On Thu, Jul 20, 2017 at 04:19:40PM +0200, Peter Zijlstra wrote:
> On Thu, Jul 20, 2017 at 05:50:54AM -0700, Paul E. McKenney wrote:
> > >
> > > static void cpuidle_idle_call()
> > > {
> > > rcu_idle_enter()
> > > ..
> > > rcu_idle_exit()
> > > }
> > >
> > > I want
> > >
> > > static
On Thu, Jul 20, 2017 at 04:19:40PM +0200, Peter Zijlstra wrote:
> On Thu, Jul 20, 2017 at 05:50:54AM -0700, Paul E. McKenney wrote:
> > >
> > > static void cpuidle_idle_call()
> > > {
> > > rcu_idle_enter()
> > > ..
> > > rcu_idle_exit()
> > > }
> > >
> > > I want
> > >
> > > static
On Thu, Jul 20, 2017 at 05:50:54AM -0700, Paul E. McKenney wrote:
> >
> > static void cpuidle_idle_call()
> > {
> > rcu_idle_enter()
> > ..
> > rcu_idle_exit()
> > }
> >
> > I want
> >
> > static void cpuidle_idle_call()
> > {
> > if (tick stopped)
> >
On Thu, Jul 20, 2017 at 05:50:54AM -0700, Paul E. McKenney wrote:
> >
> > static void cpuidle_idle_call()
> > {
> > rcu_idle_enter()
> > ..
> > rcu_idle_exit()
> > }
> >
> > I want
> >
> > static void cpuidle_idle_call()
> > {
> > if (tick stopped)
> >
On 7/20/2017 1:11 AM, Thomas Gleixner wrote:
On Thu, 20 Jul 2017, Li, Aubrey wrote:
Don't get me wrong, even if a fast path is acceptable, we still need to
figure out if the coming idle is short and when to switch. I'm just worried
about if irq timings is not an ideal statistics, we have to
On 7/20/2017 1:11 AM, Thomas Gleixner wrote:
On Thu, 20 Jul 2017, Li, Aubrey wrote:
Don't get me wrong, even if a fast path is acceptable, we still need to
figure out if the coming idle is short and when to switch. I'm just worried
about if irq timings is not an ideal statistics, we have to
On 7/20/2017 5:50 AM, Paul E. McKenney wrote:
To make this work reasonably, you would also need some way to check for
the case where the prediction idle time is short but the real idle time
is very long.
so the case where you predict very short but is actually "indefinite", the real
solution
On 7/20/2017 5:50 AM, Paul E. McKenney wrote:
To make this work reasonably, you would also need some way to check for
the case where the prediction idle time is short but the real idle time
is very long.
so the case where you predict very short but is actually "indefinite", the real
solution
On Thu, Jul 20, 2017 at 09:40:49AM +0800, Li, Aubrey wrote:
> On 2017/7/19 22:48, Paul E. McKenney wrote:
> > On Wed, Jul 19, 2017 at 01:44:06PM +0800, Li, Aubrey wrote:
> >> On 2017/7/18 23:20, Paul E. McKenney wrote:
> >>
> 2) for rcu idle enter/exit, I measured the details which Paul
On Thu, Jul 20, 2017 at 09:40:49AM +0800, Li, Aubrey wrote:
> On 2017/7/19 22:48, Paul E. McKenney wrote:
> > On Wed, Jul 19, 2017 at 01:44:06PM +0800, Li, Aubrey wrote:
> >> On 2017/7/18 23:20, Paul E. McKenney wrote:
> >>
> 2) for rcu idle enter/exit, I measured the details which Paul
On Thu, 20 Jul 2017, Li, Aubrey wrote:
> Don't get me wrong, even if a fast path is acceptable, we still need to
> figure out if the coming idle is short and when to switch. I'm just worried
> about if irq timings is not an ideal statistics, we have to skip it too.
There is no ideal solution
On Thu, 20 Jul 2017, Li, Aubrey wrote:
> Don't get me wrong, even if a fast path is acceptable, we still need to
> figure out if the coming idle is short and when to switch. I'm just worried
> about if irq timings is not an ideal statistics, we have to skip it too.
There is no ideal solution
On 2017/7/19 15:55, Thomas Gleixner wrote:
> On Wed, 19 Jul 2017, Li, Aubrey wrote:
>> On 2017/7/18 15:19, Thomas Gleixner wrote:
>>> You can very well avoid it by taking the irq timings or whatever other
>>> information into account for the NOHZ decision.
>>>
>> If I read the source correctly,
On 2017/7/19 15:55, Thomas Gleixner wrote:
> On Wed, 19 Jul 2017, Li, Aubrey wrote:
>> On 2017/7/18 15:19, Thomas Gleixner wrote:
>>> You can very well avoid it by taking the irq timings or whatever other
>>> information into account for the NOHZ decision.
>>>
>> If I read the source correctly,
On 2017/7/19 22:48, Paul E. McKenney wrote:
> On Wed, Jul 19, 2017 at 01:44:06PM +0800, Li, Aubrey wrote:
>> On 2017/7/18 23:20, Paul E. McKenney wrote:
>>
2) for rcu idle enter/exit, I measured the details which Paul provided, and
the result matches with what I have measured before,
On 2017/7/19 22:48, Paul E. McKenney wrote:
> On Wed, Jul 19, 2017 at 01:44:06PM +0800, Li, Aubrey wrote:
>> On 2017/7/18 23:20, Paul E. McKenney wrote:
>>
2) for rcu idle enter/exit, I measured the details which Paul provided, and
the result matches with what I have measured before,
On Wed, Jul 19, 2017 at 10:03:22AM -0500, Christopher Lameter wrote:
> On Wed, 19 Jul 2017, Paul E. McKenney wrote:
>
> > > Do we have any problem if we skip RCU idle enter/exit under a fast idle
> > > scenario?
> > > My understanding is, if tick is not stopped, then we don't need inform
> > >
On Wed, Jul 19, 2017 at 10:03:22AM -0500, Christopher Lameter wrote:
> On Wed, 19 Jul 2017, Paul E. McKenney wrote:
>
> > > Do we have any problem if we skip RCU idle enter/exit under a fast idle
> > > scenario?
> > > My understanding is, if tick is not stopped, then we don't need inform
> > >
On Wed, 19 Jul 2017, Paul E. McKenney wrote:
> > Do we have any problem if we skip RCU idle enter/exit under a fast idle
> > scenario?
> > My understanding is, if tick is not stopped, then we don't need inform RCU
> > in
> > idle path, it can be informed in irq exit.
>
> Indeed, the problem
On Wed, 19 Jul 2017, Paul E. McKenney wrote:
> > Do we have any problem if we skip RCU idle enter/exit under a fast idle
> > scenario?
> > My understanding is, if tick is not stopped, then we don't need inform RCU
> > in
> > idle path, it can be informed in irq exit.
>
> Indeed, the problem
On Wed, Jul 19, 2017 at 03:43:07PM +0200, Frederic Weisbecker wrote:
> On Wed, Jul 12, 2017 at 08:56:51AM -0700, Paul E. McKenney wrote:
> > On Wed, Jul 12, 2017 at 01:54:51PM +0200, Peter Zijlstra wrote:
> > > On Tue, Jul 11, 2017 at 11:09:31AM -0700, Paul E. McKenney wrote:
> > > > On Tue, Jul
On Wed, Jul 19, 2017 at 03:43:07PM +0200, Frederic Weisbecker wrote:
> On Wed, Jul 12, 2017 at 08:56:51AM -0700, Paul E. McKenney wrote:
> > On Wed, Jul 12, 2017 at 01:54:51PM +0200, Peter Zijlstra wrote:
> > > On Tue, Jul 11, 2017 at 11:09:31AM -0700, Paul E. McKenney wrote:
> > > > On Tue, Jul
On Wed, Jul 19, 2017 at 01:44:06PM +0800, Li, Aubrey wrote:
> On 2017/7/18 23:20, Paul E. McKenney wrote:
>
> >> 2) for rcu idle enter/exit, I measured the details which Paul provided, and
> >> the result matches with what I have measured before, nothing notable found.
> >> But it still makes
On Wed, Jul 19, 2017 at 01:44:06PM +0800, Li, Aubrey wrote:
> On 2017/7/18 23:20, Paul E. McKenney wrote:
>
> >> 2) for rcu idle enter/exit, I measured the details which Paul provided, and
> >> the result matches with what I have measured before, nothing notable found.
> >> But it still makes
On Tue, Jul 18, 2017 at 11:14:57AM +0800, Li, Aubrey wrote:
> On 2017/7/18 3:23, Peter Zijlstra wrote:
> > On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
> >>> And as said; Daniel has been working on a better predictor -- now he's
> >>> probably not used it on the network workload
On Tue, Jul 18, 2017 at 11:14:57AM +0800, Li, Aubrey wrote:
> On 2017/7/18 3:23, Peter Zijlstra wrote:
> > On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
> >>> And as said; Daniel has been working on a better predictor -- now he's
> >>> probably not used it on the network workload
On Wed, Jul 12, 2017 at 08:56:51AM -0700, Paul E. McKenney wrote:
> On Wed, Jul 12, 2017 at 01:54:51PM +0200, Peter Zijlstra wrote:
> > On Tue, Jul 11, 2017 at 11:09:31AM -0700, Paul E. McKenney wrote:
> > > On Tue, Jul 11, 2017 at 06:34:22PM +0200, Peter Zijlstra wrote:
> >
> > > > But I think
On Wed, Jul 12, 2017 at 08:56:51AM -0700, Paul E. McKenney wrote:
> On Wed, Jul 12, 2017 at 01:54:51PM +0200, Peter Zijlstra wrote:
> > On Tue, Jul 11, 2017 at 11:09:31AM -0700, Paul E. McKenney wrote:
> > > On Tue, Jul 11, 2017 at 06:34:22PM +0200, Peter Zijlstra wrote:
> >
> > > > But I think
On Wed, 19 Jul 2017, Li, Aubrey wrote:
> On 2017/7/18 15:19, Thomas Gleixner wrote:
> > You can very well avoid it by taking the irq timings or whatever other
> > information into account for the NOHZ decision.
> >
> If I read the source correctly, irq timing statistics computation happens at
>
On Wed, 19 Jul 2017, Li, Aubrey wrote:
> On 2017/7/18 15:19, Thomas Gleixner wrote:
> > You can very well avoid it by taking the irq timings or whatever other
> > information into account for the NOHZ decision.
> >
> If I read the source correctly, irq timing statistics computation happens at
>
On 2017/7/18 15:19, Thomas Gleixner wrote:
> On Mon, 17 Jul 2017, Andi Kleen wrote:
>> On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
>>> On Mon, 17 Jul 2017, Andi Kleen wrote:
>>>
> We need a tradeoff here IMHO. I'll check Daniel's work to understand
> how/if
> it's
On 2017/7/18 15:19, Thomas Gleixner wrote:
> On Mon, 17 Jul 2017, Andi Kleen wrote:
>> On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
>>> On Mon, 17 Jul 2017, Andi Kleen wrote:
>>>
> We need a tradeoff here IMHO. I'll check Daniel's work to understand
> how/if
> it's
On 2017/7/18 23:20, Paul E. McKenney wrote:
>> 2) for rcu idle enter/exit, I measured the details which Paul provided, and
>> the result matches with what I have measured before, nothing notable found.
>> But it still makes more sense if we can make rcu idle enter/exit hooked with
>> tick off.
On 2017/7/18 23:20, Paul E. McKenney wrote:
>> 2) for rcu idle enter/exit, I measured the details which Paul provided, and
>> the result matches with what I have measured before, nothing notable found.
>> But it still makes more sense if we can make rcu idle enter/exit hooked with
>> tick off.
On Mon, Jul 17, 2017 at 12:48:38PM -0700, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
> > Of course, this all assumes a Gaussian distribution to begin with, if we
> > get bimodal (or worse) distributions we can still get it wrong. To fix
> > that, we'd need to do
On Mon, Jul 17, 2017 at 12:48:38PM -0700, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
> > Of course, this all assumes a Gaussian distribution to begin with, if we
> > get bimodal (or worse) distributions we can still get it wrong. To fix
> > that, we'd need to do
On Tue, Jul 18, 2017 at 09:37:57AM -0700, Arjan van de Ven wrote:
> that's just a matter of fixing the C1 and later thresholds to line up right.
> shrug that's the most trivial thing to do, it's a number in a table.
Well, they represent a physical measure, namely the break-even-time. If
you go
On Tue, Jul 18, 2017 at 09:37:57AM -0700, Arjan van de Ven wrote:
> that's just a matter of fixing the C1 and later thresholds to line up right.
> shrug that's the most trivial thing to do, it's a number in a table.
Well, they represent a physical measure, namely the break-even-time. If
you go
On Tue, Jul 18, 2017 at 02:56:47PM +0800, Li, Aubrey wrote:
> 3) for tick nohz idle, we want to skip if the coming idle is short. If we can
> skip the tick nohz idle, we then skip all the items depending on it. But,
> there
> are two hard points:
>
> 3.1) how to compute the period of the coming
On Tue, Jul 18, 2017 at 02:56:47PM +0800, Li, Aubrey wrote:
> 3) for tick nohz idle, we want to skip if the coming idle is short. If we can
> skip the tick nohz idle, we then skip all the items depending on it. But,
> there
> are two hard points:
>
> 3.1) how to compute the period of the coming
On 7/18/2017 9:36 AM, Peter Zijlstra wrote:
On Tue, Jul 18, 2017 at 08:29:40AM -0700, Arjan van de Ven wrote:
the most obvious way to do this (for me, maybe I'm naive) is to add another
C state, lets call it "C1-lite" with its own thresholds and power levels etc,
and just let that be picked
On 7/18/2017 9:36 AM, Peter Zijlstra wrote:
On Tue, Jul 18, 2017 at 08:29:40AM -0700, Arjan van de Ven wrote:
the most obvious way to do this (for me, maybe I'm naive) is to add another
C state, lets call it "C1-lite" with its own thresholds and power levels etc,
and just let that be picked
On Tue, Jul 18, 2017 at 08:29:40AM -0700, Arjan van de Ven wrote:
>
> the most obvious way to do this (for me, maybe I'm naive) is to add another
> C state, lets call it "C1-lite" with its own thresholds and power levels etc,
> and just let that be picked naturally based on the heuristics.
> (if
On Tue, Jul 18, 2017 at 08:29:40AM -0700, Arjan van de Ven wrote:
>
> the most obvious way to do this (for me, maybe I'm naive) is to add another
> C state, lets call it "C1-lite" with its own thresholds and power levels etc,
> and just let that be picked naturally based on the heuristics.
> (if
On 7/18/2017 8:20 AM, Paul E. McKenney wrote:
3.2) how to determine if the idle is short or long. My current proposal is to
use a tunable value via /sys, while Peter prefers an auto-adjust mechanism. I
didn't get the details of an auto-adjust mechanism yet
the most obvious way to do this (for
On 7/18/2017 8:20 AM, Paul E. McKenney wrote:
3.2) how to determine if the idle is short or long. My current proposal is to
use a tunable value via /sys, while Peter prefers an auto-adjust mechanism. I
didn't get the details of an auto-adjust mechanism yet
the most obvious way to do this (for
On Tue, Jul 18, 2017 at 02:56:47PM +0800, Li, Aubrey wrote:
> On 2017/7/18 14:43, Thomas Gleixner wrote:
> > On Mon, 17 Jul 2017, Andi Kleen wrote:
> >
> >>> We need a tradeoff here IMHO. I'll check Daniel's work to understand
> >>> how/if
> >>> it's better than menu governor.
> >>
> >> I still
On Tue, Jul 18, 2017 at 02:56:47PM +0800, Li, Aubrey wrote:
> On 2017/7/18 14:43, Thomas Gleixner wrote:
> > On Mon, 17 Jul 2017, Andi Kleen wrote:
> >
> >>> We need a tradeoff here IMHO. I'll check Daniel's work to understand
> >>> how/if
> >>> it's better than menu governor.
> >>
> >> I still
On Tue, Jul 18, 2017 at 11:14:57AM +0800, Li, Aubrey wrote:
> On 2017/7/18 3:23, Peter Zijlstra wrote:
> > On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
> >>> And as said; Daniel has been working on a better predictor -- now he's
> >>> probably not used it on the network workload
On Tue, Jul 18, 2017 at 11:14:57AM +0800, Li, Aubrey wrote:
> On 2017/7/18 3:23, Peter Zijlstra wrote:
> > On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
> >>> And as said; Daniel has been working on a better predictor -- now he's
> >>> probably not used it on the network workload
On Mon, 17 Jul 2017, Andi Kleen wrote:
> On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
> > On Mon, 17 Jul 2017, Andi Kleen wrote:
> >
> > > > We need a tradeoff here IMHO. I'll check Daniel's work to understand
> > > > how/if
> > > > it's better than menu governor.
> > >
> >
On Mon, 17 Jul 2017, Andi Kleen wrote:
> On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
> > On Mon, 17 Jul 2017, Andi Kleen wrote:
> >
> > > > We need a tradeoff here IMHO. I'll check Daniel's work to understand
> > > > how/if
> > > > it's better than menu governor.
> > >
> >
On Mon, 17 Jul 2017, Andi Kleen wrote:
> On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
> > On Mon, 17 Jul 2017, Andi Kleen wrote:
> >
> > > > We need a tradeoff here IMHO. I'll check Daniel's work to understand
> > > > how/if
> > > > it's better than menu governor.
> > >
> >
On Mon, 17 Jul 2017, Andi Kleen wrote:
> On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
> > On Mon, 17 Jul 2017, Andi Kleen wrote:
> >
> > > > We need a tradeoff here IMHO. I'll check Daniel's work to understand
> > > > how/if
> > > > it's better than menu governor.
> > >
> >
On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
> On Mon, 17 Jul 2017, Andi Kleen wrote:
>
> > > We need a tradeoff here IMHO. I'll check Daniel's work to understand
> > > how/if
> > > it's better than menu governor.
> >
> > I still would like to see how the fast path without
On Tue, Jul 18, 2017 at 08:43:53AM +0200, Thomas Gleixner wrote:
> On Mon, 17 Jul 2017, Andi Kleen wrote:
>
> > > We need a tradeoff here IMHO. I'll check Daniel's work to understand
> > > how/if
> > > it's better than menu governor.
> >
> > I still would like to see how the fast path without
On 2017/7/18 14:43, Thomas Gleixner wrote:
> On Mon, 17 Jul 2017, Andi Kleen wrote:
>
>>> We need a tradeoff here IMHO. I'll check Daniel's work to understand how/if
>>> it's better than menu governor.
>>
>> I still would like to see how the fast path without the C1 heuristic works.
>>
>> Fast
On 2017/7/18 14:43, Thomas Gleixner wrote:
> On Mon, 17 Jul 2017, Andi Kleen wrote:
>
>>> We need a tradeoff here IMHO. I'll check Daniel's work to understand how/if
>>> it's better than menu governor.
>>
>> I still would like to see how the fast path without the C1 heuristic works.
>>
>> Fast
On Mon, 17 Jul 2017, Andi Kleen wrote:
> > We need a tradeoff here IMHO. I'll check Daniel's work to understand how/if
> > it's better than menu governor.
>
> I still would like to see how the fast path without the C1 heuristic works.
>
> Fast pathing is a different concept from a better
On Mon, 17 Jul 2017, Andi Kleen wrote:
> > We need a tradeoff here IMHO. I'll check Daniel's work to understand how/if
> > it's better than menu governor.
>
> I still would like to see how the fast path without the C1 heuristic works.
>
> Fast pathing is a different concept from a better
> We need a tradeoff here IMHO. I'll check Daniel's work to understand how/if
> it's better than menu governor.
I still would like to see how the fast path without the C1 heuristic works.
Fast pathing is a different concept from a better predictor. IMHO we need
both, but the first is likely
> We need a tradeoff here IMHO. I'll check Daniel's work to understand how/if
> it's better than menu governor.
I still would like to see how the fast path without the C1 heuristic works.
Fast pathing is a different concept from a better predictor. IMHO we need
both, but the first is likely
On 2017/7/18 3:48, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
>> Of course, this all assumes a Gaussian distribution to begin with, if we
>> get bimodal (or worse) distributions we can still get it wrong. To fix
>> that, we'd need to do something better than what we
On 2017/7/18 3:48, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
>> Of course, this all assumes a Gaussian distribution to begin with, if we
>> get bimodal (or worse) distributions we can still get it wrong. To fix
>> that, we'd need to do something better than what we
On 2017/7/18 3:23, Peter Zijlstra wrote:
> On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
>>> And as said; Daniel has been working on a better predictor -- now he's
>>> probably not used it on the network workload you're looking at, so that
>>> might be something to consider.
>>
>>
On 2017/7/18 3:23, Peter Zijlstra wrote:
> On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
>>> And as said; Daniel has been working on a better predictor -- now he's
>>> probably not used it on the network workload you're looking at, so that
>>> might be something to consider.
>>
>>
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
> On 7/17/2017 12:46 PM, Thomas Gleixner wrote:
> > That's where Daniel Lezcanos work of predicting interrupts comes in and
> > that's the right solution to the problem. The core infrastructure has been
> > merged, just the idle/cpufreq users are not
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
> On 7/17/2017 12:46 PM, Thomas Gleixner wrote:
> > That's where Daniel Lezcanos work of predicting interrupts comes in and
> > that's the right solution to the problem. The core infrastructure has been
> > merged, just the idle/cpufreq users are not
On 7/17/2017 12:53 PM, Thomas Gleixner wrote:
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Of course, this all assumes a Gaussian distribution to begin with, if we
get bimodal (or worse) distributions we can still get it wrong. To fix
that, we'd
On 7/17/2017 12:53 PM, Thomas Gleixner wrote:
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Of course, this all assumes a Gaussian distribution to begin with, if we
get bimodal (or worse) distributions we can still get it wrong. To fix
that, we'd
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
> > Of course, this all assumes a Gaussian distribution to begin with, if we
> > get bimodal (or worse) distributions we can still get it wrong. To fix
> > that, we'd need to do something better than what
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
> > Of course, this all assumes a Gaussian distribution to begin with, if we
> > get bimodal (or worse) distributions we can still get it wrong. To fix
> > that, we'd need to do something better than what
On 7/17/2017 12:46 PM, Thomas Gleixner wrote:
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Now I think the problem is that the current predictor goes for an
average idle duration. This means that we, on average, get it wrong 50%
of the time. For
On 7/17/2017 12:46 PM, Thomas Gleixner wrote:
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Now I think the problem is that the current predictor goes for an
average idle duration. This means that we, on average, get it wrong 50%
of the time. For
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Of course, this all assumes a Gaussian distribution to begin with, if we
get bimodal (or worse) distributions we can still get it wrong. To fix
that, we'd need to do something better than what we currently have.
fwiw some time ago I made a chart
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Of course, this all assumes a Gaussian distribution to begin with, if we
get bimodal (or worse) distributions we can still get it wrong. To fix
that, we'd need to do something better than what we currently have.
fwiw some time ago I made a chart
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
> > Now I think the problem is that the current predictor goes for an
> > average idle duration. This means that we, on average, get it wrong 50%
> > of the time. For performance that's bad.
>
> that's not
On Mon, 17 Jul 2017, Arjan van de Ven wrote:
> On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
> > Now I think the problem is that the current predictor goes for an
> > average idle duration. This means that we, on average, get it wrong 50%
> > of the time. For performance that's bad.
>
> that's not
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Now I think the problem is that the current predictor goes for an
average idle duration. This means that we, on average, get it wrong 50%
of the time. For performance that's bad.
that's not really what it does; it looks at next tick
and then
On 7/17/2017 12:23 PM, Peter Zijlstra wrote:
Now I think the problem is that the current predictor goes for an
average idle duration. This means that we, on average, get it wrong 50%
of the time. For performance that's bad.
that's not really what it does; it looks at next tick
and then
On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
> > And as said; Daniel has been working on a better predictor -- now he's
> > probably not used it on the network workload you're looking at, so that
> > might be something to consider.
>
> Deriving a better idle predictor is a bit
On Fri, Jul 14, 2017 at 09:26:19AM -0700, Andi Kleen wrote:
> > And as said; Daniel has been working on a better predictor -- now he's
> > probably not used it on the network workload you're looking at, so that
> > might be something to consider.
>
> Deriving a better idle predictor is a bit
On 2017/7/17 21:58, Peter Zijlstra wrote:
> On Mon, Jul 17, 2017 at 09:24:51PM +0800, Li, Aubrey wrote:
>> On 2017/7/14 12:05, Paul E. McKenney wrote:
>>>
>>> More specifically: rcu_needs_cpu(), rcu_prepare_for_idle(),
>>> rcu_cleanup_after_idle(), rcu_eqs_enter(), rcu_eqs_enter_common(),
>>>
On 2017/7/17 21:58, Peter Zijlstra wrote:
> On Mon, Jul 17, 2017 at 09:24:51PM +0800, Li, Aubrey wrote:
>> On 2017/7/14 12:05, Paul E. McKenney wrote:
>>>
>>> More specifically: rcu_needs_cpu(), rcu_prepare_for_idle(),
>>> rcu_cleanup_after_idle(), rcu_eqs_enter(), rcu_eqs_enter_common(),
>>>
On Mon, Jul 17, 2017 at 09:24:51PM +0800, Li, Aubrey wrote:
> On 2017/7/14 12:05, Paul E. McKenney wrote:
> >
> > More specifically: rcu_needs_cpu(), rcu_prepare_for_idle(),
> > rcu_cleanup_after_idle(), rcu_eqs_enter(), rcu_eqs_enter_common(),
> > rcu_dynticks_eqs_enter(),
On Mon, Jul 17, 2017 at 09:24:51PM +0800, Li, Aubrey wrote:
> On 2017/7/14 12:05, Paul E. McKenney wrote:
> >
> > More specifically: rcu_needs_cpu(), rcu_prepare_for_idle(),
> > rcu_cleanup_after_idle(), rcu_eqs_enter(), rcu_eqs_enter_common(),
> > rcu_dynticks_eqs_enter(),
On 2017/7/17 17:21, Peter Zijlstra wrote:
> On Fri, Jul 14, 2017 at 09:03:14AM -0700, Andi Kleen wrote:
>> fast idle doesn't have an upper bound.
>>
>> If the prediction exceeds the fast idle threshold any C state can be used.
>>
>> It's just another state (fast C1), but right now it has an own
On 2017/7/17 17:21, Peter Zijlstra wrote:
> On Fri, Jul 14, 2017 at 09:03:14AM -0700, Andi Kleen wrote:
>> fast idle doesn't have an upper bound.
>>
>> If the prediction exceeds the fast idle threshold any C state can be used.
>>
>> It's just another state (fast C1), but right now it has an own
On 2017/7/14 12:05, Paul E. McKenney wrote:
>
> More specifically: rcu_needs_cpu(), rcu_prepare_for_idle(),
> rcu_cleanup_after_idle(), rcu_eqs_enter(), rcu_eqs_enter_common(),
> rcu_dynticks_eqs_enter(), do_nocb_deferred_wakeup(),
> rcu_dynticks_task_enter(), rcu_eqs_exit(),
On 2017/7/14 12:05, Paul E. McKenney wrote:
>
> More specifically: rcu_needs_cpu(), rcu_prepare_for_idle(),
> rcu_cleanup_after_idle(), rcu_eqs_enter(), rcu_eqs_enter_common(),
> rcu_dynticks_eqs_enter(), do_nocb_deferred_wakeup(),
> rcu_dynticks_task_enter(), rcu_eqs_exit(),
On Fri, Jul 14, 2017 at 09:03:14AM -0700, Andi Kleen wrote:
> fast idle doesn't have an upper bound.
>
> If the prediction exceeds the fast idle threshold any C state can be used.
>
> It's just another state (fast C1), but right now it has an own threshold
> which may be different from standard
On Fri, Jul 14, 2017 at 09:03:14AM -0700, Andi Kleen wrote:
> fast idle doesn't have an upper bound.
>
> If the prediction exceeds the fast idle threshold any C state can be used.
>
> It's just another state (fast C1), but right now it has an own threshold
> which may be different from standard
> And as said; Daniel has been working on a better predictor -- now he's
> probably not used it on the network workload you're looking at, so that
> might be something to consider.
Deriving a better idle predictor is a bit orthogonal to fast idle.
It would be a good idea to set the fast idle
> And as said; Daniel has been working on a better predictor -- now he's
> probably not used it on the network workload you're looking at, so that
> might be something to consider.
Deriving a better idle predictor is a bit orthogonal to fast idle.
It would be a good idea to set the fast idle
On Fri, Jul 14, 2017 at 08:53:56AM -0700, Andi Kleen wrote:
> > > - if (cpuidle_not_available(drv, dev)) {
> > > + if (cpuidle_not_available(drv, dev) || this_is_a_fast_idle) {
> > > default_idle_call();
> > > goto exit_idle;
> > > }
> >
> > No, that's wrong. We want to fix
On Fri, Jul 14, 2017 at 08:53:56AM -0700, Andi Kleen wrote:
> > > - if (cpuidle_not_available(drv, dev)) {
> > > + if (cpuidle_not_available(drv, dev) || this_is_a_fast_idle) {
> > > default_idle_call();
> > > goto exit_idle;
> > > }
> >
> > No, that's wrong. We want to fix
On Fri, Jul 14, 2017 at 05:58:53PM +0200, Peter Zijlstra wrote:
> On Fri, Jul 14, 2017 at 08:52:28AM -0700, Arjan van de Ven wrote:
> > On 7/14/2017 8:38 AM, Peter Zijlstra wrote:
> > > No, that's wrong. We want to fix the normal C state selection process to
> > > pick the right C state.
> > >
>
On Fri, Jul 14, 2017 at 05:58:53PM +0200, Peter Zijlstra wrote:
> On Fri, Jul 14, 2017 at 08:52:28AM -0700, Arjan van de Ven wrote:
> > On 7/14/2017 8:38 AM, Peter Zijlstra wrote:
> > > No, that's wrong. We want to fix the normal C state selection process to
> > > pick the right C state.
> > >
>
On Fri, Jul 14, 2017 at 08:52:28AM -0700, Arjan van de Ven wrote:
> On 7/14/2017 8:38 AM, Peter Zijlstra wrote:
> > No, that's wrong. We want to fix the normal C state selection process to
> > pick the right C state.
> >
> > The fast-idle criteria could cut off a whole bunch of available C
> >
On Fri, Jul 14, 2017 at 08:52:28AM -0700, Arjan van de Ven wrote:
> On 7/14/2017 8:38 AM, Peter Zijlstra wrote:
> > No, that's wrong. We want to fix the normal C state selection process to
> > pick the right C state.
> >
> > The fast-idle criteria could cut off a whole bunch of available C
> >
1 - 100 of 190 matches
Mail list logo