Installation issue:ginkgo dependency exhausts system memory (OOM Killer)

[Quntized]

Steps to reproduce the issue

The build fails during the installation of ddc when its dependency ginkgo is being compiled. The full spec is very long, but the command to generate it is:

$ spack spec -I ddc

The exact steps to reproduce the failure from a clean Spack installation are:

1. Install and activate Spack 1.0.1

wget https://github.com/spack/spack/releases/download/v1.0.1/spack-1.0.1.tar.gz
tar -xvf spack-1.0.1.tar.gz
. spack-1.0.1/share/spack/setup-env.sh

2. Create a dedicated environment

spack env create ddc-env

3. Pin the repository to the recommended commit for recipe compatibility

spack --env ddc-env repo update builtin --commit 633ec00

4. Add ddc to the environment

spack --env ddc-env add ddc

5. Attempt to install (this is the command that fails)

spack --env ddc-env install

@tpadioleau
@tretre91

Error message

The final error message reported by Spack is:

==> Error: ddc-0.8.0-ugud5fchweroodm3elrlnhv6aof4kvi5: Package was not installed
==> Error: Installation request failed. Refer to reported errors for failing package(s).

The root cause is not a compilation error in the code itself, but rather the system killing the compiler processes due to memory exhaustion.

Information on your system

Spack Version: 1.0.1

OS: Linux (Arch: linux-Ubuntu)

Compiler: GCC 13.3.0

Additional information

The installation succeeds if the number of parallel build jobs is manually limited using the -j flag, which prevents memory exhaustion:

This works successfully

$ spack --env ddc-env install -j 2

General information

• I have run spack debug report and reported the version of Spack/Python/Platform
• I have run spack maintainers <name-of-the-package> and @mentioned any maintainers
• I have uploaded the build log and environment files
• I have searched the issues of this repo and believe this is not a duplicate

[tpadioleau]

Hello @Quntized, thank you for this report. As a maintainer of the DDC recipe I don't know if I can be of any help, maybe we could add a warning in the documentation for new users.

Can you tell the number of cores and how much memory do you have in your computer ?

[tretre91]

I also ran into the same issues, on a laptop with 12 cores and 16G of ram. I could go around it by limiting the number of cores to 4 (with -j 4). You can also try to combine this with -p 1 to install one package at a time.
But I don't think we can do something about this directly from the ddc recipe.