[arch-general] dev-public: Future build infrastructure for Arch Linux
What does: Future build infrastructure for Arch Linux "We decided on supporting the x86_64-v3 microarchitecture" Does that mean Arch will break on anything older than Haswell for Intel will no longer be supported on Arch? Or am I missing something? SUSE going to ALP is doing the v3 baseline for it's "containerized" Micro Leap, but what does this mean for Arch. With SUSE, it means just that, if you have an i7 Sandy Lake or older -- your just not invited to the party. That seems completely counter to what Arch should do for a main distro. -- David C. Rankin, J.D.,P.E.
Please see the RFC which should have all the details. https://gitlab.archlinux.org/archlinux/rfcs/-/blob/master/rfcs/0002-march.rs... -- Morten Linderud PGP: 9C02FF419FECBE16
On 9/5/22 04:39, Morten Linderud via arch-general wrote:
Please see the RFC which should have all the details.
https://gitlab.archlinux.org/archlinux/rfcs/-/blob/master/rfcs/0002-march.rs...
Thank you, As long as it's a 2nd port -- that's fine, but I have a lot of v2 hardware, i7 Sandy Bridge, etc.. that it would be ridiculous to need a forced replacement of. As far as the rational "Arch used to pride itself in providing optimised binaries out of the box.", I can't see compilation to v3 being more than an immeasurable instruction or two different from a build for v2 in all but rare packages. It's not like Arch is only used by a minority of users who "may" benefit from new gcc11 introduced optimizations. Will there be any testing result comparisons posted anywhere on the improvements seen between the current build and v3 builds? -- David C. Rankin, J.D.,P.E.
On Mon, Sep 05, 2022 at 04:48:59AM -0500, David C. Rankin via arch-general wrote:
As far as the rational "Arch used to pride itself in providing optimised binaries out of the box.", I can't see compilation to v3 being more than an immeasurable instruction or two different from a build for v2 in all but rare packages. It's not like Arch is only used by a minority of users who "may" benefit from new gcc11 introduced optimizations.
The original RFC was to just up the baseline from v1 to v2, but this was rejected for a separate v3 port instead. See the discussion here, https://gitlab.archlinux.org/archlinux/rfcs/-/merge_requests/2 -- Morten Linderud PGP: 9C02FF419FECBE16
participants (2)
-
David C. Rankin
-
Morten Linderud