can we deprecate this project and stop working on it?

[edwardhartnett]

Now that spack-stack has been released, is there any reason to continue with hpc-stack?

[aerorahul]

Are there templates in place in spack-stack to install new versions of libraries or request new libraries on the systems?
Also, what is the ETA on supporting python-virtual environments and/or conda environments via spack-stack. Several of our systems use that functionality.

TIA.

[climbfuji]

For testing new packages and for using Python with venv or pip with spack-stack - see Section 8.2:

https://spack-stack.readthedocs.io/en/spack-stack-1.0.0/MaintainersSection.html#testing-new-packages

[BrianCurtis-NOAA]

We should continue with hpc-stack until spack-stack is installed and implemented into the various projects that use those packages/modules. For example i can see a need for a new lib version arising before spack-stack can be integrated fully into UFSWM.

[climbfuji]

We also need to roll out spack stack on the various HPCs first. Right now, it's on discover and I on orion (building now, it was down the entire day and last night).

[aerorahul]

@climbfuji
An average user should not need to run any spack commands for creating/loading venv's or conda environments.
A user can provide the recipe for the virtual env. (or conda env), and the maintainers should provide a module to load that virtual env (or conda env).
I assume this capability will be restored in spack-stack.

[climbfuji]

I don’t see why a user runs a spack command, not according to the documentation.

…

On Jun 30, 2022, at 10:07 AM, Rahul Mahajan ***@***.***> wrote: @climbfuji <https://github.com/climbfuji> An average user should not need to run any spack commands for creating/loading venv's or conda environments. A user can provide the recipe for the virtual env. (or conda env), and the maintainers should provide a module to load that virtual env (or conda env). I assume this capability will be restored in spack-stack. — Reply to this email directly, view it on GitHub <#475 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5C2RNETTR37C7FQFCXQLTVRXA5PANCNFSM52JVLDOQ>. You are receiving this because you were mentioned.

[edwardhartnett]

I thought that spack-stack was already installed and working on all our machines. Did I misunderstand?

[aerorahul]

I don’t see why a user runs a spack command, not according to the documentation.
…
On Jun 30, 2022, at 10:07 AM, Rahul Mahajan @.***> wrote: @climbfuji https://github.com/climbfuji An average user should not need to run any spack commands for creating/loading venv's or conda environments. A user can provide the recipe for the virtual env. (or conda env), and the maintainers should provide a module to load that virtual env (or conda env). I assume this capability will be restored in spack-stack. — Reply to this email directly, view it on GitHub <#475 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB5C2RNETTR37C7FQFCXQLTVRXA5PANCNFSM52JVLDOQ. You are receiving this because you were mentioned.

May be we should discuss this offline.
Specifically, I am inquiring if there is ability in spack-stack to create virtual envs (or conda envs). The answer is likely yes. If so, a user should then just load the modulefile for that virtual env (or conda env). The user should not have to create the virtual env. individually.

[climbfuji]

Yes, let's take that offline. If you want, please find a time to chat next week, best to include Kyle as well as I don't want to give you wrong/incomplete information.

[climbfuji]

Did anyone read the documentation, the release notes, the issues? The slack messages?

…

On Jun 30, 2022, at 10:17 AM, Edward Hartnett ***@***.***> wrote: I thought that spack-stack was already installed and working on all our machines. Did I misunderstand — Reply to this email directly, view it on GitHub <#475 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5C2ROP2CNNLTNAZ75QITTVRXCCJANCNFSM52JVLDOQ>. You are receiving this because you were mentioned.

[edwardhartnett]

I have not. I am 100% focused on the GRIB libraries and that's where my focus has to stay for the next few months. ;-)

However I am happy that Alex is now here to help...

[aerorahul]

Are we any closer to deprecating this project?

[climbfuji]

I think we are, but who am I to judge this ;-) We would have liked to be done by now, obviously we aren't, but at least there is a full spack-stack solution available everywhere that should enable the UFS community figure out what works (hopefully almost everything) and what doesn't (hopefully very little) so that when we have release 1.4.0 of spack-stack end of May we can call it done. Of course, GSI issues w/ newer Intel will need to be addressed by then to call it "done done".

[jokervTv]

so what is the current status ？

[climbfuji]

We are an inch (well, one pull request) away! ufs-community/ufs-weather-model#1707

[DavidHuber-NOAA]

I should probably also mention NOAA-EMC/GSI#589.