[clas12_rgk] [EXTERNAL] Re: Trains examples
Francois-Xavier Girod
fxgirod at jlab.org
Mon Jun 15 18:00:20 EDT 2020
Thank you, I will start with 10% then
I started with RGA because I wanted to get a high statistics test and
compare the final numbers I already had for these exclusive reactions
All wagons look safe on the high energy data
Yes I realize that we still need to validate the wagons on the RGK dataset,
I think it may be enough to validate with the lowest energy 6.5 GeV
At the moment the exclusivity cuts are hard coded. In principle I could
follow the same procedure Nathan wrote to read information from the YAML
configuration instead
However I hope things are sufficiently safe that way, the cuts have all
been widened
On Mon, Jun 15, 2020 at 11:55 PM Raffaella De Vita <
Raffaella.Devita at ge.infn.it> wrote:
> Dear FX,
> I would assume that 10% of a full run would be sufficient to get a first
> estimate, i.e. ~25 DSTs files.
> I would be good to run on RG-K DSTs: maybe Nick can point to the 4 runs
> that were cooked for the review.
> Best regards,
> Raffaella
>
> On 15 Jun 2020, at 17:49, Francois-Xavier Girod <fxgirod at jlab.org> wrote:
>
> Dear Raffaella
>
> Yes I realize now that we need the correct reduction factors, it is no
> problem I can run on the DSTs instead
> This is the information I have on the file location
> /cache/clas12/rg-a/production/recon/fall2018/torus-1/pass1/v0/dst/recon/
> /cache/clas12/rg-a/production/recon/fall2018/torus-1/pass1/v0/dst/recon/
>
> Do you have a suggestion many files I will need to process to get an
> accurate reduction factor?
>
> Best regards
> FX
>
> On Mon, Jun 15, 2020 at 11:37 PM Raffaella De Vita <
> Raffaella.Devita at ge.infn.it> wrote:
>
>> Dear FX,
>> If you have time, could you run on DSTs as well? It would be helpful
>> because the reduction factors can then be compared with what Nick had from
>> the standard wagons.
>> Thanks,
>> Raffaella
>>
>>
>> On 15 Jun 2020, at 12:24, Francois-Xavier Girod <fxgirod at jlab.org> wrote:
>>
>> Dear all
>>
>> Here is a summary of the tests I performed with the new exclusive
>> reaction wagons
>>
>> All deep exclusive wagons have Q2 > 0.8 GeV^2 and W > 1.8 GeV
>>
>> RGA OUT sample was skimmed starting from skim4 (already has the condition
>> of one e in the FD)
>> CPU usage for the new wagons compared to JpsiTCS (on clara machines)
>> 2020-06-14 22:13:46.433: Benchmark results:
>> 2020-06-14 22:13:46.433: READER 210369933 events total time
>> = 72775.73 s average event time = 0.35 ms
>> 2020-06-14 22:13:46.434: JPSITCS 210369933 events total time
>> = 33937.10 s average event time = 0.16 ms
>> 2020-06-14 22:13:46.434: DVCSWagon 210369933 events total time
>> = 27440.98 s average event time = 0.13 ms
>> 2020-06-14 22:13:46.434: DVPi0PWagon 210369933 events total time
>> = 33836.22 s average event time = 0.16 ms
>> 2020-06-14 22:13:46.434: DVPipPimPWag 210369933 events total time
>> = 25329.07 s average event time = 0.12 ms
>> 2020-06-14 22:13:46.434: DVPipPimPi0P 210369933 events total time
>> = 25733.89 s average event time = 0.12 ms
>> 2020-06-14 22:13:46.434: DVKpKmPWagon 210369933 events total time
>> = 25776.63 s average event time = 0.12 ms
>> 2020-06-14 22:13:46.434: WRITER 210369933 events total time
>> = 17867.52 s average event time = 0.08 ms
>> 2020-06-14 22:13:46.435: TOTAL 210369933 events total time
>> = 262697.14 s average event time = 1.25 ms
>> 2020-06-14 22:13:46.435: Average processing time = 0.38 ms
>> 2020-06-14 22:13:46.435: Total processing time = 80223.88 s
>> 2020-06-14 22:13:46.435: Total orchestrator time = 80521.07 s
>> 2020-06-14 22:13:46.435: Processing is complete.
>>
>> Reduction factors over skim4:
>> JpsiTCS size reduction factor ~ 225
>> DVCS ~ 342
>> DVPi0P ~ 373
>> DVPipPimP ~ 754
>> DVPipPimPi0P ~ 677
>> DVKpKmP ~ 2068
>>
>> RGA IN sample was skimmed from skim8 (already has the condition of e in
>> FD and one proton)
>> CPU usage
>> 2020-06-15 09:05:28.440: Benchmark results:
>> 2020-06-15 09:05:28.440: READER 16489802 events total time =
>> 6310.92 s average event time = 0.38 ms
>> 2020-06-15 09:05:28.441: JPSITCS 16489802 events total time =
>> 2612.12 s average event time = 0.16 ms
>> 2020-06-15 09:05:28.441: DVCSWagon 16489802 events total time =
>> 2391.67 s average event time = 0.15 ms
>> 2020-06-15 09:05:28.441: DVPi0PWagon 16489802 events total time =
>> 4054.14 s average event time = 0.25 ms
>> 2020-06-15 09:05:28.442: DVPipPimPWag 16489802 events total time =
>> 2046.81 s average event time = 0.12 ms
>> 2020-06-15 09:05:28.442: DVPipPimPi0P 16489802 events total time =
>> 2184.76 s average event time = 0.13 ms
>> 2020-06-15 09:05:28.442: DVKpKmPWagon 16489802 events total time =
>> 1979.42 s average event time = 0.12 ms
>> 2020-06-15 09:05:28.442: WRITER 16489802 events total time =
>> 1649.79 s average event time = 0.10 ms
>> 2020-06-15 09:05:28.442: TOTAL 16489802 events total time
>> = 23229.64 s average event time = 1.41 ms
>> 2020-06-15 09:05:28.442: Average processing time = 0.43 ms
>> 2020-06-15 09:05:28.442: Total processing time = 7151.69 s
>> 2020-06-15 09:05:28.442: Total orchestrator time = 7650.04 s
>> 2020-06-15 09:05:28.443: Processing is complete.
>>
>> Reduction factors over skim8:
>> JpsiTCS size reduction factor ~ 53
>> DVCS ~ 51
>> DVPi0P ~ 50
>> DVPipPimP ~ 66
>> DVPipPimPi0P ~ 55
>> DVKpKmP ~ 100
>>
>> Best regards
>> FX
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/clas12_rgk/attachments/20200616/f5e97bfe/attachment.html>
More information about the clas12_rgk
mailing list