node to a different node. I would expect that since the new node took ownership of the mesh data, it would serve as the root and ought to basically be the same as the copy case….View Comment
cmoussa1 commented on issue flux-framework/flux-core#7303.
I’ll second @jameshcorbett’s vote to just add a raw=False flag…
cjvogl commented on issue E3SM-Project/E3SM#8012.
> @cjvogl can you confirm that with this change E3SM will build on LC machines? …
trws commented on issue spack/spack-packages#3088.
I have not seen that no. I’m not sure what’s causing it either, it’s a bufreadpost from your dotfiles trying to load a lua parser, but I don’t know why that would be failing there. Maybe something is out of order in the load order specifically the way we set it up with spack?…
tzanio commented on issue mfem/mfem#5110.
Merged in next for testing……
dylan-copeland commented on issue mfem/mfem#5203.
What is the policy for output in unit tests? There is std::cout << std::flush; in fem/test_var_order.cpp and other calls to cout if (verbose_tests) in test_quadinterpolator.cpp. When should cout be used with verbose_tests, vs. the more common mfem::out? There is usage of mfem::out if (verbose_tests). Should we add if (verbose_tests) in addition to the change to mfem::out? I suppose unit tests should rely on REQUIRE statements to check conditions, and output should be unnecessary but possible in the case if (verbose_tests). If so, cout could also be changed in test_quadinterpolator.cpp and test_var_order.cpp, along with use of if (verbose_tests)….
gardner48 closed issue llnl/sundials#789.
[BUG] build with -D ENABLE_OPENMP_DEVICE=ON fails
jameshcorbett commented on issue flux-framework/flux-sched#1419.
I think I’ve addressed your comments @milroy ! Thanks for the review….
garlick commented on issue flux-framework/flux-core#7321.
Restarted el8 builder after hitting …
jameshcorbett commented on issue flux-framework/flux-core#7303.
I vote B) and I would recommend not type-checking the input with isinstance but instead just converting it to bytes() (if you even need to do that) and letting any errors propagate…
alecbcs commented on issue spack/spack-packages#3123.
Looks like https://github.com/davidhalter/jedi/issues/2064 is an open issue for Python 3.14 compatibility in Jedi with a few known bugs. …
jeremylt commented on issue openjournals/joss-reviews#8388.
Ratel 1.0 has been released. I appreciate your patience and apologize again for the delay….
balos1 commented on issue llnl/sundials#796.
SteVwonder open issue gnss-sdr/gnss-sdr#972.
CMake fails with Boost 1.90+ because `boost_system` CMake package was removed
### Environment…View Comment
chu11 commented on issue flux-framework/flux-core#7191.
> It makes me lean to the idea of having a “mixed” db as I call it above. The eventlog is stored as it is above, with possibly some special handling of exceptions (maybe other events too). We have DB columns for important fields. Longer out, we can add sqlite plugin extensions for faster / simpler hostlist / rank searches in sqlite for those columns. …
alecbcs commented on issue davidhalter/jedi#2064.
Hi all, just wanted to reach out to see if any additional progress had been made on this front. Thanks!…
tzanio commented on issue GLVis/glvis#335.
@najlkin can you please resolve the conflicts?…
jeremylt commented on issue CEED/libCEED#1694.
MacOS + weak symbols fixed here: https://github.com/CEED/libCEED/pull/1919…
jedbrown commented on issue CEED/libCEED#1919.
Thank you. This looks clean and decoupled given the constraints. I agree it looks like the best option….
cyrush commented on issue visit-dav/visit#20748.
I think this was resovled in #20768…
vsoch commented on issue rootless-containers/usernetes#152.
We tried to use helm first, which is going to deploy the tigera operator, and could not get that setup working….
chu11 commented on issue networkupstools/nut#3193.
Adding -u root worked for me….
cyrush commented on issue llnl/conduit#1495.
windows path style issue on some of the tests on windows: …
mplegendre closed issue llnl/Spindle#128.
`cobo_open_tree()` creating unexpected network tree configurations
Running the first test in runTests on 12 tuolumne nodes resulted in a network layout that looks like:…View Comment
wihobbs commented on issue flux-framework/flux-core#7316.
In the last push: …
grondo commented on issue flux-framework/flux-core#7318.
Thanks! …
homijan commented on issue conservation-laws/ryujin#285.
Dead end here….
chu11 commented on issue chu11/freeipmi-mirror#78.
a few of the issues I posted above don’t seem to be resolved. Did you push the wrong update here?…
cyrush open issue HDFGroup/hdf5#6171.
hdf5 2.0.0 libhdf5.settings lists a cmake target name (Threads::Threads) instead of threading link line
Here is an example:…View Comment
becker33 commented on issue spack/spack#51835.
@FlyingSamson I was finally able to reproduce your bug (accidentally, while trying to do something else). I was also able to backport the fix and include it in the upcoming release. Thanks for bisecting that and making it easy!…
hppritcha commented on issue open-mpi/ompi#12979.
The application uses persistent send/recv requests with the comm pattern being an all-to-one pattern. The individual messages for the problem suggested above are about 40000 bytes in size when using 60 ranks. I’m seeing the hang for fewer than 64 procs. The behavior appears to be that the target rank (0) does receive all its messages but some of the senders do not “think” that all of their sends are complete. Again this appears to be btl-sm specific. turning off the fast boxes does not appear to help. investigation continues……
alecbcs closed issue llnl/hubcast#155.
LDAP account mapper
## Problem/Opportunity Statement…View Comment
JustinPrivitera commented on issue llnl/conduit#1495.
> ``` …
jakubcerveny commented on issue jakubcerveny/gilbert-paper#17.
No need to apologize. Thank you…
imguoguo closed issue imguoguo/NanoKVM-USB#6.
trws commented on issue hpsfoundation/tac#61.
Great news, thank you all!…
vsoch commented on issue hpsfoundation/tac#61.
Hooray!!
benzwick open issue SlicerRt/SlicerRT#301.
Segment Comparison module requires "Closed surface" representation
When trying to compute the metrics using Slicer 5.10.0 I get the following error…View Comment
JustinPrivitera commented on issue visit-dav/visit#20794.
To test I need to check what ends up in the analogous place for 3.5.0: /usr/gapps/visit/3.4.2/linux-x86_64/lib/python/lib/python3.9/site-packages/…
markcmiller86 commented on issue spack/spack-packages#2346.
@tldahlgren, @patrickb314 and @cyrush thanks so much!…
wihobbs commented on issue flux-framework/flux-sched#1372.
Hi @ax3l, in looking back at this issue the EINVAL was actually raised on the job in your subinstance of Flux, not the top-level allocation itself. Here’s a theory on what could have happened:…
cmoussa1 closed issue flux-framework/flux-accounting#788.
database: add `max_sched_jobs` attribute per-association
If the max number of scheduled jobs needs to be enforced per-association, then it should be a property in the flux-accounting database that can be configured with front-end commands such as add-user and edit-user….View Comment
vsoch commented on issue flux-framework/.github#7.
Updated. …
benzwick open issue coendevente/SlicerNNInteractive#83.
Errors when installing server on Windows
We tried to install the server on Windows using the following commands from the README.md file:…View Comment
cbritopacheco commented on issue cbritopacheco/rodin#163.
@copilot I have fixed the tests. Please add more 3D tests inspired from the P1 2D tests using other equations….
emily-howell closed issue Alpine-DAV/ascent#431.
Support for 64bit ids
Until now, we have gotten by without 64bit ids. 32bit worked, since no operations/filter made decisions based on the total number of cells. With filters like statistics, that has changed. There is another branch where this has been tested, e.i., using 64bit ids with vtkm works just fine….View Comment
v-dobrev commented on issue spack/spack-packages#3040.
> @v-dobrev please see #1557 and revert your llvm_amdgpu change. It’s wrong. Even if this was the case, it shouldn’t be mixed with a mfem change. …
jandrej commented on issue mfem/mfem#5192.
> Discussed in the PR meeting: instead of changing all files every year, does it make sense to have a top-level COPYRIGHT file that is updated yearly, and other source files reference that one. This is what Spack does, see https://github.com/spack/spack …
artv3 commented on issue llnl/RAJAPerf#599.
> > @rchen20 , you can also squash merge …
jameshcorbett closed issue flux-framework/flux-coral2#444.
CRD version tolerance
Problem: software that interacts with DWS software uses a fixed API version for kubernetes CRDs. Unfortunately, that forces administrators to update the flux-coral2 RPM and DWS software in sync, and if the versions are not compatible the whole ecosystem breaks. Typically this is not as much of an issue because the DWS software tolerates Flux requesting an older API version, so DWS can be updated without Flux. However, sometimes it might be useful to have Flux be updated without updating DWS, which would mean Flux asking DWS about API versions that it thinks do not exist….View Comment
benzwick commented on issue SlicerCBM/SlicerCBM#77.
## Current ruff check results (as of 2026-01-20)…
brendankeith commented on issue mfem/web#305.
This is now ready for review!…
tzanio commented on issue mfem/mfem#5198.
is this ready-for-review?…
chu11 open issue flux-framework/flux-core#7278.
no way to easily write raw value to KVS with python bindings?
There doesn’t seem to be a way to write a “raw” value to the kvs with the current KVS bindings, or I couldn’t figure out how to easily do it. I had to write this function:…View Comment
garlick commented on issue flux-framework/flux-core#7292.
Oh, hey, didn’t we have a problem a while back where the flux systemd instance was getting killed when someone logged in and out of a node? I thought we had a workaround in place for that, but don’t recall the details. Maybe the 90s default stop timeout is in effect for the system unit that contains the user instance?…
grondo commented on issue flux-framework/flux-core#7292.
A quick test on Tuolumne shows this is working in the basic case. I simulated an unkillable process by attaching gdb to the job shell then canceled the job….
markcmiller86 commented on issue visit-dav/visit#20752.
Ok, so both the HDF5 and Silo build failures are fixed by specifing -DMPI_HOME:STRING=/path/to/mpi/install/home. Now, Silo of course DOES NOT REQUIRE mpi. But, it does a find_package() on HDF5 and that will fail if the HDF5 it is finding depends on MPI but there is nothing telling Silo where to find MPI.
…
becker33 commented on issue spack/spack#51835.
@alalazo thanks! I’ve included it now…
jeremylt commented on issue openjournals/joss-reviews#8388.
My apologies, I was trying to wait for the libCEED release to finish, but unfortunately that became a much longer wait than I expected. I will proceed one way or another with a new Ratel release by next Friday no matter what libCEED does….
wihobbs commented on issue flux-framework/flux-core#7289.
> Actually, would it address this issue if we got rid of the names and affiliations in the list and just left the github IDs? …
cyrush commented on issue visit-dav/visit#20782.
We reviewed the error. The logic detects the case where int could cause badness and handles this with an error message. So the generic integer to int logic is ok, we can remove the #error.
…
alecbcs commented on issue spack/spack#51854.
Yep! Thanks for the suggestion @haampie. This works with ghcr.io/spack/centos7:
…
hppritcha commented on issue open-mpi/ompi#12979.
some notes on this. this doesn’t seem to be a problem with partial finalization in some ranks verses others. when the hang occurs I observe some process stuck in waiting for completion of the PMIx_fence_nb near the beginning of finalization while others are stuck in some waitall (on several hundred requests per call) within the application….