JustinPrivitera open issue visit-dav/visit#19605.
Make CSV an entry in the Export Options
We should add CSV to the list of export options. When you select it, it will use Blueprint to CSV to do what you want. …View Comment
tzanio commented on issue mfem/web#253.
@IdoAkkerman, can you resolve conflicts and take a final look now that https://github.com/mfem/mfem/pull/3999 is merged?…
tzanio commented on issue mfem/mfem#4354.
Merged in next for testing……
trws commented on issue flux-framework/flux-core#6041.
> Thanks! This LGTM!…
grondo commented on issue flux-framework/flux-core#6044.
MPI version is cray-mpich/8.1.23…
tpatki closed issue LLNL/variorum#553.
Newer MI200+ GPUs report GPU-domain-level (GPU socket-level) power data as opposed to individual devices
As an example from our MI200 node, here’s what rocm-smi(v6.0.2 onwards) reports. Note here that GPU devices 1,3,5,7 do not report power values but report a N/A instead. This is by design.
…View Comment
adayton1 commented on issue LLNL/RAJA#1672.
@MrBurmark, would you be willing to look at the changes to hip atomics? The previous implementation of atomicCAS relied on volatile, so I changed the implementation to use an atomic load. To avoid a circular dependency, I’ve changed in the implementation of atomicLoad to use the intrinsic if available, otherwise I fall back to atomicOr(address, 0). Does that make sense, or would it be better to fall back to atomicCAS(address, 0, 0) or something else? I think atomicOr will be better than atomicAdd, but I’m not sure if atomicCAS can avoid the write in some cases. I also modified atomicExchange to use reinterpret casting instead of atomicCAS in a loop, then made atomicStore use atomicExchange if the intrinsic is not available….
v-dobrev commented on issue mfem/mfem#4352.
One thing is to make sure all packages are built with the same compiler and that mpicc and mpicxx use the correct compiler, e.g. by running mpicc --version – this should show the real compiler used by the MPI C wrapper, mpicc.
…
jeremylt open issue freeCodeCamp/freeCodeCamp#55194.
Profile GitLab Link
### Is your feature request related to a problem? Please describe….View Comment
milroy commented on issue flux-framework/flux-sched#1222.
I can confirm that the test that simply creates and cancels 10,000 jobs doesn’t get killed by oomkiller. It does eventually complete after 47 minutes, so it may not reproduce the title issue….
garlick commented on issue flux-framework/flux-core#6040.
just pushed some tweaks to the admin description to fix my poor word choices and also add the --with-imp option to the perilog-run example….
tgamblin commented on issue flux-framework/flux-core#6041.
Ping @jefflong6 @xorjane…
jameshcorbett open issue flux-framework/flux-coral2#166.
Stop daemon on compute nodes after jobtap prolog completes
There are clientmount daemons running on every compute node to handle the mounting and unmounting of rabbit file systems. The daemons produce noise, and there have been some investigations lately into how to reduce it. In theory the daemons only need to be running when there are file systems to mount or unmount, at the beginning and end of jobs. …View Comment
garlick commented on issue flux-framework/flux-coral2#166.
I think that should be OK! A quick test might be a good idea though. We’ve been surprised on el cap before :-)…
grondo commented on issue flux-framework/flux-coral2#166.
> Thoughts @grondo ? Would there be performance issues from having every node read the eventlog? (I hope not, because I think we do some coral2 plugins do this elsewhere.) …
mplegendre open issue dyninst/dyninst#1763.
Missing entries in DWARF line information lookup
When looking up line information entries by address, SymtabAPI is failing to return some line entries that should exist. …View Comment
jedbrown commented on issue aakritisi/FreeCAD#1.
This looks good in terms of content. I think the commits should be squashed and rebased (rather than merging from upstream) into atomic parts intended for review. And then we need to write out in the PR what changes are being made and why. …
chu11 closed issue LLNL/magpie#343.
fix cornercase / functionality / default tests
It has been a long time since I have tested “default”, “cornercase” and “functionality” tests (I normally just test specific hadoop/spark versions). it appears various test errors have creeped in over the years as little things in newer versions have changed. …View Comment
cyrush open issue Alpine-DAV/ascent#1317.
replay -- parse actions before reading data
best to know if you have malformed actions before you spend effort doing I/O. …View Comment
v-dobrev open issue xiaoyeli/superlu_dist#166.
Potential issue when using PARMETIS column permutation in v8.2.1
I noticed this during testing of mfem-4.7 with SuperLU_DIST v8.2.1 – one of the mfem examples that uses SuperLU_DIST, examples/ex11p.cpp hangs when run with:
…View Comment
alecbcs commented on issue spack/spack#44420.
@tldahlgren does this PR pass your checks for homepages and versions now?…
adrienbernede commented on issue spack/spack#44571.
@white238 I think this https://github.com/spack/spack/pull/44699 is what you were looking for….
markcmiller86 commented on issue markcmiller86/hello-numerical-world#14.
Hmmm…that looks like a new warning with your changes from snprintf to sprintf. It was using snprintf before. Can you revert your changes and produce the error message the code was originally emitting?…
trws commented on issue flux-framework/flux-sched#1222.
Oh wow. They died on a cancel from resource.acquire. I’m actually not sure how we could get to a FATAL: exception not rethrown like that, it actually looks like what would happen if you hit an uncaught throw inside a destructor and terminated the process. Did it OOM or something?…
balos1 commented on issue LLNL/sundials#501.
@gardner48 The “if” version sounds like a nice addition as well. …
corbett5 closed issue ITensor/ITensorMPOConstruction.jl#17.
Follow style guides for casing of function names and keyword arguments
In Julia, it is standard to use snake casing rather than camel casing, see style guides such as https://docs.julialang.org/en/v1/manual/style-guide and https://github.com/JuliaDiff/BlueStyle….View Comment
jeremylt closed issue Eudicods/outworlds-wastes#15.
Official Release
I would like to release a stable version of the rules around the time that the Mercenaries Kickstarter boxes arrive for people. …View Comment
adrienbernede commented on issue CEED/libCEED#747.
@jeremylt I
markcmiller86 commented on issue visit-dav/visit#19593.
There are a few ways to go here. …
v-dobrev commented on issue spack/spack#44620.
I see that in develop almost all deprecated versions of openssl were removed in #44629. What do we want to do with the versions that are being deprecated by this PR: 3.0.13, 3.1.5, 3.2.1, and 3.3.0? Should we keep them as deprecated or remove them? @haampie, what do you suggest?
…
cyrush commented on issue spack/spack#44571.
Thanks @tldahlgren. …
jedbrown commented on issue rsmpi/rsmpi#183.
If this is something we can document or otherwise make more robust, feel free to open a PR to update the docs. …
jedbrown commented on issue openjournals/joss-reviews#6288.
Thanks @victorapm! …
wihobbs commented on issue flux-framework/flux-sched#1192.
Oh, oops, I had forgotten this was open as a PR. Ignore that push, that was just to sync across machines. …
chu11 commented on issue flux-framework/flux-core#4336.
This job-list w/ a database has been re-worked: …
bgunnar5 commented on issue LLNL/merlin#486.
Need to fix issue where latest changes weren’t pulled for some reason. This is a reminder to not merge this branch until that’s fixed…
hppritcha commented on issue open-mpi/ompi#12584.
Thanks for trying this. It looks like some of these may well be attributed to Open MPI. Most of the functions caught will only be run once in the call to MPI\_Init but i’m not sure about the others. Not sure about your second question.
…
cmoussa1 commented on issue flux-framework/flux-accounting#455.
OK, I’ve force-pushed up some changes (notably to the plugin: add "age" factor in priority calculation commit) based on the feedback above that I hope begins to capture the number of cases we talked about. I can break this up into multiple commits if needed, but I thought it might be simpler to have it all in just one commit.
…
jedbrown commented on issue SRI-CSL/PVS#92.
Hi :wave:, thanks for looking into this. Have you gotten a chance to make progress or would scripting around Cl
CamStan commented on issue LLNL/scr#452.
This feature was added to bootstrap.sh in #543 …
artv3 commented on issue LLNL/RAJA#1665.
Adding an example could be really nice too! You can probably just take one from the unit test. …
adrienbernede commented on issue LLNL/CHAI#263.
@adayton1 Thank you for spotting that. …
v-dobrev commented on issue GLVis/glvis#283.
We primarily use and test with Apple clang on Mac, so I’d recommend using that. …
corbett5 commented on issue GEOS-DEV/LvArray#321.
Yeah after a tad more thought I doubt it’s an OOM error since in those cases we usually see the error from Umpire directly. But certainly if it’s not terribly slow running in full debug mode will hopefully tell us more….
CamStan open issue ECP-VeloC/spath#30.
spath release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
CamStan open issue ECP-VeloC/shuffile#33.
Shuffile release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
CamStan open issue ECP-VeloC/redset#56.
Redset release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
CamStan open issue ECP-VeloC/rankstr#24.
Rankstr release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
CamStan open issue ECP-VeloC/er#42.
ER release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
CamStan open issue ECP-VeloC/KVTree#73.
KVTree release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
CamStan open issue ECP-VeloC/AXL#144.
AXL release for SCR v4.0
See SCR v4.0 project: https://github.com/orgs/LLNL/projects/26 …View Comment
hppritcha closed issue spack/spack#42651.
Installation issue: OpenMPI@5.0.{1,2} incorrect pmix dependency version range
### Steps to reproduce the issue …View Comment
BradWhitlock closed issue LLNL/axom#1337.
Calling host function in host device function.
The following was reported on Teams. A host function is being called from a host_device function. In this case, the std::numeric_limits<>::max() function is being called and it is host-only. Calls like this could affect Axom on CUDA or HIP. …View Comment
adrienbernede commented on issue LLNL/RAJA#1644.
@long58 apparently Spack does not like your new compiler. I
jakubcerveny closed issue jakubcerveny/gilbert#11.
C implementation returns 0, 0 as an index for larger dimensions.
I think it is because of this line. …View Comment
jedbrown commented on issue RDycore/RDycore#185.
I think the question is which site will be maintained when requirements change. I didn’t think of the static site when a student was running into issues building. Shall we delete the build instructions in the repo and just link, or shall we move build-from-source docs from the static site to the repo?…
correaa commented on issue include-what-you-use/include-what-you-use#1553.
I see….
grondo open issue flux-framework/flux-sched#1222.
submitting a large number of jobs and canceling them caused scheduler "hang"
On elcap a user submitted a large number of jobs, each requesting a specific node, then thought better of it and canceled them all. For the past few hours the scheduler has not scheduled any jobs and the “alloc pending” count is stuck at many thousands. …View Comment
markcmiller86 open issue visit-dav/visit#19591.
Cannot parse xdmf file
### Describe the bug …View Comment
CRobeck commented on issue triton-lang/triton#4026.
Depended on: …
vsoch commented on issue spack/spack#44602.
Thanks @alecbcs!…
trws commented on issue spack/spack#44581.
None…
cyrush commented on issue spack/spack#44571.
These are the prefered major deps: …
adayton1 commented on issue serizba/cppflow#254.
Just remove the call to std::move. You already have an rvalue reference, so the move constructor will get called without an additional call to std::move….
hppritcha commented on issue open-mpi/ompi#12601.
Could you set this env variable in the shell where the parent process is started?…
v-dobrev commented on issue mfem/mfem#4346.
I think @jandrej is right – all your source files that use mfem::forall need to be marked as CUDA sources when MFEM_USE_CUDA is true. This can be done with something like this: https://github.com/mfem/mfem/blob/c3eb769a2a316a68fd9a00448e20ca7cff4c8b7c/config/cmake/modules/MfemCmakeUtilities.cmake#L87-L90
…
jeremylt commented on issue freeCodeCamp/freeCodeCamp#55118.
Yes, we should use the phrase “single space” in all three spots…
vsoch closed issue flux-framework/spack#193.
[package-update] for flux-sched: update-package/flux-sched-2024-06-07
This is a request to open a pull request for a package update….View Comment
trws commented on issue flux-framework/flux-sched#1218.
Now that the release is together, I rebased the reformat here. Short of adding CI to verify it, which might accidentally work because it’s in the format script now that I think about it, this should be good to go….
grondo open issue flux-framework/flux-core#6038.
flux-start silently ignores `--recovery` when `-s, --test-size` is also present
``` …View Comment
grondo commented on issue flux-framework/flux-accounting#455.
Yeah, I guess if a job’s age is reset any time the job is modified then that is fair and wouldn’t require recalculating the priority for all other jobs at the time of that job’s priority update, since it would by definition have the smallest age. …
cmoussa1 commented on issue flux-framework/flux-accounting#455.
I think I see what you’re saying - sorry for it taking me until now to understand.
correaa commented on issue boostorg/serialization#309.
> The source code for the isolated test is here: https://github.com/johnwason/boost_serialization_1_84_bug . It includes an actions workflow that does a parametric test of different boost versions….
benzwick commented on issue SlicerCBM/SlicerCBM#73.
This seems to have been caused by some recent changes to the location of the documentation (Slicer is looking for a module in the docs folder). …
adrienbernede commented on issue LLNL/Umpire#795.
@davidbeckingsale +tools currently conflicts with +rocm. Depending on what you want to achieve, you
adrienbernede commented on issue LLNL/RAJAPerf#452.
@rhornung67 Sounds good. …
artv3 closed issue LLNL/RAJA#1566.
Add expt::KernelNaming forall example
We are missing an example that exercises the kernel naming capability in RAJA::forall. At the moment I can’t seem to compile the example provided in the RAJA docs (https://raja.readthedocs.io/en/develop/sphinx/user_guide/feature/reduction.html#experimental-reduction-interface). …View Comment
jeremylt commented on issue CEED/libCEED#1597.
Most of the core pieces are there, need to debug the parallelism, esp in 2/3d…
jandrej commented on issue zed-industries/zed#12644.
Yea I think it’s fixed in the preview….
gardner48 commented on issue spack/spack#44601.
This should be fixed by LLNL/sundials#434 and LLNL/sundials#437 in the sundials develop branch which fixed build problems with ROCM 6.x…
v-dobrev commented on issue spack/spack#44010.
@cnpetra, @cameronrutherford, …
jandrej commented on issue mfem/mfem#4346.
MFEM requires all files to be compiled with nvcc. Are you sure that your project files that use mfem_forall are compiled with nvcc? What’s the output of make VERBOSE=1 of your project….
tzanio commented on issue mfem/mfem#4337.
Merged in next for testing……
hppritcha commented on issue grimme-lab/xtb-python#114.
some context. we are trying to use xtb and the xtb python bindings with examol and were building it using spack and hit this issue….
garlick open issue flux-framework/flux-sched#1221.
add ability to query current schedule
Problem: as mentioned in #1219, it would be handy to be able to dump a schedule snapshot, perhaps as json. …View Comment
milroy commented on issue flux-framework/flux-sched#1220.
> Anything else you want to make sure happens @milroy? …
garlick commented on issue flux-framework/flux-core#6011.
Oh,wait. I guess probably the version of systemd I have isn’t yet using cgroup.kill so systemd running as flux cant signal the shell running as garlick. The SIGTERMs are forwarded by the IMP, and the SIGKILLs take the IMP out. So back to the original idea of having the IMP iterate over the cgroup. Sorry for the noise!…
chu11 commented on issue flux-framework/flux-core#6035.
I thought https://github.com/flux-framework/flux-core/pull/6013 was going to handle this by “buffering” the stdin RPCs?…
balos1 commented on issue LLNL/sundials#500.
@bangerth Can you try setting the matvec operation on the empty matrix object created here? It can just be a no-op.
…
markcmiller86 commented on issue LLNL/Silo#378.
Ok, I figured out how to do it same with h5py…
…
cyrush commented on issue visit-dav/visit#19582.
@tharman-git can you try building with VTK-9 support? …
jedbrown commented on issue fish-shell/fish-shell#6931.
Thanks for making this issue. The solution isn’t quite equivalent. For example, typing pacman -S xf86- and triggering insert-completions yields pacman -S xf86-xf86-input-elographics xf86-input-evdev [...]. I’m newly tinkering with Fish so don’t know how to remove the stem before inserting the completions….
bgunnar5 commented on issue LLNL/merlin#481.
So it looks like the local run will pass with or without the patch. However, the distributed run does not so I think we have to keep this sleep business embedded in the test for now. I have another branch that’s working on updating the test suite so maybe once I work on that more we can use some of those changes to resolve this….
wihobbs open issue LLNL/maestrowf#444.
feature request: "prettier" error message for missing keys or empty YAML file
Hi folx! I was working with some new-to-HPC users in a tutorial this morning, and one of them observed that when a blank YAML or YAML missing values was passed to maestro run, you get an entire Python traceback:
…View Comment
cyrush open issue Alpine-DAV/ascent#1310.
compile issue when dray stats are off?
https://github.com/Alpine-DAV/ascent/blob/0158038eb468e5a2e470f63ae271b5297d880ce2/src/libs/dray/utils/appstats.cpp#L264…View Comment
brendankeith commented on issue mfem/web#266.
> is it ready-for-review?
…
artv3 commented on issue LLNL/RAJAPerf#442.
Can you add the SYCL version too? It should be straightforward. Perhaps also call this kernel “UNROLLED_MATVEC_3D” kernel? …
jandrej commented on issue EnzymeAD/Enzyme#1904.
@GregTheMadMonk did you check the output values? I only get zeros from the fwddiff in your branch….
markcmiller86 commented on issue visit-dav/visit#19565.
Summarizing… …
correaa commented on issue seanbaxter/circle#118.
Okay, I think the problem is that Circle was the only compiler reaching the end of the priority list for adl_uninitialized_copy. This, in turn, is invalid code, and since it is an auto function, Circle was probably very confused.
…