Project

General

Profile

Actions

Feature #2230

closed

Speed up particle stack making without pre-manipulation on selected frame range on ddstack

Added by Anchi Cheng almost 12 years ago. Updated almost 9 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Category:
-
Target version:
Start date:
02/01/2013
Due date:
% Done:

0%

Estimated time:

Description

To select range of frames in ddstack movie and integrated into an image and then crop particle out currently involves saving integrated image in its full size and the cropping is done on the saved integrated image. It was designed so that ctf correction could be done on the whole image in the pipeline manner. However, for Frealign and Relion where ctf correction is refined within its own program, it would need no pre-boxing manipulation. Without the pre-manipulation, the boxing can be done much faster using a sum of memory mapping of part of the ddstack.


Related issues 1 (0 open1 closed)

Related to Appion - Feature #2604: debug print in apDisplayClosedAnchi Cheng12/18/2013

Actions
Actions #1

Updated by Anchi Cheng almost 12 years ago

  • Status changed from New to In Code Review
  • Assignee set to Neil Voss

r17371 breaks up boxParticlesFromImage so that I can add alternative path.
r17372 contains the main function.

Neil please assign David Veesler to test.

Testing:

Makes particle stack without ctf correction and with total frame not equal to zero on particle selection from ddstack image. You should see a speed increase while boxing.

Actions #2

Updated by Neil Voss over 11 years ago

  • Assignee changed from Neil Voss to Anchi Cheng

Hi Anchi,

I am worried about the line:
apDisplay.printMsg(' crop range of (x,y)=(%d,%d) to (%d,%d)' % (x1,y1,x2-1,y2-1))
While it is good for debugging, if you make a large stack it would quickly flood the log file.

Everything else looks good, so pass to David when done. Is the mmap significantly faster?

Actions #3

Updated by Anchi Cheng over 11 years ago

Neil,

Thanks for the concern. Shall we make a printDebug function in apDisplay?
Yes, mmap is faster when it comes to boxing particle of ddstack because they are 4kx4kx(number_of_frames).

Actions #4

Updated by Neil Voss over 11 years ago

Probably not a bad idea to have some sort of print debug. In my apCtf files, I have placed a True/False debug variable at the top of the file. So, a more elegant solution would be nice.

Actions #5

Updated by Anchi Cheng almost 11 years ago

  • Status changed from In Code Review to In Test
  • Assignee changed from Anchi Cheng to David Veesler

Next time when you create particle stack from frames fewer than the total number you acquired, this would take effect. Check if it prints out too much stuff.

Actions #6

Updated by Anchi Cheng almost 9 years ago

  • Status changed from In Test to Closed
  • Target version set to Appion/Leginon 3.1.0
Actions

Also available in: Atom PDF