vsoch commented on issue rse-ops/flux-compose#4.
If we put an arbitrary large number there we don’t need to rebuild - they will just appear as down. What do you think?…
vsoch commented on issue kubernetes/enhancements#2033.
Usernetes, of course!…
markcmiller86 commented on issue betterscientificsoftware/bssw.io#2563.
> What if we don’t link SSF at all where it appears in the body and just let the MIF link carry all of the responsibility? …
cyrush commented on issue Alpine-DAV/ascent#1732.
@nicolemarsaglia does this resolve#1725?…
alecbcs commented on issue spack/spack-packages#5621.
@spackbot run pipeline…
tgamblin commented on issue spack/spack#52780.
Curious, what’s the performance improvement you’re seeing?…
balos1 commented on issue spack/spack#52762.
> I would take a step back: filter: is very generic for a niche feature. Can’t you do spack env create --filter filter.yaml? Does it even compose for your spack config get idea? I’m sure there are ambiguities regarding what filter refers to if you have includes/stacking.
…
becker33 commented on issue spack/spack#52769.
@haampie I’ve rebased this and confirmed that it overwrites the package hash appropriately. …
jedbrown commented on issue openjournals/joss#1563.
I agree with @sneakers-the-rat. Automation bias is strong in this context and LLMs embody biases that the editor may not even be aware of. “Accuracy” in historical decisions does not imply validity. If we were to deploy this and then audit the resulting system, we would almost certainly find a lack of construct validity and that “bad” recommendations from the LLM unacceptably influenced decisions. I think it also violates trust of authors and the broader community, and cheapens the meaning of a JOSS publication. …
v-dobrev commented on issue mfem/mfem#5426.
@najlkin, is this ready for review? If so, please add the label ready-for-review….
alecbcs commented on issue llnl/hubcast-test#6.
@lc-hubcast help…
JustinPrivitera commented on issue llnl/conduit#1652.
@mohityadav8 why don’t you choose between #1641 or #1632…
trws commented on issue llnl/camp#218.
Agreed. Forcing the change to a ULL seems reasonable, but I would note that even if it didn’t have the qualifier on it increasing the version to a number requiring that size would have produced warnings for anyone assigning the literal to a smaller value. Honestly I’d prefer we keep the format and force the size change rather than change the format, it’s less likely to cause downstream breakage that’s undetectable by the compiler….
BradWhitlock closed issue llnl/axom#1530.
Add function to compute curvature of a NURBSCurve/BezierCurve at a given parametric location
Functions to compute local curvature within a NURBS curve was removed from the C2CReader in #1526 …View Comment
tgamblin open issue linuxfoundation/insights#2030.
Repository groups for E4S
LFX insights is only analyzing the https://github.com/E4S-Project/e4s repository for E4S, but the project has many more. It would be good to include them….View Comment
garlick commented on issue flux-framework/rfc#541.
Thanks!…
trws commented on issue flux-framework/flux-sys-rs#4.
Ok, sounds good. If we end up wanting 2024 for the higher-level crate I think all we’d need to do is be sure we can reasonably use that toolchain in the buildbot so we could deploy modules using it in LC. Something for later. …
grondo commented on issue flux-framework/flux-security#234.
Thanks! I’ve set MWP….
wihobbs commented on issue flux-framework/flux-sched#1528.
Working off of pure vibes, I want to see if this passes CI. Hence the change to draft….
jameshcorbett commented on issue flux-framework/flux-sched#1534.
> Note: it is difficult to change the phrasing of statements like opt_p.value_or (null_planner) because none of boost::optional’s operators cleanly express “empty or populated with nullptr” or the inverse in a single statement (opt_x == boost::optional (nullptr) returns false when empty and true when null), and try_at only returns references but boost::optional cannot be initialized with an rvalue reference (see Optional references).
…
trws commented on issue flux-framework/flux-sched#1535.
I honestly don’t really care, just didn’t realize it was a separate issue tickling that one. It might be worth doing that when we update to the next TOSS release so everything works in our normal production environment but I don’t see it being worth all that much right now. Thanks for this @wihobbs….
grondo commented on issue flux-framework/flux-core#6440.
Having something that could run on new code would really nice now that we’re making heavy use of LLMs which seem to forget half the time some of the formatting instructions they’re given….
trws commented on issue flux-framework/flux-core#6440.
Adding custom handling to <, == and != to the thing I linked above with never as the policy might take care of those trailing breaks that were bugging us. The last time I was seriously thinking about this, my plan was to add clang-format to pre-commit, and have it run either only on files with diffs or only on the diff itself so we could at least count on it for new code. …
garlick commented on issue flux-framework/flux-core#6440.
Closer. …
cmoussa1 commented on issue flux-framework/flux-accounting#916.
Thanks for catching that issue in the commit message @jameshcorbett - I’ll go ahead and remove that from the message altogether. The .backup DBs are automatically cleaned up in the testsuite, but you are right, the backup isn’t deleted if the update completes successfully. I guess the thinking here was that it would be an admin’s responsibility to manually clean this up for this reason:
…
jeremylt commented on issue CEED/libCEED#1830.
If you’d like to take a crack at it, feel free! It should be largely the same code copied over, but you have to make some small changes to account for the difference between CUDA and HIP syntax…
jeremylt commented on issue 4C-multiphysics/4C#2133.
I added folks who I’ve seen approve changes to these files in the git history…
cyrush closed issue visit-dav/visit#20153.
Create macOS 12+ Style Icon for VisIt?
I’ve noticed that of all the icons in my macOS Dock, VisIt is uniquely non-conformant to the icon style that has been consistent post-macOS 12. …View Comment
JustinPrivitera commented on issue visit-dav/visit#20950.
I want to make sure we error instead of crashing….
becker33 commented on issue spack/spack-packages#5711.
It might be worth showing a warning, but I think this is generally good (for two primary reasons). …
imguoguo commented on issue sipeed/NanoKVM#442.
Since we
hppritcha commented on issue open-mpi/ompi-www#569.
@jsquyres noted these missing presentations prior to HPSF meeting…
adrienbernede commented on issue llnl/radiuss-spack-configs#195.
@tdrwenski @davidbeckingsale …
adrienbernede commented on issue llnl/RAJAPerf#688.
@smithsg84 I’d appreciate a review of the changes in RAJAPerf spack package: https://github.com/spack/spack-packages/pull/5716…
artv3 commented on issue llnl/RAJA#2036.
@lc-hubcast approve…
hppritcha commented on issue hpsfoundation/tac#93.
> @hppritcha thanks for he responses above. The link https://github.com/orgs/open-mpi/teams/ompi-devel-merge is not publicly visible, unfortunately….
samuelkgutierrez closed issue hpc/quo-vadis#542.
API Simplification?
Hello,…View Comment
tpatki open issue flux-framework/flux-multi-cluster-utilities#30.
Add a performance-based policy with XOR support (reopen #30 with updates)
Add support for job performance lookup-based policy to enable evaluation of demo workflow. …View Comment
chu11 commented on issue flux-framework/flux-core#7730.
> maybe on_sigchld() with a status argument would be better callback name here? …
tpatki open issue flux-framework/PerfFlowAspect#202.
`WeavePassPlugin` INTERFACE target is missing `-fpass-plugin=`, so downstream consumers never actually run the pass
Running into this as we build gromacs for the paper for @spencer-gre and @alizalisan. …View Comment
imguoguo closed issue sipeed/NanoKVM-Pro#95.
Mark Images icon with colour when image is mounted (like it is for Power icon)
Currently to check if image is mounted you have to click on Images icon and in submenu the Mount Image text will be coloured if it is….View Comment
hppritcha commented on issue open-mpi/ompi#13133.
@kingshuk00 are you planning to work on this further? Note a related PR #14083 will be merged soon….
v-dobrev commented on issue mfem/PyMFEM#315.
Hi @sshiraiwa, thanks for reviewing this. There are CI failures, however, they are not related to this PR – are you okay if I merge this, regardless of the failures? …
BradWhitlock commented on issue llnl/conduit#1645.
Hey @cyrush , I see that you approved this PR already so I’m merging it….
jandrej commented on issue EnzymeAD/Enzyme#2964.
> @jandrej this failure is presumably from upstream mfem? …
tpatki commented on issue llnl/variorum#591.
Update from 7/21. …
bgunnar5 commented on issue llnl/mada-tools#10.
> Did the gitlab ci pipeline pass for this? I can’t find it in the pipeline list for some reason. …
eleon commented on issue hpc/quo-vadis#9.
@samuelkgutierrez , could you please try the test test-mpi-phases? …
wihobbs commented on issue flux-framework/rfc#529.
@grondo Okay with adding MWP to this?…
tpatki commented on issue flux-framework/flux-rest-server#6.
…
jameshcorbett commented on issue flux-framework/flux-accounting#912.
roger roger big boss…
cbritopacheco commented on issue cbritopacheco/rodin#311.
@copilot Fix the CI….
emily-howell commented on issue Alpine-DAV/ascent#1751.
Thank Justin for finding it and making a ticket!…
camierjs closed issue mfem/mfem#5363.
[CUDA] nvcc / cicc memory exhaustion (OOM) compiling element assembly kernels with -arch=sm_120 (CUDA 12.8)
# cicc memory ballooning / OOM compiling element-assembly kernels for sm_120 (Blackwell) with CUDA 12.8…View Comment
vsoch commented on issue flux-framework/flux-framework.github.io#205.
Thanks for the 3rd comment, Mergify bot. I almost forgot you were there after the first two. You and coffee bot must be good buds. …
JustinPrivitera commented on issue visit-dav/visit#20980.
Thanks for these changes! Everything is merged and should be released in VisIt 3.5.1….
tpatki commented on issue llnl/variorum#591.
Spent a few hours debugging the segfault with amd_smi APIs. Summary from claude below.
…
BradWhitlock open issue llnl/axom#1530.
Add function to compute curvature of a NURBSCurve/BezierCurve at a given parametric location
Functions to compute local curvature within a NURBS curve was removed from the C2CReader in #1526 …View Comment
grondo open issue flux-framework/flux-core#7732.
idea: define queue membership by hostlist with implied property == queue name
This is an alternate take on #5741, with the end result being similar for admins configuring queues….View Comment
garlick commented on issue flux-framework/flux-core#7731.
I repushed with the suggested changes, plus addressed some test failures that were a bash (redhat) vs dash (debian) thing….
alecbcs commented on issue spack/spack-packages#5620.
Package names appear to always be normalized by PyPi so we just need to match the same case / separator as is used in the actual tarball name for {name} and it’ll work….
imguoguo closed issue sipeed/NanoKVM#551.
API endpoints throwing errors when HID-Only mode is enabled
API endpoints “/api/storage/image/mounted” and “/api/storage/cdrom” throw errors when HID-Only mode is enabled…View Comment
camierjs commented on issue mfem/mfem#5387.
Really neat! Thank you @tdrwenski !…
jandrej commented on issue mfem/mfem#5387.
> We should mention this new feature in the CHANGELOG.
…
adrienbernede open issue llnl/radiuss-spack-configs#191.
Consider using /p/vast1 instead of /usr/workspace/${user} for spack install and buildcache storage
/p/vast1 now has daily snapshot….View Comment
cyrush closed issue llnl/conduit#1649.
c++20 fmt runtime param issue with relay silo paths
Another case we need to use conduit_fmt::runtime, it may be that our c++20 testing doesn’t have silo enable?…View Comment
JustinPrivitera commented on issue llnl/conduit#1651.
what is conduit_fmt::runtime and why do we need it?…
adayton1 commented on issue llnl/axom#1912.
Instead of removing the cpp14 variant, can’t you just make it conflict with the version that requires c++17 and any version newer than it?…
adrienbernede commented on issue llnl/CARE#379.
> > If you can get access to a bank (other than guests) on matrix, the best is to define it as the default one. But otherwise, yes, we can easily set the CI to use it in the alloc options (.gitlab/custom-variables.yml). …
chu11 commented on issue flux-framework/rfc#535.
doh! i totally forgot about the rfc. thanks!…
tpatki open issue flux-framework/flux-sched#1532.
Support per-xor-slot duration and tasks for XOR jobspecs?
## Problem…View Comment
tpatki open issue flux-framework/flux-multi-cluster-utilities#33.
Add a "optimize-all" policy
Design a policy that optimizes for “predicted shortest makespan” : this will combine the least-pending-jobs and the performance-based policies to demonstrate an overall optimization….View Comment
chu11 commented on issue flux-framework/flux-core#7729.
> These are vendors being motivated to be flux developers so I figured that was probably the right location? …
cmoussa1 commented on issue flux-framework/flux-accounting#556.
Going to close this for now….
cjekel commented on issue cjekel/piecewise_linear_fit_py#134.
@doronbehar should be fixed now…
cbritopacheco commented on issue cbritopacheco/rodin#314.
Superseded by #315. This branch still carried the six commits that #312 squash-merged into develop, so it conflicted with develop
jeremylt commented on issue CEED/libCEED#1988.
Yeah, go ahead and remove the one outside of the init function since it isn’t helping us anymore …
cyrush commented on issue Alpine-DAV/ascent#1742.
Yes, space is tight for static cuda builds – working on a potential fix. …
jeremylt commented on issue 4C-multiphysics/4C#2012.
Fantastic, then the user just needs to get the correct flags sent to Kokkos via building Trilinos correctly. …
tgamblin commented on issue spack/governance#6.
@rbberger yes please update Phil to https://sig.com…
imguoguo commented on issue sipeed/sipeed_wiki#988.
v-dobrev commented on issue mfem/mfem#5406.
Merged in next for testing……
artv3 commented on issue llnl/RAJA#2006.
@llnl/raja-core , is this ready to go? …
tomstitt commented on issue llnl/RAJA#2059.
from an app perspective this looks great! thanks @artv3 !…
samuelkgutierrez commented on issue hpc/quo-vadis#542.
And I guess qv_process_scope_get could become qv_process_get and qv_mpi_scope_get could become qv_mpi_get? This one isn’t as clear to me, however. Open to suggestions here….
eleon commented on issue hpc/quo-vadis#540.
Thanks for asking, @samuelkgutierrez . I prefer CLOSE. Close seems to be a more straightforward concept, while affinity preserving, in my opinion, would require explaining to users (including me) what it means :) …
sam-maloney commented on issue flux-framework/flux-core#7488.
To add the only salient info from my duplicate issue, my log was missing the “got SIGTERM” from rank 1 instead of rank 0, so it could be either order….
jameshcorbett commented on issue flux-framework/flux-coral2#161.
@behlendorf indicated that this issue and https://github.com/flux-framework/flux-coral2/issues/175 are the highest priority right now….
camierjs commented on issue GLVis/data#10.
Is this on top of #9 ? macos-10.15 was not affected then….
jandrej commented on issue EnzymeAD/Enzyme#2948.
updated and removed inactivity list file from history…
hppritcha commented on issue open-mpi/ompi#14171.
oh ignore my comments about Fortran. I forgot we’d already added a sub implementation that had the fortran interfaces….
hppritcha commented on issue open-mpi/mtt#956.
The error that started this problem is likely this based on the error log:…
trws commented on issue flux-framework/flux-sched#1530.
It might, I’d have to look at the modprobe settings. That’s my best bet for the issue cause though, if feasibility stayed loaded then the dependent so would too. Also this would always happen on systems with musl, and there’s no way to fix it there (musl doesn’t actually unload shared objects)….
sam-maloney commented on issue flux-framework/flux-sched#1529.
It’s not making my life particularly difficult, no, but I did find it a bit needlessly confusing if the listed versions aren’t actually indicating real requirements
jameshcorbett commented on issue flux-framework/flux-sched#1530.
I just ran module remove sched-fluxion-qmanager/resource && flux module reload resource && module load sched-fluxion-resource, I didn’t manually touch feasibility at all. I thought removing resource or sched-fluxion-resource would force feasibility to exit?…
sam-maloney commented on issue flux-framework/flux-accounting#885.
Everything seems to be passing well in my latest builds! Many thanks @cmoussa1
tgamblin commented on issue spack/spack#52665.
@haampie ok, caching is disabled for spliced specs, and the traversal logic is simpler. I think this is a decent backport. the diff looks big but some code just moved into a try/finally block. …
wihobbs open issue flux-framework/flux-sched#1531.
valgrind test is noisy due to hwloc errors
With some frequency on the ci / jammy - test-install (pull_request) check we’re seeing the following issue, which appears to be from hwloc and goes away. Is this a known issue? Is there something we should be doing to suppress the warning?…View Comment
wihobbs commented on issue flux-framework/rfc#529.
Ok, that seems to have consensus. Can someone push the approve button? Then I’ll go update my PR to Fluxion….
milroy commented on issue flux-framework/flux-sched#1316.
Thanks for the helpful feedback and improvements @vsoch, @jameshcorbett, and @trws! Setting MWP….