You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@mesos.apache.org by Klaus Ma <kl...@cguru.net> on 2015/09/03 16:01:27 UTC

Review Request 38102: MESOS-3046

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/
-----------------------------------------------------------

Review request for mesos and Ben Mahler.


Repository: mesos


Description
-------

MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)


Diffs
-----

  3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb0c572923a50490ecb17867dc50c6d161d 

Diff: https://reviews.apache.org/r/38102/diff/


Testing
-------

make
make check


Thanks,

Klaus Ma


Re: Review Request 38102: MESOS-3046

Posted by Mesos ReviewBot <re...@mesos.apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/#review97639
-----------------------------------------------------------


Patch looks great!

Reviews applied: [38102]

All tests passed.

- Mesos ReviewBot


On Sept. 3, 2015, 2:01 p.m., Klaus Ma wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38102/
> -----------------------------------------------------------
> 
> (Updated Sept. 3, 2015, 2:01 p.m.)
> 
> 
> Review request for mesos and Ben Mahler.
> 
> 
> Bugs: MESOS-3046
>     https://issues.apache.org/jira/browse/MESOS-3046
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb0c572923a50490ecb17867dc50c6d161d 
> 
> Diff: https://reviews.apache.org/r/38102/diff/
> 
> 
> Testing
> -------
> 
> make
> make check
> 
> 
> Thanks,
> 
> Klaus Ma
> 
>


Re: Review Request 38102: MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)

Posted by Klaus Ma <kl...@cguru.net>.

> On Sept. 14, 2015, 11:25 p.m., Ben Mahler wrote:
> > 3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp, line 31
> > <https://reviews.apache.org/r/38102/diff/1/?file=1063282#file1063282line31>
> >
> >     Please use THREAD_LOCAL from stout/thread_local.hpp and store a heap object as done in all of our other uses of THREAD_LOCAL.

Addressed :).


- Klaus


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/#review98950
-----------------------------------------------------------


On Sept. 15, 2015, 9:39 a.m., Klaus Ma wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38102/
> -----------------------------------------------------------
> 
> (Updated Sept. 15, 2015, 9:39 a.m.)
> 
> 
> Review request for mesos and Ben Mahler.
> 
> 
> Bugs: MESOS-3046
>     https://issues.apache.org/jira/browse/MESOS-3046
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> __Phenomenon:__
> Performance downgrade
> 
> __Root Cause:__
> stout's UUID abstraction is re-seeding the random generator during each call to UUID::random(), which is really expensive.
> 
> __Solution:__
> Seeding the random generator only once per thread.
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb 
> 
> Diff: https://reviews.apache.org/r/38102/diff/
> 
> 
> Testing
> -------
> 
> make
> make check
> 
> 
> Thanks,
> 
> Klaus Ma
> 
>


Re: Review Request 38102: MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)

Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/#review98950
-----------------------------------------------------------



3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp (line 31)
<https://reviews.apache.org/r/38102/#comment155715>

    Please use THREAD_LOCAL from stout/thread_local.hpp and store a heap object as done in all of our other uses of THREAD_LOCAL.


- Ben Mahler


On Sept. 5, 2015, 3:08 a.m., Klaus Ma wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38102/
> -----------------------------------------------------------
> 
> (Updated Sept. 5, 2015, 3:08 a.m.)
> 
> 
> Review request for mesos and Ben Mahler.
> 
> 
> Bugs: MESOS-3046
>     https://issues.apache.org/jira/browse/MESOS-3046
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> __Phenomenon:__
> Performance downgrade
> 
> __Root Cause:__
> stout's UUID abstraction is re-seeding the random generator during each call to UUID::random(), which is really expensive.
> 
> __Solution:__
> Seeding the random generator only once per thread.
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb0c572923a50490ecb17867dc50c6d161d 
> 
> Diff: https://reviews.apache.org/r/38102/diff/
> 
> 
> Testing
> -------
> 
> make
> make check
> 
> 
> Thanks,
> 
> Klaus Ma
> 
>


Re: Review Request 38102: MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)

Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/#review99481
-----------------------------------------------------------

Ship it!


Thanks for your patience!


3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp (line 33)
<https://reviews.apache.org/r/38102/#comment156359>

    Typically we name types with a capital in the first letter, like 'Generator', but let's just remove the typedef here, it doesn't seem to be helping a lot?


- Ben Mahler


On Sept. 15, 2015, 9:39 a.m., Klaus Ma wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38102/
> -----------------------------------------------------------
> 
> (Updated Sept. 15, 2015, 9:39 a.m.)
> 
> 
> Review request for mesos and Ben Mahler.
> 
> 
> Bugs: MESOS-3046
>     https://issues.apache.org/jira/browse/MESOS-3046
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> __Phenomenon:__
> Performance downgrade
> 
> __Root Cause:__
> stout's UUID abstraction is re-seeding the random generator during each call to UUID::random(), which is really expensive.
> 
> __Solution:__
> Seeding the random generator only once per thread.
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb 
> 
> Diff: https://reviews.apache.org/r/38102/diff/
> 
> 
> Testing
> -------
> 
> make
> make check
> 
> 
> Thanks,
> 
> Klaus Ma
> 
>


Re: Review Request 38102: MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)

Posted by Klaus Ma <kl...@cguru.net>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/
-----------------------------------------------------------

(Updated Sept. 15, 2015, 9:39 a.m.)


Review request for mesos and Ben Mahler.


Changes
-------

Address comments


Bugs: MESOS-3046
    https://issues.apache.org/jira/browse/MESOS-3046


Repository: mesos


Description
-------

__Phenomenon:__
Performance downgrade

__Root Cause:__
stout's UUID abstraction is re-seeding the random generator during each call to UUID::random(), which is really expensive.

__Solution:__
Seeding the random generator only once per thread.


Diffs (updated)
-----

  3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb 

Diff: https://reviews.apache.org/r/38102/diff/


Testing
-------

make
make check


Thanks,

Klaus Ma


Re: Review Request 38102: MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)

Posted by Klaus Ma <kl...@cguru.net>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/
-----------------------------------------------------------

(Updated Sept. 5, 2015, 3:08 a.m.)


Review request for mesos and Ben Mahler.


Changes
-------

Update summary & description


Summary (updated)
-----------------

MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)


Bugs: MESOS-3046
    https://issues.apache.org/jira/browse/MESOS-3046


Repository: mesos


Description (updated)
-------

__Phenomenon:__
Performance downgrade

__Root Cause:__
stout's UUID abstraction is re-seeding the random generator during each call to UUID::random(), which is really expensive.

__Solution:__
Seeding the random generator only once per thread.


Diffs
-----

  3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb0c572923a50490ecb17867dc50c6d161d 

Diff: https://reviews.apache.org/r/38102/diff/


Testing
-------

make
make check


Thanks,

Klaus Ma


Re: Review Request 38102: MESOS-3046

Posted by Klaus Ma <kl...@cguru.net>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38102/
-----------------------------------------------------------

(Updated Sept. 3, 2015, 2:01 p.m.)


Review request for mesos and Ben Mahler.


Bugs: MESOS-3046
    https://issues.apache.org/jira/browse/MESOS-3046


Repository: mesos


Description
-------

MESOS-3046 (Stout's UUID re-seeds a new random generator during each call to UUID::random)


Diffs
-----

  3rdparty/libprocess/3rdparty/stout/include/stout/uuid.hpp e90dabb0c572923a50490ecb17867dc50c6d161d 

Diff: https://reviews.apache.org/r/38102/diff/


Testing
-------

make
make check


Thanks,

Klaus Ma