During the board meeting, the naming issue was re-raised; “x86 SIG” just isn't that great. So I'd like to propose “x86-64 SIG” instead, with a hyphen. We use “x86_64” in the RPM architecture name and configure triplets, but only because we must, as “-” is consindered a separator in these contexts. The official vendor-neutral architecture name is x86-64.
During the meeting, I was under the impression that the board was leaning towards a narrow scope, but that is not quite what the posted minutes reflect. Per Fabian's announcement, we have at least a bit of wiggle room for non-x86 ISA experiments in CBS (ThunderX2 has LSE atomics support). Personally, I'm not interested in such experiments at this time, though. But we could call the SIG “ISA SIG” to keep open the possibility for non-x86 work, if that's what people want.
Thoughts?
Thanks, Florian
On Fri, Apr 21, 2023 at 10:14 AM Florian Weimer fweimer@redhat.com wrote:
During the board meeting, the naming issue was re-raised; “x86 SIG” just isn't that great. So I'd like to propose “x86-64 SIG” instead, with a hyphen. We use “x86_64” in the RPM architecture name and configure triplets, but only because we must, as “-” is consindered a separator in these contexts. The official vendor-neutral architecture name is x86-64.
During the meeting, I was under the impression that the board was leaning towards a narrow scope, but that is not quite what the posted minutes reflect. Per Fabian's announcement, we have at least a bit of wiggle room for non-x86 ISA experiments in CBS (ThunderX2 has LSE atomics support). Personally, I'm not interested in such experiments at this time, though. But we could call the SIG “ISA SIG” to keep open the possibility for non-x86 work, if that's what people want.
Thoughts?
I'd be cool with calling it the ISA SIG. For now, we can constrain it to x86_64 stuff, but in the future, I could envision others looking at POWER and zSystems ISA bumps too, since those happen too.
-- 真実はいつも一つ！/ Always, there's only one truth!
On Fri, 2023-04-21 at 16:14 +0200, Florian Weimer wrote:
During the board meeting, the naming issue was re-raised; “x86 SIG” just isn't that great. So I'd like to propose “x86-64 SIG” instead, with a hyphen. We use “x86_64” in the RPM architecture name and configure triplets, but only because we must, as “-” is consindered a separator in these contexts. The official vendor-neutral architecture name is x86-64.
During the meeting, I was under the impression that the board was leaning towards a narrow scope, but that is not quite what the posted minutes reflect. Per Fabian's announcement, we have at least a bit of wiggle room for non-x86 ISA experiments in CBS (ThunderX2 has LSE atomics support). Personally, I'm not interested in such experiments at this time, though. But we could call the SIG “ISA SIG” to keep open the possibility for non-x86 work, if that's what people want.
Thoughts?
Thanks, Florian
Perhaps setting up "working groups" within the SIG would help clarify the current scope of work?
An ISA SIG with an x86-64 working group would clarify that no work is currently focused on any other arches, but leave open a door if other folks wanted to form some sort of s390x working group - but they'd be on their own?
Pat
And that was the intent of the discussions. We would start the SiG with the x86-64 work and then possibly rename it down the road to be more of an umbrella for folks wanting to do similar things. The intent was never for Florian to do anything but the x86 work.
If folks are good with the ISA SiG name we can just go with that from the start.
Amy
*Amy Marrich*
She/Her/Hers
Principal Technical Marketing Manager - Cloud Platforms
Red Hat, Inc https://www.redhat.com/
amy@redhat.com
Mobile: 954-818-0514
Slack: amarrich
IRC: spotz https://www.redhat.com/
On Fri, Apr 21, 2023 at 9:51 AM Patrick Riehecky via CentOS-devel < centos-devel@centos.org> wrote:
On Fri, 2023-04-21 at 16:14 +0200, Florian Weimer wrote:
During the board meeting, the naming issue was re-raised; “x86 SIG” just isn't that great. So I'd like to propose “x86-64 SIG” instead, with a hyphen. We use “x86_64” in the RPM architecture name and configure triplets, but only because we must, as “-” is consindered a separator in these contexts. The official vendor-neutral architecture name is x86-64.
During the meeting, I was under the impression that the board was leaning towards a narrow scope, but that is not quite what the posted minutes reflect. Per Fabian's announcement, we have at least a bit of wiggle room for non-x86 ISA experiments in CBS (ThunderX2 has LSE atomics support). Personally, I'm not interested in such experiments at this time, though. But we could call the SIG “ISA SIG” to keep open the possibility for non-x86 work, if that's what people want.
Thoughts?
Thanks, Florian
Perhaps setting up "working groups" within the SIG would help clarify the current scope of work?
An ISA SIG with an x86-64 working group would clarify that no work is currently focused on any other arches, but leave open a door if other folks wanted to form some sort of s390x working group - but they'd be on their own?
Pat _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
On Fri, Apr 21, 2023 at 12:00 PM Amy Marrich amy@redhat.com wrote:
And that was the intent of the discussions. We would start the SiG with the x86-64 work and then possibly rename it down the road to be more of an umbrella for folks wanting to do similar things. The intent was never for Florian to do anything but the x86 work.
If folks are good with the ISA SiG name we can just go with that from the start.
This seems like a good compromise.
Florian, I'll find whatever documentation that says we need to open a Board ticket for the SIG this week and get that moving with the ISA SIG name.
josh
On Fri, Apr 21, 2023 at 9:51 AM Patrick Riehecky via CentOS-devel centos-devel@centos.org wrote:
On Fri, 2023-04-21 at 16:14 +0200, Florian Weimer wrote:
During the board meeting, the naming issue was re-raised; “x86 SIG” just isn't that great. So I'd like to propose “x86-64 SIG” instead, with a hyphen. We use “x86_64” in the RPM architecture name and configure triplets, but only because we must, as “-” is consindered a separator in these contexts. The official vendor-neutral architecture name is x86-64.
During the meeting, I was under the impression that the board was leaning towards a narrow scope, but that is not quite what the posted minutes reflect. Per Fabian's announcement, we have at least a bit of wiggle room for non-x86 ISA experiments in CBS (ThunderX2 has LSE atomics support). Personally, I'm not interested in such experiments at this time, though. But we could call the SIG “ISA SIG” to keep open the possibility for non-x86 work, if that's what people want.
Thoughts?
Thanks, Florian
Perhaps setting up "working groups" within the SIG would help clarify the current scope of work?
An ISA SIG with an x86-64 working group would clarify that no work is currently focused on any other arches, but leave open a door if other folks wanted to form some sort of s390x working group - but they'd be on their own?
Pat _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
CentOS-devel mailing list CentOS-devel@centos.org https://lists.centos.org/mailman/listinfo/centos-devel
On Tue, Apr 25, 2023 at 9:28 AM Josh Boyer jwboyer@redhat.com wrote:
On Fri, Apr 21, 2023 at 12:00 PM Amy Marrich amy@redhat.com wrote:
And that was the intent of the discussions. We would start the SiG with the x86-64 work and then possibly rename it down the road to be more of an umbrella for folks wanting to do similar things. The intent was never for Florian to do anything but the x86 work.
If folks are good with the ISA SiG name we can just go with that from the start.
This seems like a good compromise.
Florian, I'll find whatever documentation that says we need to open a Board ticket for the SIG this week and get that moving with the ISA SIG name.
I have the board ticket open now: https://git.centos.org/centos/board/issue/115
Thanks again for all who have contributed to the discussion so far.
josh
Hello everyone, we would like to participate to ISA SIG.
Cheers
On 28/04/2023 15:51, Josh Boyer wrote:
On Tue, Apr 25, 2023 at 9:28 AM Josh Boyer jwboyer@redhat.com wrote:
On Fri, Apr 21, 2023 at 12:00 PM Amy Marrich amy@redhat.com wrote:
And that was the intent of the discussions. We would start the SiG with the x86-64 work and then possibly rename it down the road to be more of an umbrella for folks wanting to do similar things. The intent was never for Florian to do anything but the x86 work.
If folks are good with the ISA SiG name we can just go with that from the start.
This seems like a good compromise.
Florian, I'll find whatever documentation that says we need to open a Board ticket for the SIG this week and get that moving with the ISA SIG name.
I have the board ticket open now: https://git.centos.org/centos/board/issue/115
Thanks again for all who have contributed to the discussion so far.
* Ali Erdinc Koroglu:
Hello everyone, we would like to participate to ISA SIG.
Ali,
Thank you. I look forward to working with you and your colleagues at Intel in the new CentOS ISA SIG.
In semi-related news, we have begun the process of evaluating whether we can merge your glibc performance backports for the Hyperscaler SIG into CentOS 8 Stream.
Thanks, Florian
-
Ali Erdinc Koroglu -
Amy Marrich -
Florian Weimer -
Josh Boyer -
Neal Gompa -
Patrick Riehecky