This patch proposes a performance fix for the current IPC semaphore
implementation.
There are two shortcoming in the current implementation:
try_atomic_semop() was called two times to wake up a blocked process,
once from the update_queue() (executed from the process that wakes up
the sleeping process) and once in the retry part of the blocked process
(executed from the block process that gets woken up).
A second issue is that when several sleeping processes that are eligible
for wake up, they woke up in daisy chain formation and each one in turn
to wake up next process in line. However, every time when a process
wakes up, it start scans the wait queue from the beginning, not from
where it was last scanned. This causes large number of unnecessary
scanning of the wait queue under a situation of deep wait queue.
Blocked processes come and go, but chances are there are still quite a
few blocked processes sit at the beginning of that queue.
What we are proposing here is to merge the portion of the code in the
bottom part of sys_semtimedop() (code that gets executed when a sleeping
process gets woken up) into update_queue() function. The benefit is two
folds: (1) is to reduce redundant calls to try_atomic_semop() and (2) to
increase efficiency of finding eligible processes to wake up and higher
concurrency for multiple wake-ups.
We have measured that this patch improves throughput for a large
application significantly on a industry standard benchmark.
This patch is relative to 2.5.72. Any feedback is very much
appreciated.
Some kernel profile data attached:
Kernel profile before optimization:
-----------------------------------------------
0.05 0.14 40805/529060 sys_semop [133]
0.55 1.73 488255/529060 ia64_ret_from_syscall
[2]
[52] 2.5 0.59 1.88 529060 sys_semtimedop [52]
0.05 0.83 477766/817966 schedule_timeout [62]
0.34 0.46 529064/989340 update_queue [61]
0.14 0.00 1006740/6473086 try_atomic_semop [75]
0.06 0.00 529060/989336 ipcperms [149]
-----------------------------------------------
0.30 0.40 460276/989340 semctl_main [68]
0.34 0.46 529064/989340 sys_semtimedop [52]
[61] 1.5 0.64 0.87 989340 update_queue [61]
0.75 0.00 5466346/6473086 try_atomic_semop [75]
0.01 0.11 477676/576698 wake_up_process [146]
-----------------------------------------------
0.14 0.00 1006740/6473086 sys_semtimedop [52]
0.75 0.00 5466346/6473086 update_queue [61]
[75] 0.9 0.89 0.00 6473086 try_atomic_semop [75]
-----------------------------------------------
Kernel profile with optimization:
-----------------------------------------------
0.03 0.05 26139/503178 sys_semop [155]
0.46 0.92 477039/503178 ia64_ret_from_syscall
[2]
[61] 1.2 0.48 0.97 503178 sys_semtimedop [61]
0.04 0.79 470724/784394 schedule_timeout [62]
0.05 0.00 503178/3301773 try_atomic_semop [109]
0.05 0.00 503178/930934 ipcperms [149]
0.00 0.03 32454/460210 update_queue [99]
-----------------------------------------------
0.00 0.03 32454/460210 sys_semtimedop [61]
0.06 0.36 427756/460210 semctl_main [75]
[99] 0.4 0.06 0.39 460210 update_queue [99]
0.30 0.00 2798595/3301773 try_atomic_semop [109]
0.00 0.09 470630/614097 wake_up_process [146]
-----------------------------------------------
0.05 0.00 503178/3301773 sys_semtimedop [61]
0.30 0.00 2798595/3301773 update_queue [99]
[109] 0.3 0.35 0.00 3301773 try_atomic_semop [109]
-----------------------------------------------
Both number of function calls to try_atomic_semop() and update_queue()
are reduced by 50% as a result of the merge. Execution time of
sys_semtimedop is reduced because of the reduction in the low level
functions.
Looks interesting. Give it a shot tomorrow.
Any chance you and/or Manfred Spraul can cook up a combined
patch with Manfred's changes from here :
http://marc.theaimsgroup.com/?l=linux-kernel&m=105567416526840&w=2
Margit
Chen wrote:
>What we are proposing here is to merge the portion of the code in the
>bottom part of sys_semtimedop() (code that gets executed when a sleeping
>process gets woken up) into update_queue() function. The benefit is two
>folds: (1) is to reduce redundant calls to try_atomic_semop() and (2) to
>increase efficiency of finding eligible processes to wake up and higher
>concurrency for multiple wake-ups.
>
>
Interesting. I'm not sure if you noticed it, but your patch backs out a
change from Christoph Rohland from 1998, probably from SAP DB benchmarking:
linux/ipc/sem.c:
* - The previous code had two flaws:
* 1) It actively gave the semaphore to the next waiting process
* sleeping on the semaphore. Since this process did not have the
* cpu this led to many unnecessary context switches and bad
* performance. Now we only check which process should be able to
* get the semaphore and if this process wants to reduce some
* semaphore value we simply wake it up without doing the
* operation. So it has to try to get it later. Thus e.g. the
* running process may reacquire the semaphore during the current
* time slice. If it only waits for zero or increases the semaphore,
* we do the operation in advance and wake it up.
Perhaps the O(1) scheduler is better at handling the thread switches
than the old scheduler. Could you include an update of the comments into
your patch?
--
Manfred
> Perhaps the O(1) scheduler is better at handling the thread switches
> than the old scheduler. Could you include an update of the comments
into
> your patch?
Yes, our opinion align with your observation as well. The O(1)
scheduler
handles the ctx much better.
- Ken
Hi Andrew,
> ipc/sem.c changed significantly between 2.5.72 and 2.5.73. Could
> you please update this, retest, resend?
>
> Please Cc Manfred on the update too, thanks.
Updated patch relative to 2.5.73, tested and verified. Manfred's change
is in the sem_undo area while this patch is in the wake-up path.
- Ken
Is any of you taking care of this, posted to the lkml days ago? It seems a good improvement. Is it being incorporated / tested in any of the trees? Andrew?
Con, perhaps you can measure it in the next Contest? Perhaps it can help improve both interactiveness and throughput. I think it cannot be bad. Testing it with 01int and granularity.
Thanks for taking a look a it.
Luis Miguel Garcia
>This patch proposes a performance fix for the current IPC semaphore
implementation.
There are two shortcoming in the current implementation:
try_atomic_semop() was called two times to wake up a blocked process,
once from the update_queue() (executed from the process that wakes up
the sleeping process) and once in the retry part of the blocked process
(executed from the block process that gets woken up).
A second issue is that when several sleeping processes that are eligible
for wake up, they woke up in daisy chain formation and each one in turn
to wake up next process in line. However, every time when a process
wakes up, it start scans the wait queue from the beginning, not from
where it was last scanned. This causes large number of unnecessary
scanning of the wait queue under a situation of deep wait queue.
Blocked processes come and go, but chances are there are still quite a
few blocked processes sit at the beginning of that queue.
What we are proposing here is to merge the portion of the code in the
bottom part of sys_semtimedop() (code that gets executed when a sleeping
process gets woken up) into update_queue() function. The benefit is two
folds: (1) is to reduce redundant calls to try_atomic_semop() and (2) to
increase efficiency of finding eligible processes to wake up and higher
concurrency for multiple wake-ups.
We have measured that this patch improves throughput for a large
application significantly on a industry standard benchmark.
This patch is relative to 2.5.72. Any feedback is very much
appreciated.
Some kernel profile data attached:
Kernel profile before optimization:
-----------------------------------------------
0.05 0.14 40805/529060 sys_semop [133]
0.55 1.73 488255/529060 ia64_ret_from_syscall
[2]
[52] 2.5 0.59 1.88 529060 sys_semtimedop [52]
0.05 0.83 477766/817966 schedule_timeout [62]
0.34 0.46 529064/989340 update_queue [61]
0.14 0.00 1006740/6473086 try_atomic_semop [75]
0.06 0.00 529060/989336 ipcperms [149]
-----------------------------------------------
0.30 0.40 460276/989340 semctl_main [68]
0.34 0.46 529064/989340 sys_semtimedop [52]
[61] 1.5 0.64 0.87 989340 update_queue [61]
0.75 0.00 5466346/6473086 try_atomic_semop [75]
0.01 0.11 477676/576698 wake_up_process [146]
-----------------------------------------------
0.14 0.00 1006740/6473086 sys_semtimedop [52]
0.75 0.00 5466346/6473086 update_queue [61]
[75] 0.9 0.89 0.00 6473086 try_atomic_semop [75]
-----------------------------------------------
Kernel profile with optimization:
-----------------------------------------------
0.03 0.05 26139/503178 sys_semop [155]
0.46 0.92 477039/503178 ia64_ret_from_syscall
[2]
[61] 1.2 0.48 0.97 503178 sys_semtimedop [61]
0.04 0.79 470724/784394 schedule_timeout [62]
0.05 0.00 503178/3301773 try_atomic_semop [109]
0.05 0.00 503178/930934 ipcperms [149]
0.00 0.03 32454/460210 update_queue [99]
-----------------------------------------------
0.00 0.03 32454/460210 sys_semtimedop [61]
0.06 0.36 427756/460210 semctl_main [75]
[99] 0.4 0.06 0.39 460210 update_queue [99]
0.30 0.00 2798595/3301773 try_atomic_semop [109]
0.00 0.09 470630/614097 wake_up_process [146]
-----------------------------------------------
0.05 0.00 503178/3301773 sys_semtimedop [61]
0.30 0.00 2798595/3301773 update_queue [99]
[109] 0.3 0.35 0.00 3301773 try_atomic_semop [109]
-----------------------------------------------
Both number of function calls to try_atomic_semop() and update_queue()
are reduced by 50% as a result of the merge. Execution time of
sys_semtimedop is reduced because of the reduction in the low level
functions.
unhandled content-type:application/octet-stream (sem25.patch)
--
=============================================================
Luis Miguel Garcia Mancebo
Ingenieria Tecnica en Informatica de Gestion
Universidad de Deusto / University of Deusto
Bilbao / Spain
=============================================================
Well, it doesn't apply to current tree.
Perhaps already applied or something?
Thanks!
--
=============================================================
Luis Miguel Garcia Mancebo
Ingenieria Tecnica en Informatica de Gestion
Universidad de Deusto / University of Deusto
Bilbao / Spain
=============================================================
Last patch posted was relative to 2.5.73. It is in Andrew Morton's
2.5.73-mm2 tree.
- Ken
-----Original Message-----
From: Luis Miguel Garcia [mailto:[email protected]]
Sent: Tuesday, July 01, 2003 8:06 AM
To: [email protected]
Cc: Chen, Kenneth W
Subject: Re:ipc semaphore optimization
Well, it doesn't apply to current tree.
Perhaps already applied or something?
Thanks!
--
=============================================================
Luis Miguel Garcia Mancebo
Ingenieria Tecnica en Informatica de Gestion
Universidad de Deusto / University of Deusto
Bilbao / Spain
=============================================================