Skip to content
This repository was archived by the owner on Aug 28, 2025. It is now read-only.

Upgrade to GHC-9.4 resulting in lower compile-time memory usage#148

Draft
basvandijk wants to merge 15 commits into
masterfrom
basvandijk/ghc-9.4
Draft

Upgrade to GHC-9.4 resulting in lower compile-time memory usage#148
basvandijk wants to merge 15 commits into
masterfrom
basvandijk/ghc-9.4

Conversation

@basvandijk

@basvandijk basvandijk commented Feb 28, 2023

Copy link
Copy Markdown
Contributor

Building ic-hs with GHC-9.0 requires ~13GB of memory. Upgrading to GHC-9.4 reduces this by around 30% to ~9GB:

Screenshot 2023-02-28 at 17 20 14

mraszyk
mraszyk previously approved these changes Feb 28, 2023

@mraszyk mraszyk left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

Comment thread .github/workflows/release.yml Outdated
Comment thread .github/workflows/release.yml Outdated
Comment thread .github/workflows/release.yml Outdated
Comment thread .github/workflows/release.yml Outdated
Comment thread .github/workflows/release.yml Outdated
@ggreif

ggreif commented Mar 4, 2023

Copy link
Copy Markdown
Contributor

This is nice, but we need a coherent plan how to promote such changes downstream. E.g. dfinity/motoko will need the same GHC so that we can reuse the binary cache artefacts. Hopefully nix will take care of this.

@nomeata

nomeata commented Mar 4, 2023

Copy link
Copy Markdown
Contributor

I thought dfinity/motoko imports this repo as a whole, so for ic-ref-test it uses the same nixpkgs pin and thus the cache can be shared?

@ggreif

ggreif commented Mar 4, 2023

Copy link
Copy Markdown
Contributor

Hopefully. But we also have a majority of M1 Macs in the meantime which have to build locally. So that is one difference.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants