[yt-dev] Optionally resetting the random number generator stream in Rockstar

[Matthew Turk]

Hi Cameron,

Sort of. Behavior will be unchanged unless you specifically rebuild from my
version. My version is the one that the install script grabs.

On Fri, Dec 19, 2014, 4:30 PM Cameron Hummels <chummels at gmail.com> wrote:

> Good detective work, Matt.  Thanks for this!
>
> Just to clarify, the change you made was in the version of Rockstar that
> gets automatically built/installed with yt if you ask the install_script.sh
> to install Rockstar?  So this will be fixed for people who install Rockstar
> from here on out, and for those of us who already have Rockstar installed,
> we just need to make the compile time flag you suggest in the Makefile and
> rebuild rockstar?
>
> Thanks again!
>
> Cameron
>
> On Fri, Dec 19, 2014 at 4:32 PM, Matthew Turk <matthewturk at gmail.com>
> wrote:
>
>> Hi all,
>>
>> I've just pushed a change to the fork of Rockstar I have on bitbucket
>> that adds a default-off compile-time setting that governs whether or not
>> the random stream is reset following the call to "random_sleep" that
>> governs how the clients connect to the server.
>>
>>
>> https://bitbucket.org/MatthewTurk/rockstar/commits/cf88fc3f9c1c9f0768005be5a07f816b0c79aaa8
>>
>> This addresses something that's come up on this list before.  I think it
>> should make runs completely deterministic (it does in my case) and will fix
>> the halo finding and halo analysis tests.  They are now, in my tests,
>> bitwise identical between runs whereas before I was getting definitely
>> non-bitwise results.  (This is noted in the paper.)
>>
>> I wanted to make this a parameter, but I wasn't able to figure out how to
>> do it without breaking API compatibility.  So eventually, this should make
>> its way upstream, but for now, it should (if compiled correctly!) fix the
>> tests.
>>
>> If you're building rockstar at home:
>>
>>  RNG_FLAGS=-DRESET_RNG_STREAM make clean lib
>>
>> should get it for you.
>>
>> -Matt
>>
>> _______________________________________________
>> yt-dev mailing list
>> yt-dev at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>
>>
>
> --
> Cameron Hummels
> Postdoctoral Researcher
> Steward Observatory
> University of Arizona
> http://chummels.org
>  _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-dev-spacepope.org/attachments/20141220/7720efcc/attachment.htm>
-------------- next part --------------
_______________________________________________
yt-dev mailing list
yt-dev at lists.spacepope.org
http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org