Closed dwhan89 closed 5 years ago
Currently used seeds
(set_idx, 0, 0, sim_idx)
(set_idx, 0, 1, sim_idx, 0)
(set_idx, 0, 1, sim_idx, 1)
(set_idx, 0, 1, sim_idx, 2)
(0, 0, 2, sim_idx)
(set_idx, 0, 3, sim_num) + soapack.DataModel.get_noise_sim_seed()
(set_idx, 0, 4, sim_num)
Also noise sim seeds: https://github.com/simonsobs/soapack/blob/master/soapack/interfaces.py#L140
The noise sim seed stuff should be moved from soapack to actsims, but in a way that different experiments are distinguished.
PR filed for the issue (https://github.com/ACTCollaboration/actsims/pull/28/commits)
Looks great to me, thanks for moving this stuff into a nicely-defined class! Alex
On Jun 23, 2019, at 3:53 PM, dwhan89 notifications@github.com wrote:
PR filed for the issue (https://github.com/ACTCollaboration/actsims/pull/28/commits https://github.com/ACTCollaboration/actsims/pull/28/commits)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ACTCollaboration/actsims/issues/27?email_source=notifications&email_token=ADYBOHAPI34MG52INT7WRRDP37IE5A5CNFSM4H2ZPK42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYLFYAA#issuecomment-504781824, or mute the thread https://github.com/notifications/unsubscribe-auth/ADYBOHD2T6HAHMVH4HQUT3DP37IE5ANCNFSM4H2ZPK4Q.
PS I can sign off on this, Mat could you let me know if you agree (i.e. does it work well with your noise sim stuff?)
On Jun 24, 2019, at 1:18 PM, Alexander van Engelen engelen@cita.utoronto.ca wrote:
Looks great to me, thanks for moving this stuff into a nicely-defined class! Alex
On Jun 23, 2019, at 3:53 PM, dwhan89 <notifications@github.com mailto:notifications@github.com> wrote:
PR filed for the issue (https://github.com/ACTCollaboration/actsims/pull/28/commits https://github.com/ACTCollaboration/actsims/pull/28/commits)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ACTCollaboration/actsims/issues/27?email_source=notifications&email_token=ADYBOHAPI34MG52INT7WRRDP37IE5A5CNFSM4H2ZPK42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYLFYAA#issuecomment-504781824, or mute the thread https://github.com/notifications/unsubscribe-auth/ADYBOHD2T6HAHMVH4HQUT3DP37IE5ANCNFSM4H2ZPK4Q.
Oh actually, could you delete the old simTools.py? I think it is now fully deprecated (I had been keeping it around for this index stuff).
Also, could you do me a favor since you have eyes on it, and also change the phiSeed and cmbSeed in the CMB sim generation script to call the new routine. It’s this line https://github.com/ACTCollaboration/actsims/blob/3062dc1a27fd37bf358b5638c56bab80a3e02f46/bin/signalSims.py#L85 https://github.com/ACTCollaboration/actsims/blob/3062dc1a27fd37bf358b5638c56bab80a3e02f46/bin/signalSims.py#L85
Thanks! Alex
On Jun 24, 2019, at 1:20 PM, Alexander van Engelen engelen@cita.utoronto.ca wrote:
PS I can sign off on this, Mat could you let me know if you agree (i.e. does it work well with your noise sim stuff?)
On Jun 24, 2019, at 1:18 PM, Alexander van Engelen <engelen@cita.utoronto.ca mailto:engelen@cita.utoronto.ca> wrote:
Looks great to me, thanks for moving this stuff into a nicely-defined class! Alex
On Jun 23, 2019, at 3:53 PM, dwhan89 <notifications@github.com mailto:notifications@github.com> wrote:
PR filed for the issue (https://github.com/ACTCollaboration/actsims/pull/28/commits https://github.com/ACTCollaboration/actsims/pull/28/commits)
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ACTCollaboration/actsims/issues/27?email_source=notifications&email_token=ADYBOHAPI34MG52INT7WRRDP37IE5A5CNFSM4H2ZPK42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYLFYAA#issuecomment-504781824, or mute the thread https://github.com/notifications/unsubscribe-auth/ADYBOHD2T6HAHMVH4HQUT3DP37IE5ANCNFSM4H2ZPK4Q.
No problem! It's done.
Great - thanks!
On Jun 24, 2019, at 1:57 PM, dwhan89 notifications@github.com wrote:
No problem! It's done.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ACTCollaboration/actsims/issues/27?email_source=notifications&email_token=ADYBOHDJLGOPKEF5XLPZ7QLP4EDHFA5CNFSM4H2ZPK42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYNXBMI#issuecomment-505114801, or mute the thread https://github.com/notifications/unsubscribe-auth/ADYBOHGZAXQE5LHCVKWJLE3P4EDHFANCNFSM4H2ZPK4Q.
Each type of simulation (ex: CMB, lensing potential, foregrounds etc) needs unique seed in tuple. Currently, there is no central location inside of the code keeping a track of different kinds of seeds. This is very error-prone, and messing up seed structure can introduce subtle bugs that would be hard to track down. Let's introduce a unified seed structure once for all.