Make credit2 the default scheduler
PATCH 1/3 Make credit2 the default scheduler
Credit2 was declared "supported" in 4.8, and as of 4.10 had two other critical features implemented (soft affinity / NUMA and caps).
---
CC: Ian Jackson <[email protected]>
CC: Wei Liu <[email protected]>
CC: Andrew Cooper <[email protected]>
CC: Jan Beulich <[email protected]>
CC: Tim Deegan <[email protected]>
CC: Stefano Stabellini <[email protected]>
CC: Konrad Wilk <[email protected]>
CC: Dario Faggioli <[email protected]>
--- xen/common/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
xen/common/Kconfig +2 -2
@@ -200,7 +200,7 @@ config SCHED_NULL choice prompt "Default Scheduler?" - default SCHED_CREDIT_DEFAULT + default SCHED_CREDIT2_DEFAULT config SCHED_CREDIT_DEFAULT bool "Credit Scheduler" if SCHED_CREDIT
@@ -221,7 +221,7 @@ config SCHED_DEFAULT default "rtds" if SCHED_RTDS_DEFAULT default "arinc653" if SCHED_ARINC653_DEFAULT default "null" if SCHED_NULL_DEFAULT - default "credit" + default "credit2" endmenu
Signature · 7 lines
2.15.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH 2/3 xen: Fix credit1 Kconfig entry
...so that it shows up in the menu and can be disabled.
---
CC: Ian Jackson <[email protected]>
CC: Wei Liu <[email protected]>
CC: Andrew Cooper <[email protected]>
CC: Jan Beulich <[email protected]>
CC: Tim Deegan <[email protected]>
CC: Stefano Stabellini <[email protected]>
CC: Konrad Wilk <[email protected]>
--- xen/common/Kconfig | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
xen/common/Kconfig +2 -1
@@ -164,7 +164,8 @@ menu "Schedulers" visible if EXPERT = "y" config SCHED_CREDIT - def_bool y + bool "Credit scheduler support" + default y ---help--- The traditional credit scheduler is a general purpose scheduler.
Signature · 7 lines
2.15.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH 2/3] xen: Fix credit1 Kconfig entry
Quoted 4 lines
Quoted 1 line
Quoted 1 line · On 05.02.18 at 12:36, <[email protected]> wrote:
...so that it shows up in the menu and can be disabled.
Signed-off-by George Dunlap <[email protected]>
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Dario Faggioli
Reviewed-byRe: [PATCH 2/3] xen: Fix credit1 Kconfig entry
Quoted 8 lines · Jan Beulich
Quoted 5 lines
Quoted 1 line
Quoted 1 line · On 05.02.18 at 12:36, <[email protected]> wrote:
...so that it shows up in the menu and can be disabled.
Signed-off-by George Dunlap <[email protected]>Acked-by Jan Beulich <[email protected]>
Regards,
Dario
Signature · 4 lines
<<This happens because I choose it to happen!>> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Software Engineer @ SUSE https://www.suse.com/
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH 3/3 xen: Disable ARINC653 scheduler by default for non-DEBUG builds
In reply to George DunlapThe ARINC653 scheduler is targeted at a very specific niche; typical users cannot benefit from using it. Disable it by default for non-DEBUG builds. (Enable it for DEBUG builds so that we catch any build breakages sooner rather than later.)
---
CC: Ian Jackson <[email protected]>
CC: Wei Liu <[email protected]>
CC: Andrew Cooper <[email protected]>
CC: Jan Beulich <[email protected]>
CC: Tim Deegan <[email protected]>
CC: Stefano Stabellini <[email protected]>
CC: Konrad Wilk <[email protected]>
CC: Josh Whitehead <[email protected]>
CC: Robert VanVossen <[email protected]>
--- xen/common/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
xen/common/Kconfig +1 -1
@@ -186,7 +186,7 @@ config SCHED_RTDS config SCHED_ARINC653 bool "ARINC653 scheduler support (EXPERIMENTAL)" - default y + default DEBUG ---help--- The ARINC653 scheduler is a hard real-time scheduler for single cores, targeted for avionics, drones, and medical devices.
Signature · 7 lines
2.15.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH 3/3] xen: Disable ARINC653 scheduler by default for non-DEBUG builds
Quoted 4 lines
Quoted 1 line
Quoted 1 line · On 05.02.18 at 12:36, <[email protected]> wrote:
The ARINC653 scheduler is targeted at a very specific niche; typical users cannot benefit from using it. Disable it by default for non-DEBUG builds.
The more that its prompt says "(EXPERIMENTAL)".
Quoted 1 line
Signed-off-by George Dunlap <[email protected]>
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH 3/3] xen: Disable ARINC653 scheduler by default for non-DEBUG builds
Quoted 6 lines · Jan Beulich
Quoted 4 lines
Quoted 1 line
Quoted 1 line · On 05.02.18 at 12:36, <[email protected]> wrote:
The ARINC653 scheduler is targeted at a very specific niche; typical users cannot benefit from using it. Disable it by default for non-DEBUG builds.
The more that its prompt says "(EXPERIMENTAL)".
Well I left RTDS on by default, since there would be a small chance someone might want to play with it. But AFAICT ARINC653 isn't really something designed to be used by anyone outside a very specific set of users.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Dario Faggioli
Reviewed-byRe: [PATCH 3/3] xen: Disable ARINC653 scheduler by default for non-DEBUG builds
In reply to Jan BeulichQuoted 7 lines · Jan Beulich
Quoted 6 lines
Quoted 1 line
Quoted 1 line · On 05.02.18 at 12:36, <[email protected]> wrote:
The ARINC653 scheduler is targeted at a very specific niche; typical
users cannot benefit from using it. Disable it by default for non-DEBUG builds.
Quoted 6 lines
The more that its prompt says "(EXPERIMENTAL)".
Quoted 1 line
Signed-off-by George Dunlap <[email protected]>Acked-by Jan Beulich <[email protected]>
Regards,
Dario
Signature · 4 lines
<<This happens because I choose it to happen!>> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Software Engineer @ SUSE https://www.suse.com/
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH 3/3] xen: Disable ARINC653 scheduler by default for non-DEBUG builds
Quoted 8 lines · Dario Faggioli
Quoted 7 lines · Jan Beulich
Quoted 6 lines
Quoted 1 line
Quoted 1 line · On 05.02.18 at 12:36, <[email protected]> wrote:
The ARINC653 scheduler is targeted at a very specific niche; typical
users cannot benefit from using it. Disable it by default for non-DEBUG builds.
Quoted 7 lines
Quoted 6 lines
The more that its prompt says "(EXPERIMENTAL)".
Quoted 1 line
Signed-off-by George Dunlap <[email protected]>Acked-by Jan Beulich <[email protected]>Reviewed-by Dario Faggioli <[email protected]>
Just to let people know, I was planning on waiting at least a few days for the maintainers of this scheduler to express an opinion before checking it in.
-Georeg
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 4 lines · George Dunlap
Credit2 was declared "supported" in 4.8, and as of 4.10 had two other critical features implemented (soft affinity / NUMA and caps).
Signed-off-by George Dunlap <[email protected]>
On what justification?
Switching the default involves demonstrating that the new default is better than the old one, and simply "we've implemented more features" isn't good enough IMO.
The last set of measurements I recall seeing said that credit1 and credit2 were mostly kneck-and-kneck, but there was still a noticeable difference in credit1's favour when it came to aggregate small-packet TCP throughput, and no benchmark where credit2 was clearly a winner.
~Andrew
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 11 lines · Andrew Cooper
Quoted 4 lines · George Dunlap
Credit2 was declared "supported" in 4.8, and as of 4.10 had two other critical features implemented (soft affinity / NUMA and caps).
Signed-off-by George Dunlap <[email protected]>On what justification?
Switching the default involves demonstrating that the new default is better than the old one, and simply "we've implemented more features" isn't good enough IMO.
Well the code is certainly better: more predictable, less jitter, easier to determine how modifications will affect overall behavior, easier in the future to make load-balancing behavior more subtle (e.g., taking into account the cost of powering up extra cores, &c).
Quoted 4 lines
The last set of measurements I recall seeing said that credit1 and credit2 were mostly kneck-and-kneck, but there was still a noticeable difference in credit1's favour when it came to aggregate small-packet TCP throughput, and no benchmark where credit2 was clearly a winner.
It's somewhat of a mixed bag. Unfortunately most of what I have are tests using XenServer's internal perf testing system, so I can't share the raw data (via links anyway).
Here is a summary of data from an intern e-mail Dario sent about 6 months ago:
* DVDbench: On underloaded systems, credit2 outperformed credit1 by about 4%. On overloaded systems, credit2 underperformed by about 3%.
* On a range of tests (unixbench, lmbench, &c), credit and credit2 perform within 5% of each other (up and down).
* Credit2 fairly consistently beats credit for TCP-style workloads.
* Credit2 is sometimes equal to, sometimes 5-15% worse than, credit for synthetic CPU workloads (e.g., Dhrystone).
* On LoginVSI, credit2 fairly consistently outperforms credit by about 10%.
Credit2, like credit, has a number of workloads / setups for which performance could be improved. Personally I think networking and partially-loaded systems is going to be more representative of what Xen is actually used for; so I think credit2 is on the whole the better scheduler to use by default. And in any case, making those improvements on credit2 will be easier than on credit.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 11 lines · George Dunlap
Quoted 1 line · Andrew Cooper
Here is a summary of data from an intern e-mail Dario sent about 6 months ago:
* DVDbench: On underloaded systems, credit2 outperformed credit1 by about 4%. On overloaded systems, credit2 underperformed by about 3%.
* On a range of tests (unixbench, lmbench, &c), credit and credit2 perform within 5% of each other (up and down).
Quoted 9 lines
* Credit2 fairly consistently beats credit for TCP-style workloads.
* Credit2 is sometimes equal to, sometimes 5-15% worse than, credit for
synthetic CPU workloads (e.g., Dhrystone).* On LoginVSI, credit2 fairly consistently outperforms credit by about 10%.
Thanks George for summing this up!
Quoted 7 lines
Credit2, like credit, has a number of workloads / setups for which performance could be improved. Personally I think networking and partially-loaded systems is going to be more representative of what Xen
is actually used for; so I think credit2 is on the whole the better scheduler to use by default.
I'm sure this will sound no surprise for anyone, but allow me to say it
anyway: I agree with George. :-D
Quoted 3 lines
And in any case, making those improvements
on credit2 will be easier than on credit.
And, if possible, I agree with George on this even more!
One thing I think we should consider, though, is that we've often said we would switch at the very beginning of a dev cycle, to get as much as osstest and day-by-day testing from developer as possible.
Considering we're releasing in June, but freezing in March, do we think
it is still early enough?
Regards,
Dario
Signature · 4 lines
<<This happens because I choose it to happen!>> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Software Engineer @ SUSE https://www.suse.com/
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 4 lines · Dario Faggioli
Quoted 3 lines · George Dunlap
And in any case, making those improvements
on credit2 will be easier than on credit.
Quoted 8 lines
And, if possible, I agree with George on this even more!
One thing I think we should consider, though, is that we've often said we would switch at the very beginning of a dev cycle, to get as much as osstest and day-by-day testing from developer as possible.
Considering we're releasing in June, but freezing in March, do we think
it is still early enough?
The 4.11 release is completely dominated by Meltdown/Spectre mitigation work. So either 4.11 will be a security focused version or we need to extend the development phase.
I think we should _really_ discuss release schedules rather sooner than later.
Juergen
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 11 lines · Juergen Gross
Quoted 10 lines · Dario Faggioli
Quoted 3 lines · George Dunlap
And in any case, making those improvements
on credit2 will be easier than on credit.And, if possible, I agree with George on this even more!
One thing I think we should consider, though, is that we've often said we would switch at the very beginning of a dev cycle, to get as much as osstest and day-by-day testing from developer as possible.
Quoted 6 lines
Quoted 2 lines
Considering we're releasing in June, but freezing in March, do we think
it is still early enough?The 4.11 release is completely dominated by Meltdown/Spectre mitigation work. So either 4.11 will be a security focused version or we need to extend the development phase.
Personally I could go either way on this. So unless someone wants to argue for switching now (or we significantly extend the development window), let's plan on leaving this for post-4.11.
Quoted 2 lines
I think we should _really_ discuss release schedules rather sooner than later.
Sorry I wasn't much involved in the earlier discussion -- I was rather heads-down trying to figure out what to do about speculative execution. :-) Let me see if I can go find that thread.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 13 lines · George Dunlap
Quoted 11 lines · Juergen Gross
Quoted 4 lines · Dario Faggioli
Considering we're releasing in June, but freezing in March, do we think
it is still early enough?The 4.11 release is completely dominated by Meltdown/Spectre mitigation
work. So either 4.11 will be a security focused version or we need to
extend the development phase.
Quoted 4 lines
Personally I could go either way on this. So unless someone wants to argue for switching now (or we significantly extend the development window), let's plan on leaving this for post-4.11.
Yes, I agree.
And if we don't switch now, I think we should say that, unless someone argues against your reasoning in your other email, and convince us, we'll switch as soon as 4.12 is branched (or perhaps as soon as 4.11 is released?).
Regards,
Dario
Signature · 4 lines
<<This happens because I choose it to happen!>> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Software Engineer @ SUSE https://www.suse.com/
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 18 lines · Dario Faggioli
Quoted 13 lines · George Dunlap
Quoted 11 lines · Juergen Gross
Quoted 4 lines · Dario Faggioli
Considering we're releasing in June, but freezing in March, do we think
it is still early enough?The 4.11 release is completely dominated by Meltdown/Spectre mitigation
work. So either 4.11 will be a security focused version or we need to
extend the development phase.Quoted 4 lines
Personally I could go either way on this. So unless someone wants to argue for switching now (or we significantly extend the development window), let's plan on leaving this for post-4.11.
Quoted 6 lines
Yes, I agree.
And if we don't switch now, I think we should say that, unless someone argues against your reasoning in your other email, and convince us, we'll switch as soon as 4.12 is branched (or perhaps as soon as 4.11 is released?).
Well I had planned to do that for 4.11, but something came up. :-) Hopefully nothing similar will come up this time...
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel