[OpenSPIM] OpenSPIM 64-bit [0000129]

[Michael Weber]

Hey Nikul,

a couple of thoughts regarding the speed:

- Writing a lot of small files takes a lot more time than a single big one. You might want to test your RAID system with something like ATTO Disk Benchmark, which measures data rates for different files sizes.
- Buffering the data in the RAM before writing to SSD could help, or streaming everything to RAM. Requires a lot of RAM, though.
- If you intend to run z-stacks at 100 fps and the step speed of the z-motor is limiting, you could potentially run the motor continuously while streaming data.

Michael


On Feb 3, 2014, at 5:19 PM, Nikul <nikul at ee.columbia.edu> wrote:

> Hi Luke and Johannes,
> 
> I was able to figure out what the problem was. MMCoreJJNI.java lists
> the root directory and mm/Win64 as the search directories for a 64-bit
> build while openspim.build.xml was staging the files to mm/x64. Changing
> the staging directory in  openspim.build.xml solved the problem!
> 
> On another note, I am trying to interface a 100 fps 5.5mp camera with OpenSPIM 
> and am using a RAID0 array of four ssd drives. The write speed I get through 
> OpenSPIM (or micromanager) is at the moment only able to handle about 
> 50 fps. Before I dig into the code to see if there is something I can optimize for 
> the particular setup, I wanted to check if anyone has any tips to achieve write speeds 
> to get full frame rate capture from the camera.
> 
> Thanks,
> ---
> Nikul Ukani
> 
> Doctoral Candidate,
> Dept. of Electrical Engineering,
> Columbia University.
> http://www.bionet.ee.columbia.edu/
> 
> On Jan 31, 2014, at 4:42 PM, Nikul <nikul at ee.columbia.edu> wrote:
> 
>> I'll look into it and get back. If it works, it will be great to have some resource
>> on the OpenSPIM website for compiling it in 64-bit.
>> 
>> Thanks,
>> ---
>> Nikul Ukani
>> 
>> Doctoral Candidate,
>> Dept. of Electrical Engineering,
>> Columbia University.
>> http://www.bionet.ee.columbia.edu/
>> 
>> On Jan 31, 2014, at 4:29 PM, Luke Stuyvenberg <stuyvenberg at wisc.edu> wrote:
>> 
>>> Hi Nikul,
>>> 
>>> --x64 is supposed to change "mm.architecture" to "x64" before calling Micro-Manager's build scripts; this should build MM in 64-bit. However, it's entirely probable that "build.sh" and/or "openspim.build.xml" do not copy the 64-bit files -- I would check first to see if the relevant binaries exist in "stage" or "build" subdirectories. If so, you'll have to look through the build system to see where we stage this files (I believe the relevant target is near the end of "openspim.build.xml") and make sure the correct architecture is copied from.
>>> 
>>> 
>>> Hope this helps,
>>> Luke
>>> 
>>> On 01/31/14, Nikul Ukani  wrote:
>>>> Hi Johannes,
>>>> 
>>>> That's possible as all I did was to change the platform to x64 in the build script. Can you point me to which project does that library correspond to so I can investigate whether the correct flag is being passed while compiling it?
>>>> 
>>>> Thanks,
>>>> Nikul
>>>> 
>>>>> On Jan 31, 2014, at 2:21 PM, Johannes Schindelin <Johannes.Schindelin at gmx.de> wrote:
>>>>> 
>>>>> Hi Nikul,
>>>>> 
>>>>>> On Fri, 31 Jan 2014, Nikul Ukani wrote:
>>>>>> 
>>>>>> I am currently trying to compile OpenSPIM on a 64 bit machine running
>>>>>> Windows 7. While I was able to get it to compile by modifying some of
>>>>>> the build scripts; upon trying to open it from Fiji, I get the following
>>>>>> error
>>>>>> 
>>>>>> Error: Failed to open libMMCoreJ_wrap.jnilib
>>>>>> mmcorej.MMCoreJJNI.swig_module_init()V
>>>>> 
>>>>> My guess is that the libMMCoreJ_wrap.jnilib is compiled as a 32-bit
>>>>> library. Possible?
>>>>> 
>>>>> Ciao,
>>>>> Johannes
>>>>> 
>>>> 
>>>> 
>>>> _______________________________________________
>>>> OpenSPIM mailing list
>>>> OpenSPIM at openspim.org
>>>> http://openspim.org/mailman/listinfo/openspim
>>> 
>> 
> 
> _______________________________________________
> OpenSPIM mailing list
> OpenSPIM at openspim.org
> http://openspim.org/mailman/listinfo/openspim

_____________

Michael Weber
PhD Student, Huisken lab
Max Planck Institute of Molecular Cell Biology and Genetics
Pfotenhauerstrasse 108, 01307 Dresden
Tel. 0049 351/2102837

http://www.mpi-cbg.de/huisken

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openspim.org/pipermail/openspim/attachments/20140204/c01a883d/attachment-0002.html>