Skip to contentSkip to conversation

PlugSched-6.5.1 for 2.6.22

RFC Al Boldi 5 participants 7 messages Jul 11, 2007 – Jul 16, 2007

lore
Share & tools
lore

Al Boldi

author
#0 lore raw
Quoted 2 lines · Peter Williams

Probably the last one now that CFS is in the main line :-(.

What do you mean? A pluggable scheduler framework is indispensible even in the presence of CFS or SD.

Quoted 10 lines

A patch for 2.6.22 is available at:

<http://downloads.sourceforge.net/cpuse/plugsched-6.5.1-for-2.6.22.patch>

Very Brief Documentation:

You can select a default scheduler at kernel build time. If you wish to boot with a scheduler other than the default it can be selected at boot time by adding:

Quoted 6 lines

cpusched=<scheduler>

to the boot command line where <scheduler> is one of: ingosched, ingo_ll, nicksched, staircase, spa_no_frills, spa_ws, spa_svr, spa_ebs or zaphod. If you don't change the default when you build the kernel the default scheduler will be ingosched (which is the normal scheduler).

Can't PlugSched include CFS and SD?

Thanks!

--
Al

#1 lore raw
Quoted 6 lines · Al Boldi
Quoted 2 lines · Peter Williams

Probably the last one now that CFS is in the main line :-(.

What do you mean? A pluggable scheduler framework is indispensible even in the presence of CFS or SD.

Indeed, and I hope it gets merged, giving people the chance to test out different schedulers easily, without having to do patching, de-patching, re-patching and blah blah blah.

But somehow I doubt it'll get merged now. My bet is that it won't even be taken into consideration, for the same reasons that ultimately drove Con Kolivas to giving up on development of SD. And if it's so, chances are people will get tired of trying to get it merged. :/

Quoted 10 lines
Quoted 10 lines

A patch for 2.6.22 is available at:

<http://downloads.sourceforge.net/cpuse/plugsched-6.5.1-for-2.6.22.patch>

Very Brief Documentation:

You can select a default scheduler at kernel build time. If you wish to boot with a scheduler other than the default it can be selected at boot time by adding:

Quoted 8 lines
Quoted 6 lines

cpusched=<scheduler>

to the boot command line where <scheduler> is one of: ingosched, ingo_ll, nicksched, staircase, spa_no_frills, spa_ws, spa_svr, spa_ebs or zaphod. If you don't change the default when you build the kernel the default scheduler will be ingosched (which is the normal scheduler).

Can't PlugSched include CFS and SD?

If I read the announcement correctly, those are the ones referred to as 'ingosched' and 'staircase' :)

Signature · 1 line
Giuseppe "Oblomov" Bilotta
#2 lore raw
Quoted 7 lines · Al Boldi
Quoted 6 lines
Quoted 2 lines · Peter Williams

Probably the last one now that CFS is in the main line :-(.

What do you mean? A pluggable scheduler framework is indispensible

even

Quoted 4 lines

in

Quoted 1 line

the presence of CFS or SD.

Indeed, and I hope it gets merged, giving people the chance to test

out

Quoted 1 line

different schedulers easily, without having to do patching,

de-patching,

Quoted 2 lines

re-patching and blah blah blah.

Yes, such a framework would be invaluable, especially given that scheduling often has conflicting goals and different workloads have different requirements. No single solution fits all.

Quoted 1 line

But somehow I doubt it'll get merged now. My bet is that it won't even

be

Quoted 1 line

taken into consideration, for the same reasons that ultimately drove

Con

Quoted 11 lines

Kolivas to giving up on development of SD. And if it's so, chances are people will get tired of trying to get it merged. :/

Quoted 3 lines
Quoted 3 lines

A patch for 2.6.22 is available at:

<http://downloads.sourceforge.net/cpuse/plugsched-6.5.1-for-

2.6.22.patch>

Quoted 4 lines
Quoted 4 lines

Very Brief Documentation:

You can select a default scheduler at kernel build time. If you

wish

Quoted 2 lines

to

Quoted 1 line
Quoted 1 line

boot with a scheduler other than the default it can be selected at

boot

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

time by adding:

cpusched=<scheduler>

to the boot command line where <scheduler> is one of: ingosched, ingo_ll, nicksched, staircase, spa_no_frills, spa_ws, spa_svr,

spa_ebs

Quoted 1 line
Quoted 1 line
Quoted 1 line

or zaphod. If you don't change the default when you build the

kernel

Quoted 1 line
Quoted 1 line
Quoted 1 line

the default scheduler will be ingosched (which is the normal

scheduler).

Quoted 4 lines
Quoted 2 lines

Can't PlugSched include CFS and SD?

If I read the announcement correctly, those are the ones referred to

as

Quoted 7 lines

'ingosched' and 'staircase' :)

--
Giuseppe "Oblomov" Bilotta

-
To unsubscribe from this list: send the line "unsubscribe

linux-kernel" in

Quoted 3 lines

the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#3 lore raw
Quoted 14 lines · Tong N
Quoted 7 lines · Al Boldi
Quoted 6 lines
Quoted 2 lines · Peter Williams

Probably the last one now that CFS is in the main line :-(.

What do you mean? A pluggable scheduler framework is indispensible

even

Quoted 4 lines

in

Quoted 1 line

the presence of CFS or SD.

Indeed, and I hope it gets merged, giving people the chance to test

out

Quoted 1 line

different schedulers easily, without having to do patching,

Quoted 8 lines

de-patching,

Quoted 2 lines

re-patching and blah blah blah.

Yes, such a framework would be invaluable, especially given that scheduling often has conflicting goals and different workloads have different requirements. No single solution fits all.
...

Having a framework for giving people the choice between different solutions usually sounds good in theory, but in practice there's the often underestimated high price of people using a different solution instead of reporting a problem with one solution or people adding features to only one of the solutions resulting in different solutions having different features and bugs instead of one solution with all features and bug fixes.

So you should give a good technical explanation why it's not technically possible or not desirable for one solution to work well for everyone.

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#4 lore raw
Quoted 17 lines

-----Original Message-----
From: Adrian Bunk [mailto:[email protected]]
Sent: Sunday, July 15, 2007 4:46 PM
To: Li, Tong N
Cc: Giuseppe Bilotta; [email protected]
Subject: Re: Re: [ANNOUNCE][RFC] PlugSched-6.5.1 for 2.6.22

Quoted 7 lines · Tong N
Quoted 7 lines · Al Boldi
Quoted 6 lines
Quoted 2 lines · Peter Williams

Probably the last one now that CFS is in the main line :-(.

What do you mean? A pluggable scheduler framework is

indispensible

Quoted 5 lines
Quoted 5 lines

even

Quoted 4 lines

in

Quoted 1 line

the presence of CFS or SD.

Indeed, and I hope it gets merged, giving people the chance to

test

Quoted 19 lines
Quoted 10 lines

out

Quoted 1 line

different schedulers easily, without having to do patching,

de-patching,

Quoted 2 lines

re-patching and blah blah blah.

Yes, such a framework would be invaluable, especially given that scheduling often has conflicting goals and different workloads have different requirements. No single solution fits all.
...

Having a framework for giving people the choice between different solutions usually sounds good in theory, but in practice there's the often underestimated high price of people using a different solution instead of reporting a problem with one solution or people adding features to only one of the solutions resulting in different solutions having different features and bugs instead of one solution with all features and bug fixes.

Quoted 1 line

So you should give a good technical explanation why it's not

technically

Quoted 1 line

possible or not desirable for one solution to work well for everyone.

There are various metrics a scheduler may want to optimize for, such as throughput, response time, power consumption, fairness, and so on. Each of these may also be defined differently in different environments. Take fairness as an example. People have traditionally talked about it in terms of CPU time. Now it'd also make sense to talk about scheduling that enables fair usage of other types of resources, such as shared caches. Different metrics may require different scheduling policies. Plus, many metrics may in fact conflict, e.g., a scheduling policy optimized for throughput may not be power efficient. As a result, Linux, and all general-purpose OSes, strive to achieve a balance, but it's conceivable that different hardware platforms and different application workloads may want to have different scheduling policies to meet their own needs.

Given that the scheduling policies can be diverse, the mechanisms to enforce them can also be different. The per-cpu runqueue model may be best for many scenarios, some scheduling policies (like many real-time ones) might want global knowledge about all tasks in the system and thus prefer a global task queue at the cost of being less scalable and cache-efficient. HPC systems may also want to gang-schedule. And, in terms of implementation, O(1) might be desirable for large-scale MP systems while O(log N) might be good enough for small systems. These are just examples that indicate the scheduler data structures, algorithms, and implementation can have a variety of possibilities in different usage models. A single scheduler that is easily extensible for incorporating different policies would be ideal, but IMO this is not yet the case and may not even be possible. Therefore, I think having a framework that enables multiple schedulers to co-exist would be invaluable and PlugSched seems to be one good step towards this.

tong

#5 lore raw
Quoted 8 lines · Tong N

There are various metrics a scheduler may want to optimize for, such as throughput, response time, power consumption, fairness, and so on. Each of these may also be defined differently in different environments. Take fairness as an example. People have traditionally talked about it in terms of CPU time. Now it'd also make sense to talk about scheduling that enables fair usage of other types of resources, such as shared caches. Different metrics may require different scheduling policies.

Are these real-life problems people need solutions for today or just thoughts someone might want to try and write a paper about?

Quoted 6 lines

Plus, many metrics may in fact conflict, e.g., a scheduling policy optimized for throughput may not be power efficient. As a result, Linux, and all general-purpose OSes, strive to achieve a balance, but it's conceivable that different hardware platforms and different application workloads may want to have different scheduling policies to meet their own needs.

The Linux scheduler already gives you some knobs allowing you to adjust the scheduling to your needs.

What is missing, and why wasn't it included in the current scheduler?

Quoted 7 lines

Given that the scheduling policies can be diverse, the mechanisms to enforce them can also be different. The per-cpu runqueue model may be best for many scenarios, some scheduling policies (like many real-time ones) might want global knowledge about all tasks in the system and thus prefer a global task queue at the cost of being less scalable and cache-efficient. HPC systems may also want to gang-schedule. And, in terms of implementation, O(1) might be desirable for large-scale MP

Quoted 8 lines

systems while O(log N) might be good enough for small systems. These are just examples that indicate the scheduler data structures, algorithms, and implementation can have a variety of possibilities in different usage models. A single scheduler that is easily extensible for incorporating different policies would be ideal, but IMO this is not yet the case and may not even be possible. Therefore, I think having a framework that enables multiple schedulers to co-exist would be invaluable and PlugSched seems to be one good step towards this.

Much is already possible, and e.g. when you talk about HPC you are in an area where the scope of the kernel is anyway often too small and the actual scheduling is done with a userspace batch scheduling program.

What are the real-life problems the current scheduler has, and why weren't they solved when it was designed and implemented?

People implementing a special purpose scheduler to fit their needs would really not be an ideal solution - ideally, there should be one scheduler that handles all real-life problems.

And therefore the first step for people should be to try to get their problems solved with the one scheduler, and if required enhance it, instead of saying NIH and implement their own special purpose scheduler.

Quoted 1 line

tong

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#6 lore raw
Quoted 10 lines · Adrian Bunk
Quoted 8 lines · Tong N

There are various metrics a scheduler may want to optimize for, such as throughput, response time, power consumption, fairness, and so on. Each of these may also be defined differently in different environments. Take fairness as an example. People have traditionally talked about it in terms of CPU time. Now it'd also make sense to talk about scheduling that enables fair usage of other types of resources, such as shared caches. Different metrics may require different scheduling policies.

Quoted 2 lines

Are these real-life problems people need solutions for today or just thoughts someone might want to try and write a paper about?

We are at the door step of multi core evolution. So its a bit of both.

thanks,
suresh