SRU review: Ok. From digging around I see that in one benchmark there's
a 1% performance decrease measured on ARMv8.0 processors, which seems
like an acceptable tradeoff for the very significant performance
increase on ARMv8.1 processors¹.
Since we *have* this very nice performance test, could we *also* run it
on an ARMv8.0 system as a part of validation? That would give us the
relevant performance penalty for v8.0 systems - which would be good to
verify is not significant - and also check that the code *works* on v8.0
systems!
So, in summary: this looks like an acceptable SRU, as long as we test on
both an ARMv8.0 system and an ARMv8.1 system, and the v8.0 system does
not significantly regress performance (around 1% seems ok, much more is
iffy).
¹: Although it would be *nice* to have some sort of data as to how
prevalent these chips are - is this a big win for almost everyone, with
only a few chips that don't support v8.1, or a big win only on uncommon
chips?
** Changed in: nginx (Ubuntu Focal)
Status: In Progress => Incomplete
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2024019
Title:
Add GCC atomic support (-moutline-atomics) for arm64 on Focal
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/haproxy/+bug/2024019/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
