| ... | ... | @@ -12,7 +12,7 @@ Divya |
|
|
|
|
|
|
|
Rachael
|
|
|
|
|
|
|
|
Char / Minute Taker / Focus Session [Rota](https://git.ligo.org/groups/gstlal/-/wikis/West-call/Rota-table)
|
|
|
|
Char / Minute Taker / Focus Session [Rota](https://git.ligo.org/groups/gstlal/-/wikis/West-call/Rota-table) --
|
|
|
|
|
|
|
|
## Action Items for next week
|
|
|
|
(**Leave action items here for next week.**)
|
| ... | ... | @@ -22,35 +22,155 @@ Char / Minute Taker / Focus Session [Rota](https://git.ligo.org/groups/gstlal/-/ |
|
|
|
|
|
|
|
* Announcements (5 minutes)
|
|
|
|
- Please check the rota for next week's call
|
|
|
|
- No update for EW next week. Need to find someone else!
|
|
|
|
- Confirmation of next week's focus session
|
|
|
|
- [O4 critical dev tasks from F2F in Pasadena](https://docs.google.com/document/d/1NFV9IaX4gUDPNCLBh_4AeUE7mxn4hzvPDBcoSS9YsjI/edit#heading=h.lxg2zjhsps9q)
|
|
|
|
- **F2F in Milwaukee: July 25 - July 29**
|
|
|
|
|
|
|
|
* Last week's call
|
|
|
|
- none
|
|
|
|
- Brief round table.
|
|
|
|
* Last week's East call - [agenda](https://wiki.ligo.org/CBC/EastAgenda20220510)
|
|
|
|
- dtdphi MR w/ feedback from Chad & Kipp
|
|
|
|
|
|
|
|
* Quick updates (45 minutes)
|
|
|
|
- Operations (5 minutes)
|
|
|
|
- LL CBC operations
|
|
|
|
- Plans for MDC05 (06/21):
|
|
|
|
- p(astro)
|
|
|
|
- p(astro) -- no update
|
|
|
|
- new template bank
|
|
|
|
- bank sims for sbank and manifold launched.
|
|
|
|
- idq online
|
|
|
|
- now running w/o failure in gap and no gap case
|
|
|
|
- want to make some changes to C script before MR.
|
|
|
|
- Jacob on ICDS
|
|
|
|
- when relaunching, takes hours for all jobs to move from idle --> running
|
|
|
|
- Ron will bring this up at HTC Condor meeting today
|
|
|
|
- Nothing new added currently except for patch for hanging gstlal_inspiral jobs.
|
|
|
|
- Offline CBC operations
|
|
|
|
- Shio and Prthamesh have both been running offline analyses
|
|
|
|
- LL IDQ operations
|
|
|
|
- plans in place to run offline iDQ runs to get update timeseries. This timeseries will be put in MDC frames, so pipelines have access to it during future MDCs.
|
|
|
|
- O4 Dev (30 minutes)
|
|
|
|
- Low latency integrated testing and Monitoring
|
|
|
|
- working on prepping for the MDC (becca)
|
|
|
|
- Template bank
|
|
|
|
- launched template bank generation dag on CIT by making broader bank first -> overpopulated BBH bank to fix missing template issues from F2F (Shio)
|
|
|
|
- Likelihood ratio, background and foreground sampling
|
|
|
|
- Optimization and throughput benchmarking
|
|
|
|
- Data format wrangling
|
|
|
|
- met last Fri to sort task assignment for snr/chisq distribution & specific tasks have been assigned. (leo)
|
|
|
|
- Andre and Leo may have something to present by next week.
|
|
|
|
- Optimization and throughput benchmarking -- no update
|
|
|
|
- Data format wrangling -- no update
|
|
|
|
- Enabling running on OSG/IGWN grid
|
|
|
|
- DQ dev
|
|
|
|
- HM search
|
|
|
|
- Exploratory development (5 minutes)
|
|
|
|
- Misc projects
|
|
|
|
- dag is almost done, but some jobs are not creating output files and being marked as done anyway. (Cort)
|
|
|
|
- DQ dev -- no update
|
|
|
|
- HM search -- no update
|
|
|
|
- Exploratory development -- no update
|
|
|
|
- Misc projects -- no update
|
|
|
|
|
|
|
|
* Focus session
|
|
|
|
- Data used in catalog searches - Surabhi [slides](https://docs.google.com/presentation/d/1gRaQop7riDWI64d5gRHQZ33wLN1zuno-RR4Tp8phLuQ/edit?usp=sharing)
|
|
|
|
- analyzing C00 data with search pipelines, so should be sufficient. (Jolien)
|
|
|
|
- Will DetChar be de-glitching C00 data? (Jolien)
|
|
|
|
- Push to use C00 wherever possible to avoid delays. PE tends to delay until final candidate list is available, but they will be running PE on good LL candidates during O4.(Surabhi)
|
|
|
|
- C00 will be similar to C01, but they won't be identical. (Surabhi)
|
|
|
|
- C00 will be the same as LL calibrated data, but will be different than front-end calibration data. If front end calibrated data is used, then it will be released according to this [doc](https://dcc.ligo.org/DocDB/0182/M2200101/001/o4-data-release-operations.pdf) (surabhi)
|
|
|
|
- iDQ data will be released, but unlikely that its inputs will be (Patrick)
|
|
|
|
|
|
|
|
|
|
|
|
- For O3 catalog, did run on clean data but probably didn't make too much of diff for searches (Surabhi)
|
|
|
|
- We did run on C00 first for O3a to check for sticking points/noise sources, so should have some runs that are pretty similar b/w C00 and clean data if we want a direct comparison. Only difference should be the prior. (Ryan)
|
|
|
|
- Ryan and a grad student will look into a direct comparison b/w the two.
|
|
|
|
|
|
|
|
|
|
|
|
- If we went w/ option 1 (where we rerank LL results), then every month e.g. we could already rerank events. (Jolien)
|
|
|
|
- We will continue to collect bankground month to month though, and in order to rerank only with background from relevant time, will need to re-engineer. (Surabhi)
|
|
|
|
- Benefit reranking w/ 1 month latency is we don't have to wait for O4a to finish in order to upload final results to GraceDB. (Jolien)
|
|
|
|
- Might advise against removing background in order rerank, but it was kind of hacky. To get fast results, makes sense to rerank candidates earlier in the search w/ current full background. Could even provide results after first 3 months e.g. and then update after end of O3a. (Surabhi)
|
|
|
|
- Wouldn't be hard to remove background, but did have to write stand-alone code for 190425 to do so. (Surabhi)
|
|
|
|
|
|
|
|
|
|
|
|
- Wouldn't reranking data w/ all of O4a be better than just with one month e.g. (Divya)
|
|
|
|
- Would be better, but would have to wait a full year (Jolien)
|
|
|
|
- But wouldn't have to wait a year. Could wait just the 3 months while we wait for C01 calibration data. Then could do so again at the 6 month, 9 month e.g. If we wanted to do it month by month, could do 1 month, 2 month, etc but use all the background available up until that point. (Surabhi)
|
|
|
|
|
|
|
|
- If you rerank everything w/ a year worth of background, then the difference between proposed option 1 and 2 isn't really big. (Jolien)
|
|
|
|
- When we rerank, should pick up new LR code, but based on SNR/chisq anyway, so shouldn't matter. (Surabhi)
|
|
|
|
|
|
|
|
- A problem with option 1 is that sometimes data gets dropped, and then becomes available later. This would not catch that. (Jolien)
|
|
|
|
- What is the chance that we have something special in a data drop like that? (surabhi)
|
|
|
|
- Virgo wasn't available for GW170817. (Jolien)
|
|
|
|
|
|
|
|
- Talk about it again at F2F or at later meeting.
|
|
|
|
|
|
|
|
|
|
|
|
* AOB
|
|
|
|
|
|
|
|
## Chat log |
|
|
\ No newline at end of file |
|
|
|
## Chat log
|
|
|
|
<09:30:33> "Divya Singh": https://git.ligo.org/groups/gstlal/-/wikis/West-call/060722
|
|
|
|
<09:35:46> "Rachael Huxford": PSU people, we should chat with Chad and see what the plan is on that.
|
|
|
|
<09:39:40> "Becca Ewing": i'll put it on the agenda
|
|
|
|
<09:42:57> "Becca Ewing": what's new that's been added to jacob as of now?
|
|
|
|
<09:43:02> "ShioSakon": Sorry, one more thing for template bank. I've launched a sbank template bank generation dag on cit. (This is to make a broad bank and make a overpop bbh bank from that)
|
|
|
|
<09:43:14> "ron.tapia": hand up
|
|
|
|
<09:43:28> "Becca Ewing": oh 
|
|
|
|
<09:43:42> "Becca Ewing": oh nice !
|
|
|
|
<09:43:55> "Becca Ewing": that's awesome
|
|
|
|
<09:44:17> "Rachael Huxford": Oh awesome - thanks
|
|
|
|
<09:44:44> "Becca Ewing": dont think so
|
|
|
|
<09:45:50> "Rachael Huxford": If anyone has anything they want to add to Jacob - please just message the LL channel and go for it if you're ready!
|
|
|
|
<09:46:21> "Becca Ewing": @rachael @prathamesh aree you using the count tracker in jacob? like have you tried actually removing counts?
|
|
|
|
<09:46:56> "Rachael Huxford": Not to my knowledge Becca, but the count tracker is in place
|
|
|
|
<09:47:30> "Prathamesh Joshi": I can take a look at that, trying to see if there is contamination and if the count tracker removes them
|
|
|
|
<09:47:47> "Becca Ewing": it would be nice to test that out before next mdc i think!
|
|
|
|
<09:47:57> "Prathamesh Joshi": Is Jacob burned in?
|
|
|
|
<09:48:00> "Leo Tsukada": +1 for Becca
|
|
|
|
<09:52:02> "Andre Guimaraes": After next week I will not be readily availbable. It would be great if we could present next week
|
|
|
|
<09:52:24> "Andre Guimaraes": Thank you
|
|
|
|
<09:52:26> "SurabhiSachdev": +1
|
|
|
|
<09:53:32> "SurabhiSachdev": what is the message again?
|
|
|
|
<09:54:35> "ShioSakon": I've seen it when I launched a dag on osg. Didn't investigate that time because I launched an identical dag on icds and proceeded with that
|
|
|
|
<09:54:52> "Rachael Huxford": Maybe we can follow-up at meeting this afternoon?
|
|
|
|
<09:54:58> "ShioSakon": Ok
|
|
|
|
<09:55:00> "ron.tapia": hand up
|
|
|
|
<09:55:21> "Rachael Huxford": hand up
|
|
|
|
<09:55:38> "Rachael Huxford": Oh sure no worries
|
|
|
|
<09:55:46> "Becca Ewing": +1
|
|
|
|
<09:55:48> "Rachael Huxford": Yeah you're right!
|
|
|
|
<09:55:59> "Rachael Huxford": no update
|
|
|
|
<09:56:10> "stefano.schmidt": no update
|
|
|
|
<09:57:25> "ShioSakon": I think in like 30 min
|
|
|
|
<09:57:25> "Andre Guimaraes": I'll have to leave soon today. Thank you all!
|
|
|
|
<09:57:31> "SurabhiSachdev": thanks Andre
|
|
|
|
<09:57:41> "SurabhiSachdev": https://docs.google.com/presentation/d/1gRaQop7riDWI64d5gRHQZ33wLN1zuno-RR4Tp8phLuQ/edit#slide=id.p
|
|
|
|
<10:07:18> "Leo Tsukada": I always love less work for us
|
|
|
|
<10:07:40> "Leo Tsukada": so maybe option1?
|
|
|
|
<10:08:10> "Becca Ewing": or we can just procrastinate until c001, its a future us problem 
|
|
|
|
<10:09:05> "patrick.godwin": it shouldn't
|
|
|
|
<10:09:07> "Rachael Huxford": I wouldn't think so
|
|
|
|
<10:09:40> "ryan.magee": Low-latency calibrated data is exactly the same as C00, right?
|
|
|
|
<10:09:45> "SurabhiSachdev": yes
|
|
|
|
<10:10:15> "jolien.creighton": How about front-end calibration data?
|
|
|
|
<10:10:56> "SurabhiSachdev": https://dcc.ligo.org/DocDB/0173/T2100015/005/O4-calibration-requirements.pdf
|
|
|
|
<10:11:01> "Leo Tsukada": sorry, what is front-end cal data?
|
|
|
|
<10:11:25> "SurabhiSachdev": https://dcc.ligo.org/DocDB/0182/M2200101/001/o4-data-release-operations.pdf
|
|
|
|
<10:11:54> "patrick.godwin": it's used in iDQ 
|
|
|
|
<10:11:58> "patrick.godwin": but we also use the aux channels
|
|
|
|
<10:12:01> "patrick.godwin": so...yeah
|
|
|
|
<10:12:14> "Rachael Huxford": good luck releasing all the aux channels XD
|
|
|
|
<10:12:19> "patrick.godwin": yeah, there is
|
|
|
|
<10:15:02> "ryan.magee": Yes, it was mandated in O3 by all-sky. we would probably be able to produce a few rough comparisons to c00 for searches
|
|
|
|
<10:16:17> "SurabhiSachdev": GWTC-2 was on C00 at elast partially anyway
|
|
|
|
<10:17:22> "ryan.magee": i can add that as an action for myself / a grad student that would like to help (hint)
|
|
|
|
<10:17:28> "SurabhiSachdev": perfect
|
|
|
|
<10:17:42> "ryan.magee": haha awesome. I'll message you/others
|
|
|
|
<10:18:14> "SurabhiSachdev": hand up
|
|
|
|
<10:21:40> "SurabhiSachdev": we did it for 190425 actually
|
|
|
|
<10:21:42> "SurabhiSachdev": the paper
|
|
|
|
<10:24:24> "ryan.magee": Alternatively, there's no reason we couldn't update a trigger if we were concerned about understating the significance
|
|
|
|
<10:32:00> "SurabhiSachdev": 
|
|
|
|
<10:32:51> "jolien.creighton": like GW170817
|
|
|
|
<10:33:34> "ryan.magee": yeah, the analysis timed out
|
|
|
|
<10:33:47> "Leo Tsukada": how about refiltering C00 only for the chunks that have had missing data?
|
|
|
|
<10:39:02> "Becca Ewing": byeee
|
|
|
|
<10:39:02> "Prathamesh Joshi": bye!
|
|
|
|
<10:39:07> "SurabhiSachdev": thanks bye |
|
|
\ No newline at end of file |
