markcmiller86 commented on issue visit-dav/visit#20752.
The boxlib failure appears to be repeatable and has to do with switching between 2D and 3D data. …
cyrush commented on issue visit-dav/visit#20752.
For the blueprint mfem element fields those many not be deterministic across platforms. I will take a look….
imguoguo closed issue sipeed/picoclaw#805.
[Feature]: Add TUI support via new repository picoclaw_tui
### Description…View Comment
adrienbernede commented on issue llnl/RAJAPerf#642.
@rhornung67 This is ready….
adrienbernede commented on issue llnl/RAJA#1989.
@rhornung67 This is ready….
grondo commented on issue flux-framework/flux-sched#1433.
Hm, it should. Note that ./configure is just a wrapper for cmake in this project….
grondo commented on issue flux-framework/flux-core#7410.
Oh, and one other trivial testsuite fix thrown in….
vsoch open issue feiskyer/mcp-kubernetes-server#47.
Support for `kubectl explain`
Hi @feiskyer! …View Comment
btalamini closed issue EnzymeAD/Enzyme#2715.
Custom reverse mode rule error
fn:; Function Attrs: mustprogress ssp willreturn memory(readwrite) uwtable(sync)…View Comment
tzanio commented on issue mfem/mfem#5253.
Merged in next for testing……
cjvogl commented on issue mfem/mfem#5126.
> We should also update the website to mention GSLIB and that we do support GPU builds now. …
cyrush open issue llnl/Silo#518.
hdf2.0 cmake var issue
https://github.com/llnl/Silo/blob/0cddaa6865fb49ae7b2110fa1b6f2709592abe0b/CMakeLists.txt#L237…View Comment
adayton1 commented on issue llnl/CHAI#359.
> @adayton1 I don’t see the open-source release info anywhere. Was that ever done for CHAI? For example, https://github.com/llnl/RAJA/blob/develop/COPYRIGHT#L6 …
jeremylt commented on issue CEED/libCEED#1933.
Here is a summary of different open source policies that @jedbrown found: https://blog.probabl.ai/maintaining-open-source-age-of-gen-ai…
JustinPrivitera open issue visit-dav/visit#20837.
MFEM plot limits not changing is confusing
When you make changes to LOR options in the MultiResControl operator when plotting MFEM data, you expect to see numerical differences in the data extents….View Comment
becker33 commented on issue spack/spack-packages#3505.
Additional things I’ve learned:…
hppritcha commented on issue open-mpi/ompi#13747.
could we get this cherry picked back to 6.0.x?…
balos1 closed issue llnl/sundials#820.
2026 SUNDIALS Virtual Birds-of-a-Feather
The SUNDIALS team will be holding a virtual Birds
adrienbernede commented on issue llnl/Umpire#1053.
@davidbeckingsale this is ready for merge….
adrienbernede commented on issue llnl/Caliper#687.
Merged as part of #712 …
adrienbernede commented on issue llnl/CHAI#356.
@adayton1 You may ignore the remaining failure status merge: it
cmoussa1 commented on issue flux-framework/flux-core#7408.
Done! Thanks @chu11 …
chu11 commented on issue flux-framework/flux-core#7408.
As an aside, if this typo was missed by the typo checker, you can submit it to https://github.com/crate-ci/typos/issues/1488. …
grondo commented on issue flux-framework/flux-coral2#452.
One other point is that the scheme above won’t apply for a subinstance launched with some brokers on rabbits. The rabbit property will be preserved, but the default queue in the instance will happily run stuff on the rabbits by default. Users will have to be aware that they need to run normal jobs using --requires=-rabbit, or we can look into ways automatically set up queues, exclude rabbits by default, or whatever makes the most sense for the probable use cases….
gardner48 commented on issue erf-model/ERF#2921.
@adamwise95 if you want this case to run with make test i.e., as part of the GitHub and GitLab testing. This file will need to go in a subdirectory of Tests/test_files and you’ll need to add a add_test_r() line for this case in Tests/CTestList.cmake. …
cmoussa1 commented on issue crate-ci/typos#1488.
```…
cbritopacheco commented on issue cbritopacheco/rodin#173.
@copilot Here is the problem summary. …
benzwick open issue benzwick/guix-mvox#3.
Guix Systole CI: channel fails to compile against current Guix
The Guix Systole and Guix Slicer CI workflows fail because the guix-systole channel modules don’t compile against current Guix during guix pull….View Comment
benzwick commented on issue SystoleOS/guix-systole#102.
It fails on my laptop as well:…
balos1 commented on issue AMReX-Codes/amrex#5042.
I don’t think that this is a problem. Initialize is only meant to be called once. Right now it is called in amrex::Initialize which, I thought, could only be called once. So, I believe, the case Codex is worried about will never happen. …
gardner48 commented on issue AMReX-Codes/amrex#5044.
I this change is correct, but I noticed a couple other calls that need to be fixed in the implicit MRI case. I’ll open a PR for this change and the others in a sec….
hppritcha commented on issue openpmix/openpmix#3821.
i’d ignore the mpi4py failure….
hppritcha commented on issue mpi4py/mpi4py#733.
this PR broke our open mpi github action: …
jeremylt commented on issue mfem/mfem#5247.
Hmm, interesting. This is not a flag passed directly to libCEED, so I’m not sure exactly how MFEM processes this command line option….
mplegendre commented on issue llnl/Spindle#140.
I’ve no objections to the code changes. But do you have needs for spindle on either ppc64le or aarch64? …
artv3 commented on issue llnl/RAJA#1970.
Can we add some testing and examples? Documentation would also be nice…
bhatele commented on issue hpcgroup/pssg-plots#14.
> _ No description provided. _ …
v-dobrev commented on issue mfem/mfem#5195.
Ok, thanks @dylan-copeland. I see your point about DecodeDof, so let’s keep it as it is….
dylan-copeland commented on issue mfem/mfem#5195.
@v-dobrev commit afded067a74594bb71520832ae5024f1556cbd75 replaces absdof. Thanks for pointing this out. In fespace, DecodeDof is more complicated, as it has a version returning a sign. This commit uses UnsignIndex for the implementation of the basic DecodeDof, but I am keeping DecodeDof in fespace as is. An alternative would be to add another version UnsignIndex(int i, real_t& sign) to replace DecodeDof, but I don’t see a need for that in general, outside FiniteElementSpace….
JustinPrivitera open issue llnl/conduit#1526.
Time series in one root file
It would be nice to support this….View Comment
garlick commented on issue flux-framework/flux-pmix#122.
Thanks, by chance do you know of any test code we could pull in to CI to measure success if we fix this?…
camierjs commented on issue CEED/Laghos#170.
Closing this issue as all tests are passing with MFEM, HYPRE, Laghos masters on GPU. …
wihobbs commented on issue flux-framework/flux-core#7244.
Ok @grondo, at long last this is ready for a real human to review it. :) …
brendankeith commented on issue mfem/web#305.
@mlstowell please let me know if I’ve addressed your comments, or if you have any new ones :)…
tzanio commented on issue mfem/mfem#5212.
ping: @mlstowell …
cyrush closed issue llnl/conduit#1499.
h5zzfp build script fails on redhat systems using lib64 instead of lib paths
Running build_conduit.sh on a red hat system is failing when it gets to searching for the zfp package previously installed….View Comment
markcmiller86 commented on issue llnl/H5Z-ZFP#159.
> I believe it’s also time to eliminate the make option, as the CMake implementation is now quite mature and continues to improve….
milroy commented on issue flux-framework/flux-sched#1408.
> Once this is merged, you should be able to test using cpu-affinity=dry-run and FLUX_HWLOC_XMLFILE of another system without errors by also setting FLUX_HWLOC_XMLFILE_NOT_THISSYSTEM in the environment of flux start. i.e. before you call test_under_flux. …
garlick commented on issue flux-framework/flux-core#7403.
Oh I see you Claude added
…
vsoch open issue converged-computing/fractale#5.
Issues to think about for MCP
- The way agents use algorithms: right now I’m going to design step “sub-agents” that are allowed to call tools to assist in their decision making. Those tools can be databases to get / query / summarize data (e.g., input for a model) or call to a model itself to make some decision….View Comment
cbritopacheco commented on issue cbritopacheco/rodin#173.
@copilot I have tidied up the approach and switched to a more robust scheme. However, the noSplit feature seems to not be taken into effect. Please check for errors and fix….
benzwick commented on issue swig/swig#3337.
@wsfulton I’ve pushed a branch pull-3337-bz based on your pull-3337 with the test tweaked to reproduce the issue….
cyrush commented on issue llnl/blt#755.
I think current fix is needed, and we can improve more in the future. …
BradWhitlock commented on issue llnl/axom#1790.
I fixed this specific warning in PR #1796 ….
vsoch commented on issue kubeflow/mpi-operator#12.
Thanks @garlick! …
grondo commented on issue flux-framework/flux-sched#1408.
@milroy See https://github.com/flux-framework/flux-core/pull/7402. …
chu11 commented on issue flux-framework/flux-core#7399.
lol, it’s so obvious in hindsight. Just got settled on a goal and that was it :-)…
grondo commented on issue flux-framework/flux-core#7402.
Thanks! I’ll set MWP….
jameshcorbett commented on issue flux-framework/flux-coral2#452.
> > maybe we should revert to a config on hetchy that places the rabbits in a separate queue so we can work through the fallout with configuration and prolog/epilog/housekeeping scripts in parallel with gathering more detailed requirements?…
garlick commented on issue flux-framework/flux-coral2#452.
That seems like a sensible approach and your view of the rabbits sounds right to me….
jeremylt commented on issue CEED/libCEED#1931.
I’m pondering. I want to make sure we are not violating any implicit assumptions for the user …
JustinPrivitera open issue visit-dav/visit#20824.
Add Data Binning Operator Documentation
We are missing Data Binning documentation. This page could use a link to that documentation when/if it is added….View Comment
trws commented on issue spack/spack-packages#3453.
Ah cool, I missed it when I scanned back over the package, LGTM then. :+1:…
becker33 commented on issue spack/spack#51960.
I think this change is going to make adding the per-package locking for the new installer very difficult. I think you’re going to need access to the overwrite information when you approach the install for each package, because an intervening process may have installed it between when you constructed the BuildGraph and when the build gets to that install. I think each call to worker_function needs to know whether it is an overwrite build and what time it was labelled as an overwrite build, so that it can accept if another process already overwrote in the meantime….
hppritcha commented on issue open-mpi/ompi#13734.
ompi_rte_init is only called within ompi_mpi_instance_init which can in principle be invoked via multiple threads concurrently since MPI_Session_init (correctly recalled by @rch54) can be called by individual threads. totally different from MPI_Init/MPI_INit_thread and one of the reasons for its presence in the standard.
…
garlick commented on issue kubeflow/mpi-operator#12.
Ralph is correct that pmix support in flux is minimim viable, but we do run some mpi regression tests in CI like OSU benchmarks, so the stuff most people use should work fine….
bhatele open issue hpcgroup/pssg-plots#11.
add heatmap support to pssg-plots
wihobbs commented on issue flux-framework/rfc#494.
Thanks! setting MWP…
garlick commented on issue flux-framework/rfc#493.
OK, how’s this look?…
garlick commented on issue flux-framework/flux-foundry#24.
Thanks, I’ll just press the button since we don’t have mergify set up….
cmoussa1 commented on issue flux-framework/flux-accounting#818.
I’m still seeing some unexpected and random errors from the coverage builder during the create_db() call. I wonder if it has anything to do with the syntax of creating the DB itself. The sqlite3 documentation examples use = instead of : in their .connect() call, and I wonder if this syntax error might be causing the creation of the database to not be opened with a read-write connection. Maybe this is only being caught by CI every now and then. I’m going to add another commit to change this .connect() call and see if that resolves it….
alecbcs commented on issue spack/spack-packages#3453.
No matter the targets given in the variant, the package always additionally adds the host architecture. I believe this was originally done for targets=none still building a valid installation on a users system, but it works nicely for this case too where we provide minimal cross compilation support with few architectures….
alecbcs commented on issue spack/spack#51813.
@alalazo what version of git do you have locally installed?…
hppritcha commented on issue openpmix/prrte#2407.
I like doing things in stages. So RAS component comes first. …
v-dobrev commented on issue mfem/mfem#5222.
Re-merged in next for testing……
imguoguo commented on issue sipeed/picoclaw#281.
Thank you for your PR! I’m very sorry, but another PR (#259) has already solved the same issue. …
tgamblin commented on issue spack/spack-packages#2947.
I believe this builds on spack/spack#51817, but note that it was reverted in spack/spack#51943 due to windows CI issues….
tgamblin commented on issue spack/spack#51943.
If rebuild-index passes on windows CI this should be good to merge….
markcmiller86 commented on issue llnl/Silo#243.
@brtnfld and @eschnett I just recently ran into this same issue @brtnfld reports in the above comment in Silo 4.12.0. I have a patch for it too. The PR for that patch includes some other items and will be merged soon….
brendankeith commented on issue mfem/web#305.
Thanks for your comments, @mlstowell and @acfisher! …
hppritcha commented on issue flux-framework/flux-core#7380.
Ah great! I will Check this Option out….
SteVwonder closed issue SteVwonder/dotfiles#31.
Upgrade emacs config
- Switch to built-in use-package…View Comment
alecbcs closed issue spack/spack-packages#3349.
Installation issue: adios2@2.11 race condition in cmake
### Steps to reproduce the issue…View Comment
vsoch commented on issue rootless-containers/usernetes#385.
@AkihiroSuda for your subuidless project it looks like you’ve done:…
vsoch commented on issue oras-project/oras-py#233.
Just updated the tag. It points to the commit right after the one you referenced, because there was a tributors update….
v-dobrev commented on issue mfem/mfem#5222.
Merged in next for testing……
camierjs commented on issue GLVis/glvis#106.
This issue should be resolved by the changes in PR #339, which adds support for displaying DOF numberings….
tgamblin commented on issue spack/spack-packages#3404.
> Should we also deprecate versions from e.g. 2020 and before (that would leave 2 versions non -deprecated)? …
acfisher commented on issue mfem/web#305.
> @mlstowell and @acfisher, I’d really like to use these annotated examples as a resource for my class. …
tzanio commented on issue mfem/mfem#5234.
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….
adayton1 commented on issue llnl/radiuss-shared-ci#71.
Is it worth considering “LLNL_RSCI_” as a prefix? “RSCI_” seems unlikely to conflict, but if it does happen it would be really hard to track down….
adrienbernede commented on issue llnl/radiuss-shared-ci#71.
@tdrwenski I would prefer the prefix to be “RSCI_” rather than “RSC_” because I use “RSC” all the time to refer to radiuss-spack-configs……
cyrush commented on issue llnl/conduit#1495.
Trying a few more combos of env vars to allow openmpi to oversubscribe - that is launch more MPI ranks than procs available. …
jeremylt commented on issue jeremylt/LFAToolkit.jl#94.
Ok, up the priority here for Colorado Conference…
garlick open issue flux-framework/flux-pmix#122.
support advanced PMIX features like groups
Problem: @hppritcha alluded to problems using advanced MPI 4 features under flux-pmix in https://github.com/flux-framework/flux-core/issues/7380#issuecomment-3892282986…View Comment
chu11 closed issue flux-framework/flux-core#7266.
kvs: optimize against transactions that are empty
There can be situations where an empty transaction could be committed to the KVS (see #7257). This is harmless, but a large swath of “kvs commit” code is actually done even if the ultimate result is to do “zero operations” / “zero changes”….View Comment
grondo commented on issue flux-framework/flux-core#7376.
I tried to reproduce this by preloading a uname() implementation that returns a zero-length hostname, but got a different failure:…
garlick commented on issue flux-framework/flux-core#7380.
> I am more interested in technical feedback like use of the idset api etc and not philosophical questions about why open MPI developers like me would not want prrte to be easy to use for our user community….
hppritcha commented on issue flux-framework/flux-core#7380.
No if there’s a flux RAS there’s no need for hostfiles…
cmoussa1 commented on issue flux-framework/flux-accounting#816.
Thanks @jameshcorbett! Setting MWP here…