[OpenSPIM] very slow acquisition speed

[Robert Hauschild]

Hi,

Monika:
> "What is going on" should be relatively easy if you look at the 
> timestamps in the log that you send. "Why" and "How to fix it" is a 
> different matter unfortunately.
I put together a time line that visualizes the log file I sent with the 
first email (single slice no stage movement).
 From this I learn that it somehow takes 300 ms to acquire 1 image with 
an exposure time of 100 ms. After that it takes an additional 250 ms 
until the laser is told to turn off. That is ~400 ms unaccounted for.
As I mentioned  I still don't understand why the stage is even engaged 
and the laser power  is queried during acquisition.

> Did you try to take a time series only with Micromanager, without the 
> OpenSPIM plugin? This should be possible somehow.
As mentioned I can record a time series at full camera frame rate with 
Micromanager alone. No problem whatsoever.

Johannes :
> The profiler (which one are you using? 
I am using "Profile acquisition" in the SPIM plugin.
I will try to get the results from the Flight Recorder of Java Mission 
Control.

Best regards,

Robert

>
> Regards,
> Monika
>
> W dniu .05.2015 o 19:00 <openspim-request at openspim.org> pisze:
>
>> Send OpenSPIM mailing list submissions to
>>     openspim at openspim.org
>>
>> To subscribe or unsubscribe via the World Wide Web, visit
>>     http://openspim.org/mailman/listinfo/openspim
>> or, via email, send a message with subject or body 'help' to
>>     openspim-request at openspim.org
>>
>> You can reach the person managing the list at
>>     openspim-owner at openspim.org
>>
>> When replying, please edit your Subject line so it is more specific
>> than "Re: Contents of OpenSPIM digest..."
>>
>>
>> Today's Topics:
>>
>>    1. Re: very slow acquisition speed (Robert Hauschild)
>>
>>
>> ----------------------------------------------------------------------
>>
>> Message: 1
>> Date: Tue, 12 May 2015 14:41:25 +0200
>> From: Robert Hauschild <robert.hauschild at ist.ac.at>
>> To: openspim at openspim.org
>> Subject: Re: [OpenSPIM] very slow acquisition speed
>> Message-ID: <5551F4F5.3050403 at ist.ac.at>
>> Content-Type: text/plain; charset=utf-8; format=flowed
>>
>> Dear Monika, All,
>>
>> thanks for testing! At least my system is not the only one that is so 
>> slow.
>>
>> Would be great if someone could tell me how to figure out what is going
>> on during the 75% of time that is unaccounted for in the profiler!
>>
>> thanks!
>>
>> Best regards,
>>
>> Robert
>>
>> On 07.05.2015 18:20, m.pawlowska at nencki.gov.pl wrote:
>>> Dear Robert,
>>>
>>> since I also have Orca+Picard, I decided to repeat your experiment and
>>> did 10 frames, 100 ms. I'm using the new 64-bit version, updated. I
>>> got very similar results to yours (see below). Unfortunately I don't
>>> know whether it used to be better, I never tried taking very fast
>>> measurements.
>>>
>>> Maybe somebody who is able to get 10 frames/sec could send their
>>> CoreLog for comparison...?
>>>
>>> Opened log file
>>> D:\Dane\2015\2015_05\2015_05_07\02time_series_test\log.txt
>>> Profiler("performAcquisition")[4 children]: 6,0611s
>>> 6,3337%:    Profiler("Output")[0 children]: 0,3839s
>>> 0,0000%:    Profiler("Acquisition")[0 children]: 0,0000s
>>> 29,7792%:    Profiler("Movement")[0 children]: 1,8049s
>>> 0,0005%:    Profiler("Setup")[0 children]: 0,0000s
>>>
>>> Regards,
>>> Monika
>>
>>> Hi All and especially Pete,
>>>
>>> My setup is Orca R2, Coherent Cube, Picard stage.
>>> I tested a couple of things and it it still unclear to me where the
>>> time is wasted.
>>>
>>> I attached the core log of a 10 frame long time series (100ms
>>> exposure) recorded to memory with X/Y/Z Stage/Theta all unchecked and
>>> without "Update Live preview".
>>>
>>> I tried two configurations:
>>> 1. Orca  + Cube + Picard
>>> 2. Orca  + Picard
>>>
>>> "1" took 7.6 s of which ("1"- "2") = 1.4 s are communication with
>>> laser, which, while way too much, is not the source of the problem.
>>>
>>> Profiler("performAcquisition")[4 children]: 7.6113s
>>> 0.0017%:    Profiler("Output")[0 children]: 0.0001s
>>> 0.0000%:    Profiler("Acquisition")[0 children]: 0.0000s
>>> 24.0133%:    Profiler("Movement")[0 children]: 1.8277s
>>> 0.0013%:    Profiler("Setup")[0 children]: 0.0001s
>>>
>>> So most of the time the system is doing something that the Profiler is
>>> not recording.
>>>
>>> Looking at the core log I am a bit puzzled why the laser power is
>>> queried once  per slice and why the stage is told to move to the very
>>> position it currently is even though none of the motors should move at
>>> all since X/Y/Z/Theta are all unchecked in the acquisition dialog.
>>>
>>> I would really appreciate any hint/suggestion on how to get back to
>>> reasonable acquisition speeds.
>>>
>>> Thanks again!
>>>
>>> Robert
>>
>
>


-- 

Robert Hauschild
staff scientist

Institute of Science and Technology Austria (IST Austria)
Am Campus 1
A – 3400 Klosterneuburg
Phone: +43 (0)2243 - 9000 - 1047
Email:   robert.hauschild at ist.ac.at

-------------- next part --------------
A non-text attachment was scrubbed...
Name: speedproblem.jpg
Type: image/jpeg
Size: 62746 bytes
Desc: not available
URL: <http://openspim.org/pipermail/openspim/attachments/20150513/5ff67be3/attachment-0002.jpg>