Skip to contentSkip to conversation

find_busiest_group using lots of CPU

DISCUSSION Jens Axboe 3 participants 8 messages Sep 30, 2009 – Oct 6, 2009

lore
Share & tools
lore

Jens Axboe

author
#0 lore raw

Hi,

I stuffed a few more SSDs into my text box. Running a simple workload that just does streaming reads from 10 processes (throughput is around 2.2GB/sec), find_busiest_group() is using > 10% of the CPU time. This is a 64 thread box.

The top two profile entries are:

    10.86%      fio  [kernel]                [k] find_busiest_group
                |          
                |--99.91%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                 --0.09%-- [...]

     5.78%      fio  [kernel]                [k] cpumask_next_and
                |          
                |--67.21%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                |          
                 --32.79%-- find_busiest_group
                           thread_return
                           io_schedule
                           sys_io_getevents
                           system_call_fastpath
                           0x7f4b50b61604
                           |          
                            --100.00%-- td_io_getevents
                                      io_u_queued_complete
                                      thread_main
                                      run_threads
                                      main
                                      __libc_start_main

This is with SCHED_DEBUG=y and SCHEDSTATS=y enabled, I just tried with both disabled but that yields the same result (well actually worse, 22% spent in there. dunno if that's normal "fluctuation"). GROUP_SCHED is not set. This seems way excessive!

Signature · 1 line
Jens Axboe
#1 lore raw
Quoted 54 lines · Jens Axboe

Hi,

I stuffed a few more SSDs into my text box. Running a simple workload that just does streaming reads from 10 processes (throughput is around 2.2GB/sec), find_busiest_group() is using > 10% of the CPU time. This is a 64 thread box.

The top two profile entries are:

    10.86%      fio  [kernel]                [k] find_busiest_group
                |          
                |--99.91%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                 --0.09%-- [...]

     5.78%      fio  [kernel]                [k] cpumask_next_and
                |          
                |--67.21%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                |          
                 --32.79%-- find_busiest_group
                           thread_return
                           io_schedule
                           sys_io_getevents
                           system_call_fastpath
                           0x7f4b50b61604
                           |          
                            --100.00%-- td_io_getevents
                                      io_u_queued_complete
                                      thread_main
                                      run_threads
                                      main
                                      __libc_start_main
Quoted 4 lines

This is with SCHED_DEBUG=y and SCHEDSTATS=y enabled, I just tried with both disabled but that yields the same result (well actually worse, 22% spent in there. dunno if that's normal "fluctuation"). GROUP_SCHED is not set. This seems way excessive!

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

	val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

[ assuming SCHED_DEBUG=y ]

Cure things?

If so, then its spending time looking for work, which there might not be on your machine, since everything is waiting for IO or somesuch.

Not really sure what to do about it though, this is a quad socket nehalem, right? We could possibly disable SD_BALANCE_NEWIDLE on the NODE level, but that would again decrease throughput in things like kbuild.

author
#2 lore raw
Quoted 63 lines · Peter Zijlstra
Quoted 54 lines · Jens Axboe

Hi,

I stuffed a few more SSDs into my text box. Running a simple workload that just does streaming reads from 10 processes (throughput is around 2.2GB/sec), find_busiest_group() is using > 10% of the CPU time. This is a 64 thread box.

The top two profile entries are:

    10.86%      fio  [kernel]                [k] find_busiest_group
                |          
                |--99.91%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                 --0.09%-- [...]

     5.78%      fio  [kernel]                [k] cpumask_next_and
                |          
                |--67.21%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                |          
                 --32.79%-- find_busiest_group
                           thread_return
                           io_schedule
                           sys_io_getevents
                           system_call_fastpath
                           0x7f4b50b61604
                           |          
                            --100.00%-- td_io_getevents
                                      io_u_queued_complete
                                      thread_main
                                      run_threads
                                      main
                                      __libc_start_main
Quoted 4 lines

This is with SCHED_DEBUG=y and SCHEDSTATS=y enabled, I just tried with both disabled but that yields the same result (well actually worse, 22% spent in there. dunno if that's normal "fluctuation"). GROUP_SCHED is not set. This seems way excessive!

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

Quoted 8 lines

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

	val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

[ assuming SCHED_DEBUG=y ]

Cure things?

I can try, as mentioned it doesn't look any better with SCHED_DEBUG=n

Quoted 2 lines

If so, then its spending time looking for work, which there might not be on your machine, since everything is waiting for IO or somesuch.

OK, just seems way excessive for something which is only 10 tasks and not even that context switch intensive.

Quoted 3 lines

Not really sure what to do about it though, this is a quad socket nehalem, right? We could possibly disable SD_BALANCE_NEWIDLE on the NODE level, but that would again decrease throughput in things like kbuild.

Yes, it's a quad socket nehalem. I'll see if disabling NEWIDLE makes a difference, I need to run some other tests on that box today anyway.

Signature · 1 line
Jens Axboe
author
#3 lore raw
Quoted 69 lines · Jens Axboe
Quoted 68 lines · Peter Zijlstra
Quoted 54 lines · Jens Axboe

Hi,

I stuffed a few more SSDs into my text box. Running a simple workload that just does streaming reads from 10 processes (throughput is around 2.2GB/sec), find_busiest_group() is using > 10% of the CPU time. This is a 64 thread box.

The top two profile entries are:

    10.86%      fio  [kernel]                [k] find_busiest_group
                |          
                |--99.91%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                 --0.09%-- [...]

     5.78%      fio  [kernel]                [k] cpumask_next_and
                |          
                |--67.21%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                |          
                 --32.79%-- find_busiest_group
                           thread_return
                           io_schedule
                           sys_io_getevents
                           system_call_fastpath
                           0x7f4b50b61604
                           |          
                            --100.00%-- td_io_getevents
                                      io_u_queued_complete
                                      thread_main
                                      run_threads
                                      main
                                      __libc_start_main
Quoted 4 lines

This is with SCHED_DEBUG=y and SCHEDSTATS=y enabled, I just tried with both disabled but that yields the same result (well actually worse, 22% spent in there. dunno if that's normal "fluctuation"). GROUP_SCHED is not set. This seems way excessive!

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

	val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

Quoted 5 lines
Quoted 3 lines

[ assuming SCHED_DEBUG=y ]

Cure things?

I can try, as mentioned it doesn't look any better with SCHED_DEBUG=n

It does, it's gone from the profiles.

Signature · 1 line
Jens Axboe
#4 lore raw
Quoted 70 lines · Jens Axboe
Quoted 69 lines · Jens Axboe
Quoted 68 lines · Peter Zijlstra
Quoted 54 lines · Jens Axboe

Hi,

I stuffed a few more SSDs into my text box. Running a simple workload that just does streaming reads from 10 processes (throughput is around 2.2GB/sec), find_busiest_group() is using > 10% of the CPU time. This is a 64 thread box.

The top two profile entries are:

    10.86%      fio  [kernel]                [k] find_busiest_group
                |          
                |--99.91%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                 --0.09%-- [...]

     5.78%      fio  [kernel]                [k] cpumask_next_and
                |          
                |--67.21%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                |          
                 --32.79%-- find_busiest_group
                           thread_return
                           io_schedule
                           sys_io_getevents
                           system_call_fastpath
                           0x7f4b50b61604
                           |          
                            --100.00%-- td_io_getevents
                                      io_u_queued_complete
                                      thread_main
                                      run_threads
                                      main
                                      __libc_start_main
Quoted 4 lines

This is with SCHED_DEBUG=y and SCHEDSTATS=y enabled, I just tried with both disabled but that yields the same result (well actually worse, 22% spent in there. dunno if that's normal "fluctuation"). GROUP_SCHED is not set. This seems way excessive!

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

	val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

[ assuming SCHED_DEBUG=y ]

Cure things?

I can try, as mentioned it doesn't look any better with SCHED_DEBUG=n

It does, it's gone from the profiles.

Peter mentioned SCHED_DEBUG=y to have /proc/sys/kernel/sched_domain/cpu*/domain*/flags available.

Ingo

author
#5 lore raw
Quoted 75 lines · Ingo Molnar
Quoted 72 lines · Jens Axboe
Quoted 71 lines · Jens Axboe
Quoted 70 lines · Peter Zijlstra
Quoted 54 lines · Jens Axboe

Hi,

I stuffed a few more SSDs into my text box. Running a simple workload that just does streaming reads from 10 processes (throughput is around 2.2GB/sec), find_busiest_group() is using > 10% of the CPU time. This is a 64 thread box.

The top two profile entries are:

    10.86%      fio  [kernel]                [k] find_busiest_group
                |          
                |--99.91%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                 --0.09%-- [...]

     5.78%      fio  [kernel]                [k] cpumask_next_and
                |          
                |--67.21%-- thread_return
                |          io_schedule
                |          sys_io_getevents
                |          system_call_fastpath
                |          0x7f4b50b61604
                |          |          
                |           --100.00%-- td_io_getevents
                |                     io_u_queued_complete
                |                     thread_main
                |                     run_threads
                |                     main
                |                     __libc_start_main
                |          
                 --32.79%-- find_busiest_group
                           thread_return
                           io_schedule
                           sys_io_getevents
                           system_call_fastpath
                           0x7f4b50b61604
                           |          
                            --100.00%-- td_io_getevents
                                      io_u_queued_complete
                                      thread_main
                                      run_threads
                                      main
                                      __libc_start_main
Quoted 4 lines

This is with SCHED_DEBUG=y and SCHEDSTATS=y enabled, I just tried with both disabled but that yields the same result (well actually worse, 22% spent in there. dunno if that's normal "fluctuation"). GROUP_SCHED is not set. This seems way excessive!

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

	val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

[ assuming SCHED_DEBUG=y ]

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines
Quoted 1 line

Cure things?

I can try, as mentioned it doesn't look any better with SCHED_DEBUG=n

It does, it's gone from the profiles.

Peter mentioned SCHED_DEBUG=y to have /proc/sys/kernel/sched_domain/cpu*/domain*/flags available.

In case it wasn't clear, what I did was run the 'for...' from Peter (with a kernel with SCHED_DEBUG=y) and it cured the "problem". Peter asked whether it cured things, to which I replied "It does, it's gone from the profiles".

Signature · 1 line
Jens Axboe
#6 lore raw
In reply to Jens Axboe
Quoted 8 lines · Jens Axboe
Quoted 8 lines
Quoted 8 lines

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

    val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

[ assuming SCHED_DEBUG=y ]

Cure things?

I can try,

It does, it's gone from the profiles.

Hrmm,. ok. I'll go prod at the NEWIDLE bits to see if anything stands out.

If only perf-annotate could generate annotated C... maybe I'll look at that.

Thanks!

#7 lore raw
Quoted 15 lines · Peter Zijlstra
Quoted 14 lines · Jens Axboe
Quoted 13 lines
Quoted 11 lines

io_schedule() straight into find_busiest_group() leads me to think this could be SD_BALANCE_NEWIDLE, does something like:

for i in /proc/sys/kernel/sched_domain/cpu*/domain*/flags; do

    val=`cat $i`; echo $((val & ~0x02)) > $i; 

done

[ assuming SCHED_DEBUG=y ]

Cure things?

I can try,

Quoted 7 lines
Quoted 1 line

It does, it's gone from the profiles.

Hrmm,. ok. I'll go prod at the NEWIDLE bits to see if anything stands out.

If only perf-annotate could generate annotated C... maybe I'll look at that.

hey it can do that - build a vmlinux with DEBUG_INFO=y and it should just work :-)

Ingo